1 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
3 PR rtl-optimization/58438
4 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
5 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
6 from previous subpasses.
8 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
10 * arm.c (arm_get_frame_offsets): Validate architecture supports
11 LDRD/STRD before accepting the tuning preference.
12 (arm_expand_prologue): Likewise.
13 (arm_expand_epilogue): Likewise.
15 2013-09-18 Richard Biener <rguenther@suse.de>
17 PR tree-optimization/58417
18 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
19 have chrecs with symbols defined in the loop as operands.
20 (chrec_fold_multiply): Likewise.
21 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
22 parameters before folding binary operations.
23 (struct instantiate_cache_entry_hasher): Remove.
24 (struct instantiate_cache_type): Use a pointer-map.
25 (instantiate_cache_type::instantiate_cache_type): New function.
26 (instantiate_cache_type::get): Likewise.
27 (instantiate_cache_type::set): Likewise.
28 (instantiate_cache_type::~instantiate_cache_type): Adjust.
29 (get_instantiated_value_entry): Likewise.
30 (global_cache): New global.
31 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
32 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
33 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
34 (instantiate_scev_name): Adjust.
35 (instantiate_scev): Construct global instead of local cache.
36 (resolve_mixers): Likewise.
38 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
39 Paolo Carlini <paolo.carlini@oracle.com>
42 * doc/implement-cxx.texi: Fix references to the C++ standards.
44 2013-09-18 Jakub Jelinek <jakub@redhat.com>
46 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
47 * tree-vect-data-refs.c (vect_analyze_data_refs): For
48 simd_lane_access drs, update also DR_ALIGNED_TO.
50 2013-09-18 Marek Polacek <polacek@redhat.com>
53 * doc/extend.texi: Document no_sanitize_undefined attribute.
54 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
55 no_sanitize_undefined attribute.
57 2013-09-18 Nick Clifton <nickc@redhat.com>
59 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
60 (ASM_DECLARE_FUNCTION_NAME): Define.
62 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
64 * compare-elim.c (find_comparison_dom_walker): New class
65 (find_comparisons_in_bb): Rename to
66 find_comparison_dom_walker::before_dom_children
67 (find_comparisons): Adjust
68 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
70 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
71 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
72 (init_walk_dominator_tree): Remove declaration.
73 (fini_walk_dominator_tree): Remove declaration.
74 * fwprop.c (single_def_use_dom_walker): New class
75 (single_def_use_enter_block): Convert to
76 single_def_use_dom_walker::before_dom_children.
77 (single_def_use_leave_block): Convert to
78 single_def_use_dom_walker::after_dom_children.
79 (build_single_def_use_links): Adjust.
80 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
82 (find_candidates_in_block): Convert to
83 find_candidates_dom_walker::before_dom_children.
84 (execute_strength_reduction): Adjust.
85 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
86 (sese_dom_walker): New class.
87 (sese_dom_walker::sese_dom_walker): New constructor.
88 (sese_dom_walker::~sese_dom_walker): New destructor.
89 (build_sese_conditions_before): Convert to
90 sese_dom_walker::before_dom_children.
91 (build_sese_conditions_after): Convert to
92 sese_dom_walker::after_dom_children.
93 (build_poly_scop): Adjust
94 * tree-into-ssa.c (rewrite_dom_walker): New class
95 (rewrite_enter_block): Convert to
96 rewrite_dom_walker::before_dom_children.
97 (rewrite_leave_block): Convert to
98 rewrite_dom_walker::after_dom_children.
99 (rewrite_update_dom_walker): New class.
100 (rewrite_update_enter_block): Convert to
101 rewrite_update_dom_walker::before_dom_children.
102 (rewrite_update_leave_block): Convert to
103 rewrite_update_dom_walker::after_dom_children.
104 (rewrite_blocks, rewrite_into_ssa): Adjust.
105 (mark_def_dom_walker): New class.
106 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
107 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
108 (mark_def_sites_blocks): Convert to
109 mark_def_dom_walker::before_dom_children.
110 (mark_def_site_blocks): Remove.
111 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
112 (tree_ssa_dominator_optimize): Adjust.
113 (dom_thread_across_edge): Convert to method
114 dom_opt_dom_walker::thread_across_edge.
115 (dom_opt_enter_block): Convert to member function
116 dom_opt_dom_walker::before_dom_children.
117 (dom_opt_leave_block): Convert to member function
118 dom_opt_dom_walker::after_dom_children.
119 * tree-ssa-dse.c (dse_dom_walker): New class.
120 (dse_enter_block): Convert to member function
121 dse_dom_walker::before_dom_children.
122 (tree_ssa_dse): Adjust.
123 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
124 (determine_invariantness_stmt): Convert to method
125 invariantness_dom_walker::before_dom_children.
126 (determine_invariantness): Remove
127 (move_computations_dom_walker): New class.
128 (move_computations_stmt): Convert to method
129 move_computations_dom_walker::before_dom_children.
130 (move_computations, tree_ssa_lim): Adjust.
131 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
132 (nt_init_block): Make method
133 notrappping_dom_walker::before_dom_children.
134 (nt_fini_block): Make
135 method nontrapping_dom_walker::after_dom_children.
136 (get_non_trapping): Adjust.
137 * tree-ssa-pre.c (eliminate_dom_walker): New class.
138 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
139 (eliminate_leave_block): Make method.
140 eliminate_dom_walker::after_dom_children.
142 * tree-ssa-strlen.c (strlen_dom_walker): New class.
143 (strlen_enter_block): Make method
144 strlen_dom_walker::before_dom_children.
145 (strlen_leave_block): Make
146 method strlen_dom_walker::after_dom_children.
147 (tree_ssa_strlen): Adjust.
148 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
149 (tree_ssa_uncprop): Adjust.
150 (uncprop_leave_block): Make method
151 uncprop_dom_walker::after_dom_children.
152 (uncprop_leave_block): Make method
153 uncprop_dom_walker::before_dom_children.
155 2013-09-18 Bin Cheng <bin.cheng@arm.com>
157 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
158 before branch in same basic block. Check both src and dest of
161 2013-09-17 Nick Clifton <nickc@redhat.com>
163 * config/rl78/rl78-real.md (bf): New pattern.
165 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
166 (rl78_print_operand): Do not put a # before a %B.
167 * config/rl78/rl78.opt: Tweak doc strings.
169 2013-09-17 DJ Delorie <dj@redhat.com>
171 * config/rl78/constraints.md (Wcv): Allow up to $r31.
172 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
173 (rl78_option_override): Likewise, if -mallregs.
174 (is_virtual_register): Likewise.
175 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
176 (REGNO_OK_FOR_BASE_P): Likewise.
177 * config/rl78/rl78.opt (-mallregs): New.
179 2013-09-17 Nick Clifton <nickc@redhat.com>
181 * config/rl78/rl78.c (need_to_save): Change return type to bool.
182 For interrupt functions: save all call clobbered registers if the
183 interrupt handler is not a leaf function.
184 (rl78_expand_prologue): Always recompute the frame information.
185 For interrupt functions: only select bank 0 if one of the bank 0
186 registers is going to be psuhed.
188 2013-09-17 DJ Delorie <dj@redhat.com>
190 * config/rl78/constraints.md: For each W* constraint, rename to C*
191 and create a W* constraint that checks for an optional ES: prefix
193 * config/rl78/rl78.md (UNS_ES_ADDR): New.
194 (es_addr): New. Used to wrap far addresses.
195 * config/rl78/rl78-protos.h (rl78_es_addr): New.
197 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
198 wrapped far addresses.
199 (rl78_print_operand_1): Unwrap far addresses before processing.
200 (rl78_lo16): Wrap far addresses in unspecs.
203 (insn_ok_now): Check for not-yet-wrapped far addresses.
204 (transcode_memory_rtx): Properly re-wrap far addresses.
206 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
208 * config/sparc/t-rtems: Add leon3 multilibs.
210 2013-09-17 Cong Hou <congh@google.com>
212 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
213 when checking the dot production pattern. The type of rhs operand
214 of multiply is now checked correctly.
216 2013-09-17 Jeff Law <law@redhat.com>
218 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
219 edge implied equivalences into successor phis.
220 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
221 here from tree-ssa-threadedge.c.
222 (mark_threaded_blocks): When threading through a joiner, if both
223 successors of the joiner's clone reach the same block, verify the
224 PHI arguments are equal. If not, cancel the jump threading request.
225 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
226 tree-ssa-threadupdate.c
227 (thread_across_edge): Don't check PHI argument equality when
228 threading through joiner block here.
230 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
232 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
233 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
234 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
235 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
236 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
237 (has_undefined_value_p): New. Pass dependant parts of
238 ssa_undefined_value_p.
239 (uninit_undefined_value_p): Use has_undefined_value_p.
240 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
241 make_pass_early_warn_uninitialized): Move from tree-ssa.c
242 * tree-ssa.h: Adjust prototypes
244 2013-09-17 Jan Hubicka <jh@suse.cz>
247 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
248 * ipa-inline.c (can_inline_edge_p): Do not downgrade
249 FUNCTION_NOT_OPTIMIZED.
250 * ipa-inline-analysis.c (compute_inline_parameters): Function
251 not optimized is not inlinable unless it is alwaysinline.
252 (inline_analyze_function): Force calls in not optimized
253 function not inlinable.
255 2013-09-17 Jan Hubicka <jh@suse.cz>
258 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
260 * ipa.c (function_and_variable_visibility): Likewise.
261 * ipa-profile.c (ipa_profile): Likewise.
263 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
266 * cgraph.c (cgraph_function_body_availability): Check for ifunc
267 attribute, and don't inline the resolver in this case.
269 2013-09-17 Teresa Johnson <tejohnson@google.com>
271 * coverage.c (get_coverage_counts): Add missing newline.
273 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
275 PR tree-optimization/58088
276 * fold-const.c (mask_with_trailing_zeros): New function.
277 (fold_binary_loc): Make sure we don't recurse infinitely
278 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
279 Use mask_with_trailing_zeros where appropriate.
281 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
283 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
284 of 'prev' var to get better distance estimation.
286 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
288 * tree-inline.h (struct copy_body_data): Add transform_parameter.
289 * tree-inline.c (is_parameter_of): New predicate.
290 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
291 a parameter has been remapped.
292 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
293 (optimize_inline_calls): Initialize transform_parameter.
294 (copy_gimple_seq_and_replace_locals): Likewise.
295 (tree_function_versioning): Likewise.
296 (maybe_inline_call_in_expr): Likewise.
298 2013-09-17 Nick Clifton <nickc@redhat.com>
300 * config/msp430/msp430-protos.h: Add prototypes for new functions.
301 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
303 (is_attr_func): New function.
304 (msp430_is_interrupt_func): New function.
305 (is_naked_func): New function.
306 (is_reentrant_func): New function.
307 (is_critical_func): New function.
308 (msp430_start_function): Add annotations for function attributes.
309 (msp430_attr): New function.
310 (msp430_attribute_table): New.
311 (msp430_function_section): New function.
312 (TARGET_ASM_FUNCTION_SECTION): Define.
313 (msp430_builtin): New enum.
314 (msp430_init_builtins): New function.
315 (msp430_builtin_devl): New function.
316 (msp430_expand_builtin): New function.
317 (TARGET_INIT_BUILTINS): Define.
318 (TARGET_EXPAND_BUILTINS): Define.
319 (TARGET_BUILTIN_DECL): Define.
320 (msp430_expand_prologue): Add support for naked, interrupt,
321 critical and reentrant functions.
322 (msp430_expand_epilogue): Likewise.
323 (msp430_print_operand): Handle 'O' character.
324 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
326 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
327 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
328 (pushm): Use a 'n' rather than an 'i' constraint.
329 (msp_return): Add generation of the interrupt return instruction.
330 (disable_interrupts): New pattern.
331 (enable_interrupts): New pattern.
332 (push_intr_state): New pattern.
333 (pop_intr_state): New pattern.
334 (bic_SR): New pattern.
335 (bis_SR): New pattern.
336 * doc/extend.texi: Document MSP430 function attributes and builtin
339 2013-09-17 Richard Biener <rguenther@suse.de>
341 PR tree-optimization/58432
342 * tree-loop-distribution.c (tree_loop_distribution): Also
343 scan PHIs for outside loop uses and seed a partition from them.
345 2013-09-17 Bin Cheng <bin.cheng@arm.com>
347 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
348 (restructure_reference): Call backtrace_base_for_ref.
350 2013-09-17 Alan Modra <amodra@gmail.com>
353 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
355 2013-09-16 DJ Delorie <dj@redhat.com>
357 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
358 vregs location for RL78/G10.
359 (rl78_expand_prologue): Avoid SEL on G10.
360 (rl78_expand_epilogue): Likewise.
361 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
362 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
363 __RL78_G10__ when appropriate.
364 (ASM_SPEC): Pass -mg10 along to the assembler.
365 * config/rl78/rl78.md (sel_rb): Disable for G10.
366 * config/rl78/rl78.opt: Add -mg10 option.
367 * config/rl78/t-rl78: Add -mg10 multilib.
369 2013-09-16 Xinliang David Li <davidxl@google.com>
371 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
372 * omp-low.c (omp_max_vf): Ditto.
373 (expand_omp_simd): Ditto.
374 * tree-vectorizer.c (vectorize_loops): Ditto.
375 (gate_vect_slp): Ditto.
376 (gate_increase_alignment): Ditto.
377 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
378 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
379 (gate_tree_vectorize): Name change.
380 (tree_vectorize): Ditto.
381 (pass_vectorize::gate): Call new function.
382 (pass_vectorize::execute): Ditto.
383 opts.c: O3 default setting change.
384 (finish_options): Check new flag.
385 * doc/invoke.texi: Document new flags.
386 * common.opt: New flags.
388 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
390 * doc/tm.texi.in (Cond Exec Macros): Remove node.
391 (Condition Code): Don't reference it.
392 * doc/tm.texi: Regenerate.
394 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
397 * lra-constraints.c (undo_optional_reloads): Consider all optional
398 reload even if it did not get a hard reg.
400 2013-09-16 Teresa Johnson <tejohnson@google.com>
402 * dumpfile.c (dump_loc): Remove newline emission.
403 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
404 emission to dump_printf_loc calls where missing.
405 (vect_mark_for_runtime_alias_test): Ditto.
406 (vect_analyze_data_ref_dependence): Ditto.
407 (vect_analyze_data_ref_dependences): Ditto.
408 (vect_slp_analyze_data_ref_dependence): Ditto.
409 (vect_slp_analyze_data_ref_dependences): Ditto.
410 (vect_compute_data_ref_alignment): Ditto.
411 (vect_update_misalignment_for_peel): Ditto.
412 (vect_verify_datarefs_alignment): Ditto.
413 (vector_alignment_reachable_p): Ditto.
414 (vect_get_data_access_cost): Ditto.
415 (vect_enhance_data_refs_alignment): Ditto.
416 (vect_find_same_alignment_drs): Ditto.
417 (vect_analyze_data_refs_alignment): Ditto.
418 (vect_analyze_group_access): Ditto.
419 (vect_analyze_data_ref_access): Ditto.
420 (vect_analyze_data_ref_accesses): Ditto.
421 (vect_prune_runtime_alias_test_list): Ditto.
422 (vect_analyze_data_refs): Ditto.
423 (vect_create_addr_base_for_vector_ref): Ditto.
424 (vect_create_data_ref_ptr): Ditto.
425 (vect_grouped_store_supported): Ditto.
426 (vect_grouped_load_supported): Ditto.
427 * value-prof.c (check_counter): Ditto.
428 (check_ic_target): Ditto.
429 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
430 (vect_recog_widen_mult_pattern): Ditto.
431 (vect_recog_widen_sum_pattern): Ditto.
432 (vect_recog_over_widening_pattern): Ditto.
433 (vect_recog_widen_shift_pattern): Ditto.
434 (vect_recog_rotate_pattern): Ditto.
435 (vect_recog_vector_vector_shift_pattern): Ditto.
436 (vect_recog_divmod_pattern): Ditto.
437 (vect_recog_mixed_size_cond_pattern): Ditto.
438 (vect_recog_bool_pattern): Ditto.
439 (vect_pattern_recog_1): Ditto.
440 (vect_pattern_recog): Ditto.
441 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
442 (vect_is_simple_iv_evolution): Ditto.
443 (vect_analyze_scalar_cycles_1): Ditto.
444 (vect_get_loop_niters): Ditto.
445 (vect_analyze_loop_1): Ditto.
446 (vect_analyze_loop_form): Ditto.
447 (vect_analyze_loop_operations): Ditto.
448 (vect_analyze_loop_2): Ditto.
449 (vect_analyze_loop): Ditto.
450 (report_vect_op): Ditto.
451 (vect_is_slp_reduction): Ditto.
452 (vect_is_simple_reduction_1): Ditto.
453 (vect_get_known_peeling_cost): Ditto.
454 (vect_estimate_min_profitable_iters): Ditto.
455 (vect_model_reduction_cost): Ditto.
456 (vect_model_induction_cost): Ditto.
457 (get_initial_def_for_induction): Ditto.
458 (vect_create_epilog_for_reduction): Ditto.
459 (vectorizable_reduction): Ditto.
460 (vectorizable_induction): Ditto.
461 (vectorizable_live_operation): Ditto.
462 (vect_loop_kill_debug_uses): Ditto.
463 (vect_transform_loop): Ditto.
464 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
465 (vect_stmt_relevant_p): Ditto.
466 (process_use): Ditto.
467 (vect_mark_stmts_to_be_vectorized): Ditto.
468 (vect_model_simple_cost): Ditto.
469 (vect_model_promotion_demotion_cost): Ditto.
470 (vect_model_store_cost): Ditto.
471 (vect_get_store_cost): Ditto.
472 (vect_model_load_cost): Ditto.
473 (vect_get_load_cost): Ditto.
474 (vect_init_vector_1): Ditto.
475 (vect_get_vec_def_for_operand): Ditto.
476 (vect_finish_stmt_generation): Ditto.
477 (vectorizable_call): Ditto.
478 (vectorizable_conversion): Ditto.
479 (vectorizable_assignment): Ditto.
480 (vectorizable_shift): Ditto.
481 (vectorizable_operation): Ditto.
482 (vectorizable_store): Ditto.
483 (vectorizable_load): Ditto.
484 (vectorizable_condition): Ditto.
485 (vect_analyze_stmt): Ditto.
486 (vect_transform_stmt): Ditto.
487 (vect_is_simple_use): Ditto.
488 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
489 (vect_can_advance_ivs_p): Ditto.
490 (vect_update_ivs_after_vectorizer): Ditto.
491 (vect_do_peeling_for_loop_bound): Ditto.
492 (vect_gen_niters_for_prolog_loop): Ditto.
493 (vect_update_inits_of_drs): Ditto.
494 (vect_create_cond_for_alias_checks): Ditto.
495 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
496 (vect_build_slp_tree_1): Ditto.
497 (vect_supported_load_permutation_p): Ditto.
498 (vect_analyze_slp_instance): Ditto.
499 (vect_analyze_slp): Ditto.
500 (vect_make_slp_decision): Ditto.
501 (vect_detect_hybrid_slp): Ditto.
502 (vect_bb_vectorization_profitable_p): Ditto.
503 (vect_slp_analyze_bb_1): Ditto.
504 (vect_update_slp_costs_according_to_vf): Ditto.
505 (vect_get_mask_element): Ditto.
506 (vect_transform_slp_perm_load): Ditto.
507 (vect_schedule_slp_instance): Ditto.
508 (vect_schedule_slp): Ditto.
509 (vect_slp_transform_bb): Ditto.
510 * profile.c (read_profile_edge_counts): Ditto.
511 (compute_branch_probabilities): Ditto.
512 * coverage.c (get_coverage_counts): Ditto.
514 2013-09-16 Diego Novillo <dnovillo@google.com>
516 * tree-core.h: Add missing comment lines from refactoring
519 2013-09-16 Jan Hubicka <jh@suse.cz>
521 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
522 abstract functions; for static functions check the presence
525 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
527 * config/aarch64/aarch64-simd-builtins.def (fma): New.
528 * config/aarch64/aarch64-simd.md
529 (aarch64_mla_elt<mode>): New.
530 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
531 (aarch64_mls_elt<mode>): Likewise.
532 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
533 (aarch64_fma4_elt<mode>): Likewise.
534 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
535 (aarch64_fma4_elt_to_128v2df): Likewise.
536 (aarch64_fma4_elt_to_64df): Likewise.
537 (fnma<mode>4): Likewise.
538 (aarch64_fnma4_elt<mode>): Likewise.
539 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
540 (aarch64_fnma4_elt_to_128v2df): Likewise.
541 (aarch64_fnma4_elt_to_64df): Likewise.
542 * config/aarch64/iterators.md (VDQSF): New.
543 * config/aarch64/arm_neon.h
544 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
545 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
547 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
549 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
550 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
551 (aarch64_mul3_elt_to_128df): Likewise.
552 (aarch64_mul3_elt_to_64v2df): Likewise.
553 * config/aarch64/iterators.md (VEL): Also handle DFmode.
555 (VMUL_CHANGE_NLANES) Likewise.
558 * config/aarch64/arm_neon.h
559 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
561 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
563 * config/aarch64/arm_neon.h
564 (vcvtx_high_f32_f64): Fix parameters.
566 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
567 Uros Bizjak <ubizjak@gmail.com>
569 * config/alpha.c: Include tree-ssanames.h.
571 2013-09-16 Richard Biener <rguenther@suse.de>
573 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
574 (dot_rdg_1): Handle control_dd.
575 (create_edge_for_control_dependence): New function.
576 (create_rdg_edges): Add control dependences if asked for.
577 (build_rdg): Likewise.
578 (generate_loops_for_partition): If there are not necessary
579 control stmts remove all their dependencies.
580 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
581 (distribute_loop): Pass on control dependences.
582 (tree_loop_distribution): Compute control dependences and remove
583 restriction on number of loop nodes.
585 2013-09-16 Jakub Jelinek <jakub@redhat.com>
587 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
590 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
592 * cse.c (try_const_anchors): Punt on CC modes.
594 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
596 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
598 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
601 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
604 2013-09-14 Iain Sandoe <iain@codesourcery.com>
607 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
609 (darwin_objc1_section): Likewise.
610 (darwin_file_end): Emit Image Info section when required.
612 2013-09-14 Jan Hubicka <jh@suse.cz>
614 * tree-into-ssa.c (gate_into_ssa): New.
615 (pass_data_build_ssa): Use it.
616 * cgraph.h (expand_thunk): Update prototype.
617 * cgraphunit.c (analyze_function): Expand thunks early.
618 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
619 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
620 set lowered flag; do not add new function.
621 (assemble_thunks_and_aliases): Update.
622 * tree-ssa.c (gate_init_datastructures): New gate.
623 (pass_data_init_datastructures): Use it.
625 2013-09-14 Iain Sandoe <iain@codesourcery.com>
628 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
629 xmm register set described in the psABI.
631 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
633 * dwarf2out.c (should_emit_struct_debug): Add check
634 for type_decl variable is not NULL.
636 2013-09-13 Jacek Caban <jacek@codeweavers.com>
638 * config.gcc: Use new winnt-c.c target hooks
639 * config/t-winnt: New file
640 * config/winnt-c.c: New file
641 * doc/tm.texi.in: Document new hook
642 * doc/tm.texi: Regenerated
644 2013-09-13 Jan Hubicka <jh@suse.cz>
647 * ipa-inline.c (check_callers): New function.
648 (check_caller_edge): Remove.
649 (want_inline_function_to_all_callers_p): Also permit alises that are
651 (inline_to_all_callers): Terminate the walk when devirtualization
654 2013-09-13 Jan Hubicka <jh@suse.cz>
656 * ipa-inline-analysis.c (struct growth_data): Add node.
657 (do_estimate_growth_1): Fix detection of recursion.
659 2013-09-13 Jakub Jelinek <jakub@redhat.com>
661 PR tree-optimization/58392
662 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
663 to avoid shadowing of outer loop variable. If
664 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
665 replace_by_duplicate_decl simduid of loops that have it set and
666 set dest_cfun->has_simduid_loops and/or
667 dest_cfun->has_force_vect_loops.
668 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
669 instead of maybe_lookup_decl.
670 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
671 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
672 call. Copy over force_vect and copy and remap simduid. Set
673 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
674 (copy_cfg_body): Remove blocks_to_copy argument. Use
675 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
676 caller. Don't set cfun->has_simduid_loops and/or
677 cfun->has_force_vect_loops here.
678 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
680 (expand_call_inline, tree_function_versioning): Adjust copy_body
683 2013-09-13 Martin Jambor <mjambor@suse.cz>
686 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
687 the assert if the edge was a speculative one.
689 2013-09-13 Richard Biener <rguenther@suse.de>
691 * tree-data-ref.h (known_dependences_p): Move here ...
692 * tree-loop-distribution.c (known_dependences_p): ... from here.
693 (dump_rdg_component, debug_rdg_component): Remove.
695 (generate_loops_for_partition): Use gimple_uid instead of
696 relying on matching stmt visit order.
697 (rdg_build_partitions): Take starting stmt vector.
698 (ldist_gen): Merge into ...
699 (distribute_loop): ... this function. Do not compute starting
701 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
703 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
705 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
706 Set type attribute correctly. Set predicable_short_it attribute.
707 (cmpsi_shiftsi): Remove %? from output template.
709 2013-09-13 Richard Biener <rguenther@suse.de>
711 * tree-loop-distribution.c (struct rdg_component,
712 rdg_defs_used_in_other_loops_p, free_rdg_components,
713 rdg_build_components): Remove.
714 (stmts_from_loop): Do not record virtual PHIs.
715 (generate_loops_for_partition): Skip virtual PHIs.
716 (build_rdg_partition_for_component): Rename to ...
717 (build_rdg_partition_for_vertex): ... this and adjust.
718 (rdg_build_partitions): Take a vector of starting vertices
719 instead of components. Remove unnecessary leftover handling.
720 (ldist_gen): Do not build components or record other stores.
721 (distribute_loop): Do not distribute loops containing stmts
724 2013-09-13 Christian Bruel <christian.bruel@st.com>
727 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
729 2013-09-13 Kai Tietz <ktietz@redhat.com>
731 * config.gcc: Separate cases for mingw and cygwin targets,
732 and add 64-bit cygwin target case.
734 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
735 dll-export inline-functions.
736 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
739 2013-09-13 Jeff Law <law@redhat.com>
743 2013-09-06 Jeff Law <law@redhat.com>
745 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
746 edge implied equivalences into successor phis.
748 2013-09-12 DJ Delorie <dj@redhat.com>
750 * config/rl78/rl78-virt.md: Change from | to \; for asm line
753 2013-09-12 Brooks Moses <bmoses@google.com>
756 * Makefile.in: Do not install driver binaries in $(target)/bin.
758 2013-09-12 DJ Delorie <dj@redhat.com>
760 * config/rl78/rl78.opt (mrelax): New.
761 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
762 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
764 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
766 (rl78_expand_epilogue): Likewise.
768 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
771 * lra-eliminations.c (remove_reg_equal_offset_note): New.
772 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
773 pointer elimination with using remove_reg_equal_offset_note.
775 2013-09-12 DJ Delorie <dj@redhat.com>
777 * config/msp430/: New port.
778 * config.gcc (msp430): Added.
779 * doc/invoke.texi: Document MSP430 options.
780 * doc/install.texi: Document msp430-elf
781 * doc/md.texi: Document msp430-elf
782 * doc/contrib.texi: Document msp430-elf
784 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
787 2013-09-12 Martin Jambor <mjambor@suse.cz>
790 * ipa-prop.c (remove_described_reference): Give up if the edge in the
791 reference descriptor is NULL.
792 (ipa_edge_removal_hook): If owning a reference descriptor, set its
795 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
797 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
798 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
799 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
800 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
801 make_temp_ssa_name): move to tree-ssanames.h
802 * tree-ssa-alias.h: Move prototype.
803 * tree-ssa.h: Include tree-ssanames.h.
804 * tree-ssanames.c (FREE_SSANAMES): Move to here.
805 * tree-ssanames.h: New. Move items from tree-flow*.h
806 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
808 2013-09-12 Richard Biener <rguenther@suse.de>
810 PR tree-optimization/58404
811 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
812 propagate non-invariant addresses into dereferences wrapped
813 in component references.
815 2013-09-12 Richard Biener <rguenther@suse.de>
817 PR tree-optimization/58402
818 * passes.def: Move pass_late_warn_uninitialized later.
820 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
822 * tree-ssa.h: New. Move content from tree-flow.h and
824 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
825 Move prototypes belonging to tree-ssa.c.
826 * tree-flow-inline.h (redirect_edge_var_map_def,
827 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
829 * gimple.h: Adjust prototypes.
830 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
832 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
833 * tree.h: Move prototype to tree-ssa.h.
834 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
835 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
836 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
837 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
838 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
839 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
840 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
841 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
842 graphite-blocking.c, graphite-clast-to-gimple.c,
843 graphite-dependences.c, graphite-interchange.c,
844 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
845 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
846 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
847 ipa-reference.c, ipa-split.c, ipa-utils.c,
848 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
849 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
850 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
851 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
852 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
853 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
854 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
855 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
856 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
857 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
858 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
859 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
860 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
861 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
862 value-prof.c, var-tracking.c,
863 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
865 2013-09-12 Richard Biener <rguenther@suse.de>
867 PR tree-optimization/58396
868 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
869 (build_rdg): Take a loop-nest parameter, fix memleaks.
870 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
872 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
874 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
877 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
879 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
880 The FMA instruction names should have a 'v' prefix.
882 2013-09-12 Richard Biener <rguenther@suse.de>
884 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
885 (dot_rdg): Use popen instead of system in optional code.
886 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
887 (already_processed_vertex_p): Adjust.
888 (has_anti_or_output_dependence, predecessor_has_mem_write,
889 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
890 rdg_flag_uses): Remove.
891 (rdg_flag_vertex): Simplify.
892 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
894 (build_rdg_partition_for_component): Process the first vertex
896 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
898 2013-09-12 Alan Modra <amodra@gmail.com>
900 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
902 2013-09-11 DJ Delorie <dj@redhat.com>
903 Nick Clifton <nickc@redhat.com>
905 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
906 (rl78_stack_based_mem): New.
907 * config/rl78/constraints.md (Iv08): New.
916 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
918 (movsi): Change from expand to insn-and-split.
919 (ashrsi3): Clobber AX.
923 * config/rl78/rl78.md (CC_REG): Fix.
924 (addsi3): Allow memory and immediate operands.
925 (addsi3_internal): Split into...
926 (addsi3_internal_virt): ...new, and ...
927 (addsi3_internal_real): ...new.
929 (subsi3_internal_virt): New.
930 (subsi3_internal_real): New.
931 (mulsi3): Add memory operand.
932 (mulsi3_rl78): Likewise.
933 (mulsi3_g13): Likewise.
934 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
935 (cbranchqi4_real): Add more constraint options.
936 (cbranchhi4_real): Expand pattern.
937 (cbranchhi4_real_signed): New.
938 (cbranchhi4_real_inverted): New.
939 (cbranchsi4_real_lt): New.
940 (cbranchsi4_real_ge): New.
941 (cbranchsi4_real_signed): New.
942 (cbranchsi4_real): New.
944 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
946 (lshrsi3_virt): Likewise.
947 (ashlsi3_virt): Likewise.
948 (cbranchqi4_virt_signed): New.
949 (cbranchhi4_virt_signed): New.
950 (cbranchsi4_virt): New.
951 * config/rl78/rl78.c: Whitespace fixes throughout.
952 (move_elim_pass): New.
953 (pass_data_rl78_move_elim): New.
954 (pass_rl78_move_elim): New.
955 (make_pass_rl78_move_elim): New.
956 (rl78_devirt_info): Run devirt earlier.
957 (rl78_move_elim_info): New.
958 (rl78_asm_file_start): Register it.
959 (rl78_split_movsi): New.
960 (rl78_as_legitimate_address): Allow virtual base registers when
962 (rl78_addr_space_convert): Remove spurious debug stuff.
963 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
964 (rl78_print_operand): More cases for not printing '#'.
965 (rl78_expand_compare): Remove most of the logic.
966 (content_memory): New.
967 (clear_content_memory): New.
968 (get_content_index): New.
969 (get_content_name): New.
970 (display_content_memory): New.
971 (update_content): New.
972 (record_content): New.
973 (already_contains): New.
974 (insn_ok_now): Re-recog insns with virtual registers.
975 (add_postponed_content_update): New.
976 (process_postponed_content_update): New.
977 (gen_and_emit_move): New.
978 (transcode_memory_rtx): Record new location content.
979 Use gen_and_emit_move.
980 (force_into_acc): New.
981 (move_to_acc): Use gen_and_emit_move.
982 (move_from_acc): Likewise.
983 (move_acc_to_reg): Likewise.
984 (move_to_x): Likewise.
985 (move_to_hl): Likewise.
986 (move_to_de): Likewise.
987 (rl78_alloc_physical_registers_op1): Record location content.
988 (has_constraint): New.
989 (rl78_alloc_physical_registers_op2): Record location content.
991 (rl78_alloc_physical_registers_ro1): Likewise.
992 (rl78_alloc_physical_registers_cmp): Likewise.
993 (rl78_alloc_physical_registers_umul): Likewise.
994 (rl78_alloc_address_registers_macax): New.
995 (rl78_alloc_physical_registers): Initialize and set location
996 content memory as needed.
997 (rl78_reorg): Make sure split2 is called.
998 (rl78_rtx_costs): New.
1000 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1002 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1003 for (not (neg ...)) and (neg (not ...)) cases.
1005 2013-09-11 Richard Biener <rguenther@suse.de>
1008 * passes.def: Split critical edges before late uninit warning passes.
1009 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1011 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1013 PR tree-optimization/58385
1014 * fold-const.c (build_range_check): If both low and high are NULL,
1015 use omit_one_operand_loc to preserve exp side-effects.
1017 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1019 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1021 2013-09-11 Richard Biener <rguenther@suse.de>
1023 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1024 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1025 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1026 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1027 stmts_from_loop, known_dependences_p, build_empty_rdg,
1028 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1029 * tree-loop-distribution.c: ... here.
1030 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1031 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1032 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1033 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1034 * tree-loop-distribution.c: ... here.
1035 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1036 (struct partition_s): Add loops member.
1037 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1038 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1039 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1041 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1042 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1043 Sergey Lega <sergey.s.lega@intel.com>
1044 Anna Tikhonova <anna.tikhonova@intel.com>
1045 Ilya Tocar <ilya.tocar@intel.com>
1046 Andrey Turetskiy <andrey.turetskiy@intel.com>
1047 Ilya Verbin <ilya.verbin@intel.com>
1048 Kirill Yukhin <kirill.yukhin@intel.com>
1049 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1051 * config/i386/constraints.md (k): New.
1053 * config/i386/i386.c (const regclass_map): Add new mask registers.
1054 (dbx_register_map): Ditto.
1055 (dbx64_register_map): Ditto.
1056 (svr4_dbx_register_map): Ditto.
1057 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
1059 (ix86_preferred_reload_class): Disable constants for mask registers.
1060 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
1061 (ix86_hard_regno_mode_ok): Support new mask registers.
1062 (x86_order_regs_for_local_alloc): Ditto.
1063 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
1064 (FIXED_REGISTERS): Add new mask registers.
1065 (CALL_USED_REGISTERS): Ditto.
1066 (REG_ALLOC_ORDER): Ditto.
1067 (VALID_MASK_REG_MODE): New.
1068 (FIRST_MASK_REG): Ditto.
1069 (LAST_MASK_REG): Ditto.
1070 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
1071 (MAYBE_MASK_CLASS_P): New.
1072 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
1073 (REG_CLASS_CONTENTS): Ditto.
1074 (MASK_REGNO_P): New.
1075 (ANY_MASK_REG_P): Ditto.
1076 (HI_REGISTER_NAMES): Add new mask registers.
1077 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
1078 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
1080 (attribute "type"): Add mskmov, msklog.
1081 (attribute "length_immediate"): Support them.
1082 (attribute "memory"): Ditto.
1083 (attribute "prefix_0f"): Ditto.
1084 (*movhi_internal): Support new mask registers.
1085 (*movqi_internal): Ditto.
1086 (define_split): Split out clobber pattern is a logic
1087 insn on mask registers.
1088 (*k<logic><mode>): New.
1089 (*andhi_1): Extend to support mask regs.
1090 (*andqi_1): Extend to support mask regs.
1092 (define_split): Split and-not to and and not if operands
1094 (*<code><mode>_1): Separate HI mode to new pattern...
1095 (*<code>hi_1): This.
1096 (*<code>qi_1): Extend to support mask regs.
1098 (kortestzhi): Ditto.
1099 (kortestchi): Ditto.
1101 (*one_cmpl<mode>2_1): Remove HImode and handle it...
1102 (*one_cmplhi2_1): ...Here, now with mask registers support.
1103 (*one_cmplqi2_1): Support new mask registers.
1104 (HI/QImode arithmetics splitter): Don't split if mask registers
1106 (HI/QImode not splitter): Ditto.
1107 * config/i386/predicated.md (mask_reg_operand): New.
1108 (general_reg_operand): Ditto.
1110 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1112 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
1113 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
1115 2013-09-10 Jeff Law <law@redhat.com>
1117 PR tree-optimization/58380
1118 * tree-ssa-threadupdate.c (thread_block): Recognize another case
1119 of threading through a buried loop header.
1121 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
1122 return value for single successor case.
1124 2013-09-10 Jan Hubicka <jh@suse.cz>
1126 * ipa-devirt.c (ipa_devirt): Enable with LTO.
1128 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
1131 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
1132 support conditional execution.
1133 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
1135 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
1137 * lra.c (lra): Clear lra_optional_reload_pseudos before every
1139 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
1140 Check destination too to check move insn.
1141 (undo_optional_reloads): Add check that the original peudo did not
1142 changed its allocation and the optional reload was inherited on last
1143 inheritance pass. Break loop after deciding to keep optional reload.
1144 (lra_undo_inheritance): Add check that inherited pseudo still in
1147 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
1149 * config/aarch64/aarch64.md (generic_sched): New.
1150 * config/aarch64/aarch64-generic.md (load): Make conditional
1151 on generic_sched attribute.
1152 (nonload): Likewise.
1154 2013-09-10 Jan Hubicka <jh@suse.cz>
1156 * lto-cgraph.c: Include ipa-utils.h.
1157 (compute_ltrans_boundary): Also add possible targets into the boundary.
1159 2013-09-10 Jan Hubicka <jh@suse.cz>
1161 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
1162 VAR_DECL of vtable rather than full expression.
1164 2013-09-10 Jan Hubicka <jh@suse.cz>
1165 Paolo Carlini <paolo.carlini@oracle.com>
1167 * cgraphunit.c (analyze_functions): Save input_location, set it
1168 to UNKNOWN_LOCATION and restore it at the end.
1170 2013-09-10 Martin Jambor <mjambor@suse.cz>
1172 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
1173 represented by a thunk.
1175 2013-09-10 Jeff Law <law@redhat.com>
1177 PR tree-optimization/58343
1178 * tree-ssa-threadupdate.c (thread_block): Identify and disable
1179 jump threading requests through loop headers buried in the middle
1180 of a jump threading path.
1182 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
1183 in return value/type.
1185 2013-09-10 Jakub Jelinek <jakub@redhat.com>
1187 PR rtl-optimization/58365
1188 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
1189 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
1192 2013-09-10 Richard Biener <rguenther@suse.de>
1194 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
1195 * tree-data-ref.c (create_rdg_vertices): Collect all data
1196 references, signal failure to the caller, use data-ref API.
1197 (build_rdg): Compute data references only once. Maintain lifetime
1198 of data references and data dependences from within RDG.
1199 (free_rdg): Free dependence relations.
1200 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
1201 inventing extra dependences.
1202 (distribute_loop): Update for RDG API changes.
1204 2013-09-10 Kai Tietz <ktietz@redhat.com>
1206 * doc/invoke.texi (fms-extensions): Document changed
1207 behavior for ms-abi targets.
1208 * config/i386/i386.c (ix86_option_override_internal):
1209 Set default value of option -fms-extension for ms-abi targets.
1211 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
1213 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
1215 2013-09-10 Alan Modra <amodra@gmail.com>
1218 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
1220 2013-09-10 Alan Modra <amodra@gmail.com>
1222 * config/rs6000/predicates.md (add_cint_operand): New.
1223 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
1224 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
1225 using add_cint_operand.
1226 (largetoc_high_plus_aix): Likewise.
1228 2013-09-09 Jakub Jelinek <jakub@redhat.com>
1230 PR tree-optimization/58364
1231 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
1232 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
1233 the current range can't be an unconditional true or false.
1235 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
1237 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
1239 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
1241 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
1243 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
1246 * doc/invoke.texi (-Wdelete-incomplete): Document it.
1248 2013-09-09 Ian Bolton <ian.bolton@arm.com>
1250 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
1251 NO_REGS for immediate that can't be moved directly into FP_REGS.
1253 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1255 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
1256 comparison with negated operand.
1257 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
1260 2013-09-09 Richard Biener <rguenther@suse.de>
1263 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
1264 of a subloop record all its block as affecting loop-closed SSA form.
1266 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1268 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
1269 of an rtx/bitpos pair.
1270 (store_fixed_bit_field): Update accordingly.
1272 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1274 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
1276 * builtins.c (expand_errno_check): Likewise.
1277 * dwarf2cfi.c (init_return_column_size): Likewise.
1278 * except.c (sjlj_mark_call_sites): Likewise.
1279 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
1280 * lra-constraints.c (emit_inc): Likewise.
1281 * ree.c (combine_set_extension): Likewise.
1282 * regmove.c (fixup_match_2): Likewise.
1283 * reload1.c (inc_for_reload): Likewise.
1285 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1287 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
1288 (force_to_mode, simplify_shift_const_1, simplify_comparison):
1289 Use gen_int_mode with the mode of the associated simplify_* call.
1290 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
1291 * expmed.c (expand_shift_1): Likewise.
1292 * function.c (instantiate_virtual_regs_in_insn): Likewise.
1293 * loop-iv.c (iv_number_of_iterations): Likewise.
1294 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
1295 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
1297 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1299 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
1300 of the associated expand_* call.
1301 (asan_emit_stack_protection): Likewise.
1302 * builtins.c (round_trampoline_addr): Likewise.
1303 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
1304 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
1305 (emit_store_flag): Likewise.
1306 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
1307 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
1309 * function.c (instantiate_virtual_regs_in_insn): Likewise.
1310 * ifcvt.c (noce_try_store_flag_constants): Likewise.
1311 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
1312 * modulo-sched.c (generate_prolog_epilog): Likewise.
1313 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
1314 (expand_ctz, expand_ffs, expand_unop): Likewise.
1316 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1318 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
1319 of the associated gen_rtx_* call.
1320 * caller-save.c (init_caller_save): Likewise.
1321 * combine.c (find_split_point, make_extraction): Likewise.
1322 (make_compound_operation): Likewise.
1323 * dwarf2out.c (mem_loc_descriptor): Likewise.
1324 * explow.c (plus_constant, probe_stack_range): Likewise.
1325 * expmed.c (expand_mult_const): Likewise.
1326 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
1327 * reload1.c (init_reload): Likewise.
1328 * valtrack.c (cleanup_auto_inc_dec): Likewise.
1329 * var-tracking.c (adjust_mems): Likewise.
1330 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
1331 rather than gen_rtx_fmt_ee.
1333 2013-09-09 Jan Hubicka <jh@suse.cz>
1336 * value-prof.c (gimple_ic): Copy also abnormal edges.
1338 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1340 * asan.c (asan_shadow_cst): Use gen_int_mode.
1342 2013-09-08 Jan Hubicka <jh@suse.cz>
1344 * ipa-profile.c: Add toplevel comment.
1345 (ipa_propagate_frequency_1): Be more conservative when profile is read.
1346 (contains_hot_call_p): New function.
1347 (ipa_propagate_frequency): Set frequencies based on counts when
1349 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
1350 profile; do not tamper with profile after inlining if it is read.
1352 2013-09-08 Jan Hubicka <jh@suse.cz>
1354 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
1357 2013-09-08 Jan Hubicka <jh@suse.cz>
1359 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
1362 2013-09-08 Jeff Law <law@redhat.com>
1365 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
1368 2013-09-08 Andi Kleen <ak@linux.intel.com>
1370 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
1372 2013-09-08 Jan Hubicka <jh@suse.cz>
1374 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
1376 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
1378 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
1379 for non-debug insns.
1380 * lra.c (new_insn_reg): Take the containing insn as a parameter.
1381 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
1382 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
1385 2013-09-08 Jan Hubicka <jh@suse.cz>
1387 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
1388 targets and devirtualize to BUILT_IN_UNREACHABLE.
1389 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
1390 * ipa.c (walk_polymorphic_call_targets): New function.
1391 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
1392 functions; use the new timevar.
1393 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
1394 was removed from the program.
1395 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
1396 not consider it in the walk when its vtable is dead.
1397 (possible_polymorphic_call_targets_1): Pass anonymous flag to
1399 (devirt_variable_node_removal_hook): New function.
1400 (possible_polymorphic_call_targets): Also register
1401 devirt_variable_node_removal_hook.
1402 (ipa_devirt): Do not do non-speculative devirtualization.
1403 (gate_ipa_devirt): One execute if devirtualizing speculatively.
1405 2013-09-08 Jan Hubicka <jh@suse.cz>
1407 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
1408 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
1409 varpool_remove_variable_insertion_hook): Declare.
1410 * varpool.c (varpool_node_hook_list): New structure.
1411 (first_varpool_node_removal_hook,
1412 first_varpool_variable_insertion_hook): New variables.
1413 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
1414 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
1415 varpool_remove_variable_insertion_hook,
1416 varpool_call_variable_insertion_hooks): New functions.
1417 (varpool_remove_node): Use it.
1419 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
1422 * diagnostic.c (diagnostic_build_prefix): When s.file is
1423 "<built-in>" don't output line and column numbers.
1425 2013-09-06 Jan Hubicka <jh@suse.cz>
1427 * cgraphunit.c (expand_thunk): Get body before touching arguments.
1428 * lto-streamer-out.c: Stream thunks, too.
1429 * lto-streamer-in.c (input_function): Pop cfun here
1430 (lto_read_body): Instead of here.
1432 2013-09-06 Caroline Tice <cmtice@google.com>
1434 * doc/install.texi: Add documentation for the --enable-vtable-verify
1435 and the --disable-libvtv configure options.
1437 2013-09-06 Jeff Law <law@redhat.com>
1439 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1440 edge implied equivalences into successor phis.
1442 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
1444 * resource.c (mark_referenced_resources): Handle COND_EXEC.
1446 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
1448 * resource.c (mark_target_live_regs): Compute resources taking
1449 into account if a call is predicated or not.
1451 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1453 * toplev.c (output_stack_usage): Be prepared for suffixes created by
1454 the compiler in the function names.
1456 2013-09-06 Jan Hubicka <jh@suse.cz>
1459 * ipa-inline.c (has_caller_p): New function.
1460 (want_inline_function_to_all_callers_p): Use it.
1461 (sum_callers, inline_to_all_callers): Break out from ...
1462 (ipa_inline): ... here.
1464 2013-09-06 Jan Hubicka <jh@suse.cz>
1466 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
1467 only when AVX is enabled.
1469 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1471 * config/aarch64/aarch64.md
1472 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
1473 is fpsimd_<load/store>2.
1474 (load_pair<mode>): Likewise.
1475 (store_pair<mode>): Likewise.
1477 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1479 * config/arm/types.md (type): Add "mrs" type.
1480 * config/aarch64/aarch64.md
1481 (aarch64_load_tp_hard): Make type "mrs".
1483 (load_tp_hard): Make type "mrs".
1484 * config/arm/cortex-a15.md: Update with new attributes.
1485 * config/arm/cortex-a5.md: Update with new attributes.
1486 * config/arm/cortex-a53.md: Update with new attributes.
1487 * config/arm/cortex-a7.md: Update with new attributes.
1488 * config/arm/cortex-a8.md: Update with new attributes.
1489 * config/arm/cortex-a9.md: Update with new attributes.
1490 * config/arm/cortex-m4.md: Update with new attributes.
1491 * config/arm/cortex-r4.md: Update with new attributes.
1492 * config/arm/fa526.md: Update with new attributes.
1493 * config/arm/fa606te.md: Update with new attributes.
1494 * config/arm/fa626te.md: Update with new attributes.
1495 * config/arm/fa726te.md: Update with new attributes.
1497 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1499 * config/aarch64/aarch64.md
1500 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
1501 (*movtf_aarch64): Likewise.
1503 (thumb1_movdi_insn): Use "multiple" for type where more than one
1504 instruction is used for a move.
1505 (*arm32_movhf): Likewise.
1506 (*thumb_movdf_insn): Likewise.
1508 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1510 * config/arm/types.md (type): Rename fcpys to fmov.
1512 (*arm_movsi_vfp): Rename type fcpys as fmov.
1513 (*thumb2_movsi_vfp): Likewise
1514 (*movhf_vfp_neon): Likewise
1515 (*movhf_vfp): Likewise
1516 (*movsf_vfp): Likewise
1517 (*thumb2_movsf_vfp): Likewise
1518 (*movsfcc_vfp): Likewise
1519 (*thumb2_movsfcc_vfp): Likewise
1520 * config/aarch64/aarch64-simd.md
1521 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
1522 * config/aarch64/aarch64.md
1523 (*movsi_aarch64): Replace type mov_reg with fmovs.
1524 (*movdi_aarch64): Likewise
1525 (*movsf_aarch64): Likewise
1526 (*movdf_aarch64): Likewise
1528 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
1529 * config/arm/iwmmxt.md
1530 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
1531 * config/arm/arm1020e.md: Update with new attributes.
1532 * config/arm/cortex-a15-neon.md: Update with new attributes.
1533 * config/arm/cortex-a5.md: Update with new attributes.
1534 * config/arm/cortex-a53.md: Update with new attributes.
1535 * config/arm/cortex-a7.md: Update with new attributes.
1536 * config/arm/cortex-a8-neon.md: Update with new attributes.
1537 * config/arm/cortex-a9.md: Update with new attributes.
1538 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1539 * config/arm/cortex-r4f.md: Update with new attributes.
1540 * config/arm/marvell-pj4.md: Update with new attributes.
1541 * config/arm/vfp11.md: Update with new attributes.
1543 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1545 * config/aarch64/aarch64.md
1546 (*madd<mode>): Fix type attribute.
1547 (*maddsi_uxtw): Likewise.
1548 (*msub<mode>): Likewise.
1549 (*msubsi_uxtw): Likewise.
1550 (<su_optab>maddsidi4): Likewise.
1551 (<su_optab>msubsidi4): Likewise.
1553 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1555 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
1556 * config/arm/arm.md (core_cycles): Remove fdiv.
1557 * config/arm/vfp.md:
1558 (*sqrtsf2_vfp): Update for attribute changes.
1559 (*sqrtdf2_vfp): Likewise.
1560 * config/aarch64/aarch64.md:
1561 (sqrt<mode>2): Update for attribute changes.
1562 * config/arm/arm1020e.md: Update with new attributes.
1563 * config/arm/cortex-a15-neon.md: Update with new attributes.
1564 * config/arm/cortex-a5.md: Update with new attributes.
1565 * config/arm/cortex-a53.md: Update with new attributes.
1566 * config/arm/cortex-a7.md: Update with new attributes.
1567 * config/arm/cortex-a8-neon.md: Update with new attributes.
1568 * config/arm/cortex-a9.md: Update with new attributes.
1569 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1570 * config/arm/cortex-r4f.md: Update with new attributes.
1571 * config/arm/marvell-pj4.md: Update with new attributes.
1572 * config/arm/vfp11.md: Update with new attributes.
1574 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1576 * config/arm/types.md
1577 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
1578 * config/aarch64/aarch64.md
1579 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
1581 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
1582 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
1583 (float<GPI:mode><GPF:mode>2): Likewise.
1585 (*truncsisf2_vfp): Update with new attributes.
1586 (*truncsidf2_vfp): Likewise.
1587 (fixuns_truncsfsi2): Likewise.
1588 (fixuns_truncdfsi2): Likewise.
1589 (*floatsisf2_vfp): Likewise.
1590 (*floatsidf2_vfp): Likewise.
1591 (floatunssisf2): Likewise.
1592 (floatunssidf2): Likewise.
1593 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
1594 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
1595 * config/arm/arm1020e.md: Update with new attributes.
1596 * config/arm/cortex-a15-neon.md: Update with new attributes.
1597 * config/arm/cortex-a5.md: Update with new attributes.
1598 * config/arm/cortex-a53.md: Update with new attributes.
1599 * config/arm/cortex-a7.md: Update with new attributes.
1600 * config/arm/cortex-a8-neon.md: Update with new attributes.
1601 * config/arm/cortex-a9.md: Update with new attributes.
1602 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1603 * config/arm/cortex-r4f.md: Update with new attributes.
1604 * config/arm/marvell-pj4.md: Update with new attributes.
1605 * config/arm/vfp11.md: Update with new attributes.
1607 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1609 * config/aarch64/arm_neon.h
1610 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
1611 (vqtbx<1,2,3,4><q>_s8): Likewise.
1613 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1615 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
1616 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
1617 (add<mode>3): Add type attribute.
1618 (add<mode>3): Likewise.
1619 (usadd<mode>3): Likewise.
1620 (ssadd<mode>3): Likewise.
1621 (sub<mode>3): Likewise.
1622 (sub<mode>3): Likewise.
1623 (ussub<mode>3): Likewise.
1624 (sssub<mode>3): Likewise.
1625 (ssmulsa3): Likewise.
1626 (usmulusa3): Likewise.
1627 (arm_usatsihi): Likewise.
1629 (*movdi_vfp): Add types for all instructions.
1630 (*movdi_vfp_cortexa8): Likewise.
1631 (*movhf_vfp_neon): Likewise.
1632 (*movhf_vfp): Likewise.
1633 (*movdf_vfp): Likewise.
1634 (*thumb2_movdf_vfp): Likewise.
1635 (*thumb2_movdfcc_vfp): Likewise.
1636 * config/arm/arm.md: Add type attribute to all insn patterns.
1637 (*thumb1_adddi3): Add type attribute.
1638 (*arm_adddi3): Likewise.
1639 (*adddi_sesidi_di): Likewise.
1640 (*adddi_zesidi_di): Likewise.
1641 (*thumb1_addsi3): Likewise.
1642 (addsi3_compare0): Likewise.
1643 (*addsi3_compare0_scratch): Likewise.
1644 (*compare_negsi_si): Likewise.
1645 (cmpsi2_addneg): Likewise.
1646 (*addsi3_carryin_<optab>): Likewise.
1647 (*addsi3_carryin_alt2_<optab>): Likewise.
1648 (*addsi3_carryin_clobercc_<optab>): Likewise.
1649 (*subsi3_carryin): Likewise.
1650 (*subsi3_carryin_const): Likewise.
1651 (*subsi3_carryin_compare): Likewise.
1652 (*subsi3_carryin_compare_const): Likewise.
1653 (*arm_subdi3): Likewise.
1654 (*thumb_subdi3): Likewise.
1655 (*subdi_di_zesidi): Likewise.
1656 (*subdi_di_sesidi): Likewise.
1657 (*subdi_zesidi_di): Likewise.
1658 (*subdi_sesidi_di): Likewise.
1659 (*subdi_zesidi_ze): Likewise.
1660 (thumb1_subsi3_insn): Likewise.
1661 (*arm_subsi3_insn): Likewise.
1662 (*anddi3_insn): Likewise.
1663 (*anddi_zesidi_di): Likewise.
1664 (*anddi_sesdi_di): Likewise.
1665 (*ne_zeroextracts): Likewise.
1666 (*ne_zeroextracts): Likewise.
1667 (*ite_ne_zeroextr): Likewise.
1668 (*ite_ne_zeroextr): Likewise.
1669 (*anddi_notdi_di): Likewise.
1670 (*anddi_notzesidi): Likewise.
1671 (*anddi_notsesidi): Likewise.
1672 (andsi_notsi_si): Likewise.
1673 (thumb1_bicsi3): Likewise.
1674 (*iordi3_insn): Likewise.
1675 (*iordi_zesidi_di): Likewise.
1676 (*iordi_sesidi_di): Likewise.
1677 (*thumb1_iorsi3_insn): Likewise.
1678 (*xordi3_insn): Likewise.
1679 (*xordi_zesidi_di): Likewise.
1680 (*xordi_sesidi_di): Likewise.
1681 (*arm_xorsi3): Likewise.
1682 (*andsi_iorsi3_no): Likewise.
1683 (*smax_0): Likewise.
1684 (*smax_m1): Likewise.
1685 (*arm_smax_insn): Likewise.
1686 (*smin_0): Likewise.
1687 (*arm_smin_insn): Likewise.
1688 (*arm_umaxsi3): Likewise.
1689 (*arm_uminsi3): Likewise.
1690 (*minmax_arithsi): Likewise.
1691 (*minmax_arithsi_): Likewise.
1692 (*satsi_<SAT:code>): Likewise.
1693 (arm_ashldi3_1bit): Likewise.
1694 (arm_ashrdi3_1bit): Likewise.
1695 (arm_lshrdi3_1bit): Likewise.
1696 (*arm_negdi2): Likewise.
1697 (*thumb1_negdi2): Likewise.
1698 (*arm_negsi2): Likewise.
1699 (*thumb1_negsi2): Likewise.
1700 (*negdi_extendsid): Likewise.
1701 (*negdi_zero_extend): Likewise.
1702 (*arm_abssi2): Likewise.
1703 (*thumb1_abssi2): Likewise.
1704 (*arm_neg_abssi2): Likewise.
1705 (*thumb1_neg_abss): Likewise.
1706 (one_cmpldi2): Likewise.
1707 (extend<mode>di2): Likewise.
1708 (*compareqi_eq0): Likewise.
1709 (*arm_extendhisi2addsi): Likewise.
1710 (*arm_movdi): Likewise.
1711 (*thumb1_movdi_insn): Likewise.
1712 (*arm_movt): Likewise.
1713 (*thumb1_movsi_insn): Likewise.
1714 (pic_add_dot_plus_four): Likewise.
1715 (pic_add_dot_plus_eight): Likewise.
1716 (tls_load_dot_plus_eight): Likewise.
1717 (*thumb1_movhi_insn): Likewise.
1718 (*thumb1_movsf_insn): Likewise.
1719 (*movdf_soft_insn): Likewise.
1720 (*thumb_movdf_insn): Likewise.
1721 (cbranchsi4_insn): Likewise.
1722 (cbranchsi4_scratch): Likewise.
1723 (*negated_cbranchsi4): Likewise.
1724 (*tbit_cbranch): Likewise.
1725 (*tlobits_cbranch): Likewise.
1726 (*tstsi3_cbranch): Likewise.
1727 (*cbranchne_decr1): Likewise.
1728 (*addsi3_cbranch): Likewise.
1729 (*addsi3_cbranch_scratch): Likewise.
1730 (*arm_cmpdi_insn): Likewise.
1731 (*arm_cmpdi_unsig): Likewise.
1732 (*arm_cmpdi_zero): Likewise.
1733 (*thumb_cmpdi_zero): Likewise.
1734 (*deleted_compare): Likewise.
1735 (*mov_scc): Likewise.
1736 (*mov_negscc): Likewise.
1737 (*mov_notscc): Likewise.
1738 (*cstoresi_eq0_thumb1_insn): Likewise.
1739 (cstoresi_nltu_thumb1): Likewise.
1740 (cstoresi_ltu_thu): Likewise.
1741 (thumb1_addsi3_addgeu): Likewise.
1742 (*arm_jump): Likewise.
1743 (*thumb_jump): Likewise.
1744 (*check_arch2): Likewise.
1745 (arm_casesi_internal): Likewise.
1746 (thumb1_casesi_dispatch): Likewise.
1747 (*arm_indirect_jump): Likewise.
1748 (*thumb1_indirect_jump): Likewise.
1750 (*and_scc): Likewise.
1751 (*ior_scc): Likewise.
1752 (*compare_scc): Likewise.
1753 (*cond_move): Likewise.
1754 (*cond_arith): Likewise.
1755 (*cond_sub): Likewise.
1756 (*cmp_ite0): Likewise.
1757 (*cmp_ite1): Likewise.
1758 (*cmp_and): Likewise.
1759 (*cmp_ior): Likewise.
1760 (*ior_scc_scc): Likewise.
1761 (*ior_scc_scc_cmp): Likewise.
1762 (*and_scc_scc): Likewise.
1763 (*and_scc_scc_cmp): Likewise.
1764 (*and_scc_scc_nod): Likewise.
1765 (*negscc): Likewise.
1766 (movcond_addsi): Likewise.
1767 (movcond): Likewise.
1768 (*ifcompare_plus_move): Likewise.
1769 (*if_plus_move): Likewise.
1770 (*ifcompare_move_plus): Likewise.
1771 (*if_move_plus): Likewise.
1772 (*ifcompare_arith_arith): Likewise.
1773 (*if_arith_arith): Likewise.
1774 (*ifcompare_arith_move): Likewise.
1775 (*if_arith_move): Likewise.
1776 (*ifcompare_move_arith): Likewise.
1777 (*if_move_arith): Likewise.
1778 (*ifcompare_move_not): Likewise.
1779 (*if_move_not): Likewise.
1780 (*ifcompare_not_move): Likewise.
1781 (*if_not_move): Likewise.
1782 (*ifcompare_shift_move): Likewise.
1783 (*if_shift_move): Likewise.
1784 (*ifcompare_move_shift): Likewise.
1785 (*if_move_shift): Likewise.
1786 (*ifcompare_shift_shift): Likewise.
1787 (*ifcompare_not_arith): Likewise.
1788 (*ifcompare_arith_not): Likewise.
1789 (*if_arith_not): Likewise.
1790 (*ifcompare_neg_move): Likewise.
1791 (*if_neg_move): Likewise.
1792 (*ifcompare_move_neg): Likewise.
1793 (*if_move_neg): Likewise.
1794 (prologue_thumb1_interwork): Likewise.
1795 (*cond_move_not): Likewise.
1796 (*sign_extract_onebit): Likewise.
1797 (*not_signextract_onebit): Likewise.
1798 (stack_tie): Likewise.
1799 (align_4): Likewise.
1800 (align_8): Likewise.
1801 (consttable_end): Likewise.
1802 (consttable_1): Likewise.
1803 (consttable_2): Likewise.
1804 (consttable_4): Likewise.
1805 (consttable_8): Likewise.
1806 (consttable_16): Likewise.
1807 (*thumb1_tablejump): Likewise.
1808 (prefetch): Likewise.
1809 (force_register_use): Likewise.
1810 (thumb_eh_return): Likewise.
1811 (load_tp_hard): Likewise.
1812 (load_tp_soft): Likewise.
1813 (tlscall): Likewise.
1814 (*arm_movtas_ze): Likewise.
1815 (*arm_rev): Likewise.
1816 (*arm_revsh): Likewise.
1817 (*arm_rev16): Likewise.
1818 * config/arm/thumb2.md
1819 (*thumb2_smaxsi3): Likewise.
1820 (*thumb2_sminsi3): Likewise.
1821 (*thumb32_umaxsi3): Likewise.
1822 (*thumb2_uminsi3): Likewise.
1823 (*thumb2_negdi2): Likewise.
1824 (*thumb2_abssi2): Likewise.
1825 (*thumb2_neg_abss): Likewise.
1826 (*thumb2_movsi_insn): Likewise.
1827 (tls_load_dot_plus_four): Likewise.
1828 (*thumb2_movhi_insn): Likewise.
1829 (*thumb2_mov_scc): Likewise.
1830 (*thumb2_mov_negs): Likewise.
1831 (*thumb2_mov_negs): Likewise.
1832 (*thumb2_mov_nots): Likewise.
1833 (*thumb2_mov_nots): Likewise.
1834 (*thumb2_movsicc_): Likewise.
1835 (*thumb2_movsfcc_soft_insn): Likewise.
1836 (*thumb2_indirect_jump): Likewise.
1837 (*thumb2_and_scc): Likewise.
1838 (*thumb2_ior_scc): Likewise.
1839 (*thumb2_ior_scc_strict_it): Likewise.
1840 (*thumb2_cond_move): Likewise.
1841 (*thumb2_cond_arith): Likewise.
1842 (*thumb2_cond_ari): Likewise.
1843 (*thumb2_cond_sub): Likewise.
1844 (*thumb2_negscc): Likewise.
1845 (*thumb2_movcond): Likewise.
1846 (thumb2_casesi_internal): Likewise.
1847 (thumb2_casesi_internal_pic): Likewise.
1848 (*thumb2_alusi3_short): Likewise.
1849 (*thumb2_mov<mode>_shortim): Likewise.
1850 (*thumb2_addsi_short): Likewise.
1851 (*thumb2_subsi_short): Likewise.
1852 (thumb2_addsi3_compare0): Likewise.
1853 (*thumb2_cbz): Likewise.
1854 (*thumb2_cbnz): Likewise.
1855 (*thumb2_one_cmplsi2_short): Likewise.
1856 (*thumb2_negsi2_short): Likewise.
1857 (*orsi_notsi_si): Likewise.
1858 * config/arm/arm1020e.md: Update with new attributes.
1859 * config/arm/arm1026ejs.md: Update with new attributes.
1860 * config/arm/arm1136jfs.md: Update with new attributes.
1861 * config/arm/arm926ejs.md: Update with new attributes.
1862 * config/arm/cortex-a15.md: Update with new attributes.
1863 * config/arm/cortex-a5.md: Update with new attributes.
1864 * config/arm/cortex-a53.md: Update with new attributes.
1865 * config/arm/cortex-a7.md: Update with new attributes.
1866 * config/arm/cortex-a8.md: Update with new attributes.
1867 * config/arm/cortex-a9.md: Update with new attributes.
1868 * config/arm/cortex-m4.md: Update with new attributes.
1869 * config/arm/cortex-r4.md: Update with new attributes.
1870 * config/arm/fa526.md: Update with new attributes.
1871 * config/arm/fa606te.md: Update with new attributes.
1872 * config/arm/fa626te.md: Update with new attributes.
1873 * config/arm/fa726te.md: Update with new attributes.
1875 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1877 * config/aarch64/aarch64-simd.md
1878 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
1879 <vwx> iterator to ensure correct register choice.
1880 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1881 (aarch64_sqdmull_n<mode>): Likewise.
1882 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1883 * config/aarch64/arm_neon.h
1884 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
1885 (vml<as><q>_n_<su>16): Likewise.
1886 (vml<as>l_high_lane<q>_<su>16): Likewise.
1887 (vml<as>l_high_n_<su>16): Likewise.
1888 (vml<as>l_lane<q>_<su>16): Likewise.
1889 (vml<as>l_n_<su>16): Likewise.
1890 (vmul<q>_lane<q>_<su>16): Likewise.
1891 (vmul<q>_n_<su>16): Likewise.
1892 (vmull_lane<q>_<su>16): Likewise.
1893 (vmull_n_<su>16): Likewise.
1894 (vmull_high_lane<q>_<su>16): Likewise.
1895 (vmull_high_n_<su>16): Likewise.
1896 (vqrdmulh<q>_n_s16): Likewise.
1898 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
1900 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
1901 have the correct lane parameter.
1903 2013-09-06 Richard Biener <rguenther@suse.de>
1905 * cfganal.c (control_dependences::~control_dependences):
1906 Properly free all of the vector.
1908 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
1911 * config/i386/i386.c (ix86_conditional_register_usage):
1912 Proper initialize extended SSE registers.
1914 2013-09-06 Jan Hubicka <jh@suse.cz>
1916 PR tree-optimization/58311
1917 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
1919 2013-09-06 Jan Hubicka <jh@suse.cz>
1921 * Makefile.in (tree-sra.o): Update dependencies.
1922 * tree-sra.c: Include ipa-utils.h
1923 (scan_function): Use recursive_call_p.
1924 (has_caller_p): New function.
1925 (cgraph_for_node_and_aliases): Count also callers of aliases.
1927 2013-09-06 Jan Hubicka <jh@suse.cz>
1930 * cgraph.h (symtab_semantically_equivalent_p): Declare.
1931 * tree-tailcall.c: Include ipa-utils.h.
1932 (find_tail_calls): Use it.
1933 * ipa-pure-const.c (check_call): Likewise.
1934 * ipa-utils.c (recursive_call_p): New function.
1935 * ipa-utils.h (recursive_call_p): Dclare.
1936 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
1937 (symtab_semantically_equivalent_p): New function.
1938 * Makefile.in (tree-tailcall.o): Update dependencies.
1940 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1942 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
1945 2013-09-06 Richard Biener <rguenther@suse.de>
1947 * lto-streamer.h (lto_global_var_decls): Remove.
1948 * Makefile.in (OBJS): Remove lto-symtab.o.
1949 (lto-symtab.o): Remove.
1950 (GTFILES): Remove lto-symtab.c
1951 * lto-symtab.c: Move to lto/
1953 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1955 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
1956 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
1957 (UNSPEC_FPINT_RINT): New constant definitions.
1958 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
1959 definition with 2 attributes.
1960 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
1961 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
1964 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1966 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
1967 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
1968 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
1969 attribute to "z196_alone".
1970 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
1973 2013-09-06 Richard Biener <rguenther@suse.de>
1975 * basic-block.h (class control_dependences): New.
1976 * tree-ssa-dce.c (control_dependence_map): Remove.
1978 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
1979 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
1980 find_pdom, find_control_dependence, find_all_control_dependences):
1982 (mark_control_dependent_edges_necessary,
1983 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
1984 tree_dce_done, perform_tree_ssa_dce): Adjust.
1985 * cfganal.c (set_control_dependence_map_bit,
1986 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
1987 find_all_control_dependences): Move from tree-ssa-dce.c and
1988 implement as methods of control_dependences class.
1989 (control_dependences::control_dependences): New.
1990 (control_dependences::~control_dependences): Likewise.
1991 (control_dependences::get_edges_dependent_on): Likewise.
1992 (control_dependences::get_edge): Likewise.
1994 2013-09-04 Jan Hubicka <jh@suse.cz>
1996 * tree.c (types_same_for_odr): Drop overactive check.
1997 * ipa-devirt.c (hash_type_name): Likewise.
1999 2013-09-04 Jan Hubicka <jh@suse.cz>
2001 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2002 (analyze_functions): ... here.
2004 2013-09-04 Jan Hubicka <jh@suse.cz>
2007 * cgraphunit.c (analyze_functions): Clear AUX fields
2008 after processing; initialize assembler name has.
2010 2013-09-05 Jeff Law <law@redhat.com>
2012 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2013 from thread_around_empty_block. Record threading path into PATH.
2014 Recurse if threading through the initial block is successful.
2015 (thread_across_edge): Corresponding changes to slightly simplify.
2017 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2019 * config/aarch64/aarch64.md
2020 (type): Remove frecpe, frecps, frecpx.
2021 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2022 fix to be a TARGET_SIMD instruction.
2023 (aarch64_frecps): Remove.
2024 * config/aarch64/aarch64-simd.md
2025 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2026 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2028 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2029 Sofiane Naci <sofiane.naci@arm.com>
2031 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2032 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2033 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2034 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2035 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2036 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2037 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2038 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2039 into "clz, "rbit". Rename "shift" to "shift_imm".
2040 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2041 changes. Update for attribute changes all occurrences of arlo_* and
2043 * config/arm/arm-fixed.md: Update for attribute changes
2044 all occurrences of arlo_* types.
2045 * config/arm/thumb2.md: Update for attribute changes all occurrences
2047 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2048 (cortexa7_older_only): Likewise.
2049 (cortexa7_younger): Likewise.
2050 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2051 (1020alu_shift_op): Likewise.
2052 (1020alu_shift_reg_op): Likewise.
2053 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2054 (alu_shift_op): Likewise.
2055 (alu_shift_reg_op): Likewise.
2056 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2057 (11_alu_shift_op): Likewise.
2058 (11_alu_shift_reg_op): Likewise.
2059 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2060 (9_alu_shift_reg_op): Likewise.
2061 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
2063 (cortex_a15_alu_shift): Likewise.
2064 (cortex_a15_alu_shift_reg): Likewise.
2065 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
2067 (cortex_a5_alu_shift): Likewise.
2068 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
2070 (cortex_a53_alu_shift): Likewise.
2071 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
2073 (cortex_a7_alu_reg): Likewise.
2074 (cortex_a7_alu_shift): Likewise.
2075 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
2077 (cortex_a8_alu_shift): Likewise.
2078 (cortex_a8_alu_shift_reg): Likewise.
2079 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2080 (cortex_a9_dp_shift): Likewise.
2081 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
2083 * config/arm/cortex-r4.md
2084 (cortex_r4_alu): Update for attribute changes.
2085 (cortex_r4_mov): Likewise.
2086 (cortex_r4_alu_shift_reg): Likewise.
2087 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2088 (526_alu_shift_op): Likewise.
2089 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2090 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2091 (626te_alu_shift_op): Likewise.
2092 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
2093 (726te_alu_shift_op): Likewise.
2094 (726te_alu_shift_reg_op): Likewise.
2095 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2096 (mp626_alu_shift_op): Likewise.
2097 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
2098 (pj4_alu_conds): Likewise.
2099 (pj4_shift): Likewise.
2100 (pj4_shift_conds): Likewise.
2101 (pj4_alu_shift): Likewise.
2102 (pj4_alu_shift_conds): Likewise.
2103 * config/aarch64/aarch64.md: Update for attribute change
2104 all occurrences of arlo_* and shift* types.
2106 2013-09-05 Mike Stump <mikestump@comcast.net>
2108 * tree.h: Move documentation for tree_function_decl to tree-core.h
2109 with the declaration.
2111 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
2114 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
2115 looking for widest mode.
2117 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
2119 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
2121 2013-09-05 Richard Biener <rguenther@suse.de>
2123 PR tree-optimization/58137
2124 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
2125 Do not create vectors of pointers.
2126 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
2127 types for the components of the vector initializer.
2128 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
2129 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
2131 2013-09-05 Martin Jambor <mjambor@suse.cz>
2133 * ipa-prop.c (remove_described_reference): Accept missing references,
2134 return false if that hppens, otherwise return true.
2135 (cgraph_node_for_jfunc): New function.
2136 (try_decrement_rdesc_refcount): Likewise.
2137 (try_make_edge_direct_simple_call): Use them.
2138 (ipa_edge_removal_hook): Remove references from rdescs.
2139 (ipa_edge_duplication_hook): Clone rdescs and their references
2140 when the new edge has the same caller as the old one.
2141 * cgraph.c (cgraph_resolve_speculation): Remove speculative
2142 reference before removing any edges.
2144 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
2146 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
2148 * thumb2.md (thumb2_storewb_parisi): New pattern.
2150 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
2152 * config/aarch64/aarch64-option-extensions.def: Add
2153 AARCH64_OPT_EXTENSION of 'crc'.
2154 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
2155 (AARCH64_ISA_CRC): Ditto.
2156 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
2157 description of the CRC extension.
2159 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
2161 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
2162 * config/rs6000/linux.h: Ditto.
2163 * alpha/linux.h: Ditto.
2164 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
2165 no_c99_libc_has_function.
2166 * config/c6x/uclinux-elf.h: Ditto.
2167 * config/lm32/uclinux-elf.h: Ditto.
2168 * config/m68k/uclinux.h: Ditto.
2169 * config/moxie/uclinux.h: Ditto.
2170 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
2171 (crisv32-*-linux*, cris-*-linux*): Ditto.
2172 * config/bfin/bfin.c: Include "tm_p.h".
2174 2013-09-05 Richard Biener <rguenther@suse.de>
2176 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
2177 check for a definition without a basic-block.
2179 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2180 Sofiane Naci <sofiane.naci@arm.com>
2182 * config/aarch64/aarch64.md
2183 (*movti_aarch64): Rename r_2_f and f_2_r.
2184 (*movsf_aarch64): Likewise.
2185 (*movdf_aarch64): Likewise.
2186 (*movtf_aarch64): Likewise.
2187 (aarch64_movdi_<mode>low): Likewise.
2188 (aarch64_movdi_<mode>high): Likewise.
2189 (aarch64_mov<mode>high_di): Likewise.
2190 (aarch64_mov<mode>low_di): Likewise.
2191 (aarch64_movtilow_tilow): Likewise.
2192 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
2193 values to config/arm/types.md
2194 (attribute "conds"): Update for attribute change.
2195 (anddi3_insn): Likewise.
2196 (iordi3_insn): Likewise.
2197 (xordi3_insn): Likewise.
2198 (one_cmpldi2): Likewise.
2199 * config/arm/types.md (type): Add Neon types.
2200 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
2201 use "type" attribute.
2202 (movmisalign<mode>_neon_store): Likewise.
2203 (movmisalign<mode>_neon_load): Likewise.
2204 (vec_set<mode>_internal): Likewise.
2205 (vec_setv2di_internal): Likewise.
2206 (vec_extract<mode>): Likewise.
2207 (vec_extractv2di): Likewise.
2208 (add<mode>3_neon): Likewise.
2209 (adddi3_neon): Likewise.
2210 (sub<mode>3_neon): Likewise.
2211 (subdi3_neon): Likewise.
2212 (mul<mode>3_neon): Likewise.
2213 (mul<mode>3add<mode>_neon): Likewise.
2214 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
2215 (fma<VCVTF:mode>4)): Likewise.
2216 (fma<VCVTF:mode>4_intrinsic): Likewise.
2217 (fmsub<VCVTF:mode>4)): Likewise.
2218 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
2219 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
2220 (ior<mode>3): Likewise.
2221 (and<mode>3): Likewise.
2222 (anddi3_neon): Likewise.
2223 (orn<mode>3_neon): Likewise.
2224 (orndi3_neon): Likewise.
2225 (bic<mode>3_neon): Likewise.
2226 (bicdi3_neon): Likewise.
2227 (xor<mode>3): Likewise.
2228 (one_cmpl<mode>2): Likewise.
2229 (abs<mode>2): Likewise.
2230 (neg<mode>2): Likewise.
2231 (umin<mode>3_neon): Likewise.
2232 (umax<mode>3_neon): Likewise.
2233 (smin<mode>3_neon): Likewise.
2234 (smax<mode>3_neon): Likewise.
2235 (vashl<mode>3): Likewise.
2236 (vashr<mode>3_imm): Likewise.
2237 (vlshr<mode>3_imm): Likewise.
2238 (ashl<mode>3_signed): Likewise.
2239 (ashl<mode>3_unsigned): Likewise.
2240 (neon_load_count): Likewise.
2241 (ashldi3_neon_noclobber): Likewise.
2242 (signed_shift_di3_neon): Likewise.
2243 (unsigned_shift_di3_neon): Likewise.
2244 (ashrdi3_neon_imm_noclobber): Likewise.
2245 (lshrdi3_neon_imm_noclobber): Likewise.
2246 (widen_ssum<mode>3): Likewise.
2247 (widen_usum<mode>3): Likewise.
2248 (quad_halves_<code>v4si): Likewise.
2249 (quad_halves_<code>v4sf): Likewise.
2250 (quad_halves_<code>v8hi): Likewise.
2251 (quad_halves_<code>v16qi): Likewise.
2252 (reduc_splus_v2di): Likewise.
2253 (neon_vpadd_internal<mode>): Likewise.
2254 (neon_vpsmin<mode>): Likewise.
2255 (neon_vpsmax<mode>): Likewise.
2256 (neon_vpumin<mode>): Likewise.
2257 (neon_vpumax<mode>): Likewise.
2258 (ss_add<mode>_neon): Likewise.
2259 (us_add<mode>_neon): Likewise.
2260 (ss_sub<mode>_neon): Likewise.
2261 (us_sub<mode>_neon): Likewise.
2262 (neon_vadd<mode>_unspec): Likewise.
2263 (neon_vaddl<mode>): Likewise.
2264 (neon_vaddw<mode>): Likewise.
2265 (neon_vhadd<mode>): Likewise.
2266 (neon_vqadd<mode>): Likewise.
2267 (neon_vaddhn<mode>): Likewise.
2268 (neon_vmul<mode>): Likewise.
2269 (neon_vmla<mode>): Likewise.
2270 (neon_vmlal<mode>): Likewise.
2271 (neon_vmls<mode>): Likewise.
2272 (neon_vmlsl<mode>): Likewise.
2273 (neon_vqdmulh<mode>): Likewise.
2274 (neon_vqdmlal<mode>): Likewise.
2275 (neon_vqdmlsl<mode>): Likewise.
2276 (neon_vmull<mode>): Likewise.
2277 (neon_vqdmull<mode>): Likewise.
2278 (neon_vsub<mode>_unspec): Likewise.
2279 (neon_vsubl<mode>): Likewise.
2280 (neon_vsubw<mode>): Likewise.
2281 (neon_vqsub<mode>): Likewise.
2282 (neon_vhsub<mode>): Likewise.
2283 (neon_vsubhn<mode>): Likewise.
2284 (neon_vceq<mode>): Likewise.
2285 (neon_vcge<mode>): Likewise.
2286 (neon_vcgeu<mode>): Likewise.
2287 (neon_vcgt<mode>): Likewise.
2288 (neon_vcgtu<mode>): Likewise.
2289 (neon_vcle<mode>): Likewise.
2290 (neon_vclt<mode>): Likewise.
2291 (neon_vcage<mode>): Likewise.
2292 (neon_vcagt<mode>): Likewise.
2293 (neon_vtst<mode>): Likewise.
2294 (neon_vabd<mode>): Likewise.
2295 (neon_vabdl<mode>): Likewise.
2296 (neon_vaba<mode>): Likewise.
2297 (neon_vabal<mode>): Likewise.
2298 (neon_vmax<mode>): Likewise.
2299 (neon_vmin<mode>): Likewise.
2300 (neon_vpaddl<mode>): Likewise.
2301 (neon_vpadal<mode>): Likewise.
2302 (neon_vpmax<mode>): Likewise.
2303 (neon_vpmin<mode>): Likewise.
2304 (neon_vrecps<mode>): Likewise.
2305 (neon_vrsqrts<mode>): Likewise.
2306 (neon_vqabs<mode>): Likewise.
2307 (neon_vqneg<mode>): Likewise.
2308 (neon_vcls<mode>): Likewise.
2309 (clz<mode>2): Likewise.
2310 (popcount<mode>2): Likewise.
2311 (neon_vrecpe): Likewise.
2312 (neon_vrsqrte): Likewise.
2313 (neon_vget_lane<mode>_sext_internal): Likewise.
2314 (neon_vget_lane<mode>_zext_internal): Likewise.
2315 (neon_vdup_n<mode>): Likewise.
2316 (neon_vdup_nv2di): Likewise.
2317 (neon_vdpu_lane<mode>_internal): Likewise.
2318 (neon_vswp<mode>): Likewise.
2319 (float<mode><V_cvtto>2): Likewise.
2320 (floatuns<mode><V_cvtto>2): Likewise.
2321 (fix_trunc<mode><V_cvtto>)2): Likewise
2322 (fixuns_trunc<mode><V_cvtto)2): Likewise.
2323 (neon_vcvt<mode>): Likewise.
2324 (neon_vcvtv4sfv4hf): Likewise.
2325 (neon_vcvtv4hfv4sf): Likewise.
2326 (neon_vcvt_n<mode>): Likewise.
2327 (neon_vmovn<mode>): Likewise.
2328 (neon_vqmovn<mode>): Likewise.
2329 (neon_vqmovun<mode>): Likewise.
2330 (neon_vmovl<mode>): Likewise.
2331 (neon_vmul_lane<mode>): Likewise.
2332 (neon_vmull_lane<mode>): Likewise.
2333 (neon_vqdmull_lane<mode>): Likewise.
2334 (neon_vqdmulh_lane<mode>): Likewise.
2335 (neon_vmla_lane<mode>): Likewise.
2336 (neon_vmlal_lane<mode>): Likewise.
2337 (neon_vqdmlal_lane<mode>): Likewise.
2338 (neon_vmls_lane<mode>): Likewise.
2339 (neon_vmlsl_lane<mode>): Likewise.
2340 (neon_vqdmlsl_lane<mode>): Likewise.
2341 (neon_vext<mode>): Likewise.
2342 (neon_vrev64<mode>): Likewise.
2343 (neon_vrev32<mode>): Likewise.
2344 (neon_vrev16<mode>): Likewise.
2345 (neon_vbsl<mode>_internal): Likewise.
2346 (neon_vshl<mode>): Likewise.
2347 (neon_vqshl<mode>): Likewise.
2348 (neon_vshr_n<mode>): Likewise.
2349 (neon_vshrn_n<mode>): Likewise.
2350 (neon_vqshrn_n<mode>): Likewise.
2351 (neon_vqshrun_n<mode>): Likewise.
2352 (neon_vshl_n<mode>): Likewise.
2353 (neon_vqshl_n<mode>): Likewise.
2354 (neon_vqshlu_n<mode>): Likewise.
2355 (neon_vshll_n<mode>): Likewise.
2356 (neon_vsra_n<mode>): Likewise.
2357 (neon_vsri_n<mode>): Likewise.
2358 (neon_vsli_n<mode>): Likewise.
2359 (neon_vtbl1v8qi): Likewise.
2360 (neon_vtbl2v8qi): Likewise.
2361 (neon_vtbl3v8qi): Likewise.
2362 (neon_vtbl4v8qi): Likewise.
2363 (neon_vtbx1v8qi): Likewise.
2364 (neon_vtbx2v8qi): Likewise.
2365 (neon_vtbx3v8qi): Likewise.
2366 (neon_vtbx4v8qi): Likewise.
2367 (neon_vtrn<mode>_internal): Likewise.
2368 (neon_vzip<mode>_internal): Likewise.
2369 (neon_vuzp<mode>_internal): Likewise.
2370 (neon_vld1<mode>): Likewise.
2371 (neon_vld1_lane<mode>): Likewise.
2372 (neon_vld1_dup<mode>): Likewise.
2373 (neon_vld1_dupv2di): Likewise.
2374 (neon_vst1<mode>): Likewise.
2375 (neon_vst1_lane<mode>): Likewise.
2376 (neon_vld2<mode>): Likewise.
2377 (neon_vld2_lane<mode>): Likewise.
2378 (neon_vld2_dup<mode>): Likewise.
2379 (neon_vst2<mode>): Likewise.
2380 (neon_vst2_lane<mode>): Likewise.
2381 (neon_vld3<mode>): Likewise.
2382 (neon_vld3qa<mode>): Likewise.
2383 (neon_vld3qb<mode>): Likewise.
2384 (neon_vld3_lane<mode>): Likewise.
2385 (neon_vld3_dup<mode>): Likewise.
2386 (neon_vst3<mode>): Likewise.
2387 (neon_vst3qa<mode>): Likewise.
2388 (neon_vst3qb<mode>): Likewise.
2389 (neon_vst3_lane<mode>): Likewise.
2390 (neon_vld4<mode>): Likewise.
2391 (neon_vld4qa<mode>): Likewise.
2392 (neon_vld4qb<mode>): Likewise.
2393 (neon_vld4_lane<mode>): Likewise.
2394 (neon_vld4_dup<mode>): Likewise.
2395 (neon_vst4<mode>): Likewise.
2396 (neon_vst4qa<mode>): Likewise.
2397 (neon_vst4qb<mode>): Likewise.
2398 (neon_vst4_lane<mode>): Likewise.
2399 (neon_vec_unpack<US>_lo_<mode>): Likewise.
2400 (neon_vec_unpack<US>_hi_<mode>): Likewise.
2401 (neon_vec_<US>mult_lo_<mode>): Likewise.
2402 (neon_vec_<US>mult_hi_<mode>): Likewise.
2403 (neon_vec_<US>shiftl_<mode>): Likewise.
2404 (neon_unpack<US>_<mode>): Likewise.
2405 (neon_vec_<US>mult_<mode>): Likewise.
2406 (vec_pack_trunc_<mode>): Likewise.
2407 (neon_vec_pack_trunk_<mode>): Likewise.
2408 (neon_vabd<mode>_2): Likewise.
2409 (neon_vabd<mode>_3): Likewise.
2410 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
2411 (thumb2_movsi_vfp): Likewise.
2412 (movdi_vfp): Likewise.
2413 (movdi_vfp_cortexa8): Likewise.
2414 (movhf_vfp_neon): Likewise.
2415 (movhf_vfp): Likewiwse.
2416 (movsf_vfp): Likewiwse.
2417 (thumb2_movsf_vfp): Likewiwse.
2418 (movdf_vfp): Likewise.
2419 (thumb2_movdf_vfp): Likewise.
2420 (movsfcc_vfp): Likewise.
2421 (thumb2_movsfcc_vfp): Likewise.
2422 (movdfcc_vfp): Likewise.
2423 (thumb2_movdfcc_vfp): Likewise.
2424 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
2425 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
2426 (v10_v2c): Likewise.
2427 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
2429 (cortex_a15_neon_int_2): Likewise.
2430 (cortex_a15_neon_int_3): Likewise.
2431 (cortex_a15_neon_int_4): Likewise.
2432 (cortex_a15_neon_int_5): Likewise.
2433 (cortex_a15_neon_vqneg_vqabs): Likewise.
2434 (cortex_a15_neon_vmov): Likewise.
2435 (cortex_a15_neon_vaba): Likewise.
2436 (cortex_a15_neon_vaba_qqq): Likewise.
2437 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2438 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2439 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
2440 scalar_64_32_long_scalar): Likewise.
2441 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2442 (cortex_a15_neon_mla_qqq_8_16): Likewise.
2443 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2444 lotype_qdd_64_32_long): Likewise.
2445 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2446 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2447 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
2448 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2449 (cortex_a15_neon_shift_1): Likewise.
2450 (cortex_a15_neon_shift_2): Likewise.
2451 (cortex_a15_neon_shift_3): Likewise.
2452 (cortex_a15_neon_vshl_ddd): Likewise.
2453 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2454 (cortex_a15_neon_vsra_vrsra): Likewise.
2455 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
2456 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
2457 (cortex_a15_neon_fp_vmul_ddd): Likewise.
2458 (cortex_a15_neon_fp_vmul_qqd): Likewise.
2459 (cortex_a15_neon_fp_vmla_ddd): Likewise.
2460 (cortex_a15_neon_fp_vmla_qqq): Likewise.
2461 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
2462 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
2463 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2464 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2465 (cortex_a15_neon_bp_simple): Likewise.
2466 (cortex_a15_neon_bp_2cycle): Likewise.
2467 (cortex_a15_neon_bp_3cycle): Likewise.
2468 (cortex_a15_neon_vld1_1_2_regs): Likewise.
2469 (cortex_a15_neon_vld1_3_4_regs): Likewise.
2470 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2471 (cortex_a15_neon_vld2_4_regs): Likewise.
2472 (cortex_a15_neon_vld3_vld4): Likewise.
2473 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2474 (cortex_a15_neon_vst1_3_4_regs): Likewise.
2475 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
2476 (cortex_a15_neon_vst3_vst4): Likewise.
2477 (cortex_a15_neon_vld1_vld2_lane): Likewise.
2478 (cortex_a15_neon_vld3_vld4_lane" 10
2479 (cortex_a15_neon_vst1_vst2_lane): Likewise.
2480 (cortex_a15_neon_vst3_vst4_lane): Likewise.
2481 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
2482 (cortex_a15_neon_ldm_2): Likewise.0
2483 (cortex_a15_neon_stm_2): Likewise.
2484 (cortex_a15_neon_mcr): Likewise.
2485 (cortex_a15_neon_mcr_2_mcrr): Likewise.
2486 (cortex_a15_neon_mrc): Likewise.
2487 (cortex_a15_neon_mrrc): Likewise.
2488 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2490 (cortex_a15_alu_shift): Likewise.
2491 (cortex_a15_alu_shift_reg): Likewise.
2492 (cortex_a15_mult32): Likewise.
2493 (cortex_a15_mult64): Likewise.
2494 (cortex_a15_block): Likewise.
2495 (cortex_a15_branch): Likewise.
2496 (cortex_a15_load1): Likewise.
2497 (cortex_a15_load3): Likewise.
2498 (cortex_a15_store1): Likewise.
2499 (cortex_a15_store3): Likewise.
2500 (cortex_a15_call): Likewise.
2501 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
2502 (cortex_a5_f2r): Likewise.
2503 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
2505 (cortex_a53_f2r): Likewise.
2506 * config/arm/cortex-a7.md
2507 (cortex_a7_branch): Update for attribute change.
2508 (cortex_a7_call): Likewise.
2509 (cortex_a7_alu_imm): Likewise.
2510 (cortex_a7_alu_reg): Likewise.
2511 (cortex_a7_alu_shift): Likewise.
2512 (cortex_a7_mul): Likewise.
2513 (cortex_a7_load1): Likewise.
2514 (cortex_a7_store1): Likewise.
2515 (cortex_a7_load2): Likewise.
2516 (cortex_a7_store2): Likewise.
2517 (cortex_a7_load3): Likewise.
2518 (cortex_a7_store3): Likewise.
2519 (cortex_a7_load4): Likewise.
2520 (cortex_a7_store4): Likewise.
2521 (cortex_a7_fpalu): Likewise.
2522 (cortex_a7_fconst): Likewise.
2523 (cortex_a7_fpmuls): Likewise.
2524 (cortex_a7_neon_mul): Likewise.
2525 (cortex_a7_fpmacs): Likewise.
2526 (cortex_a7_neon_mla: Likewise.
2527 (cortex_a7_fpmuld: Likewise.
2528 (cortex_a7_fpmacd: Likewise.
2529 (cortex_a7_fpfmad: Likewise.
2530 (cortex_a7_fdivs: Likewise.
2531 (cortex_a7_fdivd: Likewise.
2532 (cortex_a7_r2f: Likewise.
2533 (cortex_a7_f2r: Likewise.
2534 (cortex_a7_f_flags: Likewise.
2535 (cortex_a7_f_loads: Likewise.
2536 (cortex_a7_f_loadd: Likewise.
2537 (cortex_a7_f_stores: Likewise.
2538 (cortex_a7_f_stored: Likewise.
2539 (cortex_a7_neon): Likewise.
2540 * config/arm/cortex-a8-neon.md
2541 (cortex_a8_neon_mrc): Update for attribute change.
2542 (cortex_a8_neon_mrrc): Likewise.
2543 (cortex_a8_neon_int_1): Likewise.
2544 (cortex_a8_neon_int_2): Likewise.
2545 (cortex_a8_neon_int_3): Likewise.
2546 (cortex_a8_neon_int_4): Likewise.
2547 (cortex_a8_neon_int_5): Likewise.
2548 (cortex_a8_neon_vqneg_vqabs): Likewise.
2549 (cortex_a8_neon_vmov): Likewise.
2550 (cortex_a8_neon_vaba): Likewise.
2551 (cortex_a8_neon_vaba_qqq): Likewise.
2552 (cortex_a8_neon_vsma): Likewise.
2553 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2554 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2555 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
2556 long_scalar): Likewise.
2557 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2558 (cortex_a8_neon_mla_qqq_8_16): Likewise.
2559 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2560 long_scalar_qdd_64_32_long): Likewise.
2561 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2562 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2563 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
2564 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2565 (cortex_a8_neon_shift_1): Likewise.
2566 (cortex_a8_neon_shift_2): Likewise.
2567 (cortex_a8_neon_shift_3): Likewise.
2568 (cortex_a8_neon_vshl_ddd): Likewise.
2569 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2570 (cortex_a8_neon_vsra_vrsra): Likewise.
2571 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
2572 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
2573 (cortex_a8_neon_fp_vsum): Likewise.
2574 (cortex_a8_neon_fp_vmul_ddd): Likewise.
2575 (cortex_a8_neon_fp_vmul_qqd): Likewise.
2576 (cortex_a8_neon_fp_vmla_ddd): Likewise.
2577 (cortex_a8_neon_fp_vmla_qqq): Likewise.
2578 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
2579 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
2580 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2581 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2582 (cortex_a8_neon_bp_simple): Likewise.
2583 (cortex_a8_neon_bp_2cycle): Likewise.
2584 (cortex_a8_neon_bp_3cycle): Likewise.
2585 (cortex_a8_neon_ldr): Likewise.
2586 (cortex_a8_neon_str): Likewise.
2587 (cortex_a8_neon_vld1_1_2_regs): Likewise.
2588 (cortex_a8_neon_vld1_3_4_regs): Likewise.
2589 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2590 (cortex_a8_neon_vld2_4_regs): Likewise.
2591 (cortex_a8_neon_vld3_vld4): Likewise.
2592 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2593 (cortex_a8_neon_vst1_3_4_regs): Likewise.
2594 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
2595 (cortex_a8_neon_vst3_vst4): Likewise.
2596 (cortex_a8_neon_vld1_vld2_lane): Likewise.
2597 (cortex_a8_neon_vld3_vld4_lane): Likewise.
2598 (cortex_a8_neon_vst1_vst2_lane): Likewise.
2599 (cortex_a8_neon_vst3_vst4_lane): Likewise.
2600 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
2601 (cortex_a8_neon_mcr): Likewise.
2602 (cortex_a8_neon_mcr_2_mcrr): Likewise.
2603 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2604 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
2606 (ca9_neon_mrrc): Likewise.
2607 (cortex_a9_neon_int_1): Likewise.
2608 (cortex_a9_neon_int_2): Likewise.
2609 (cortex_a9_neon_int_3): Likewise.
2610 (cortex_a9_neon_int_4): Likewise.
2611 (cortex_a9_neon_int_5): Likewise.
2612 (cortex_a9_neon_vqneg_vqabs): Likewise.
2613 (cortex_a9_neon_vmov): Likewise.
2614 (cortex_a9_neon_vaba): Likewise.
2615 (cortex_a9_neon_vaba_qqq): Likewise.
2616 (cortex_a9_neon_vsma): Likewise.
2617 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2618 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2619 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
2620 long_scalar): Likewise.
2621 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2622 (cortex_a9_neon_mla_qqq_8_16): Likewise.
2623 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2624 long_scalar_qdd_64_32_long): Likewise.
2625 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2626 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2627 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
2628 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2629 (cortex_a9_neon_shift_1): Likewise.
2630 (cortex_a9_neon_shift_2): Likewise.
2631 (cortex_a9_neon_shift_3): Likewise.
2632 (cortex_a9_neon_vshl_ddd): Likewise.
2633 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2634 (cortex_a9_neon_vsra_vrsra): Likewise.
2635 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
2636 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
2637 (cortex_a9_neon_fp_vsum): Likewise.
2638 (cortex_a9_neon_fp_vmul_ddd): Likewise.
2639 (cortex_a9_neon_fp_vmul_qqd): Likewise.
2640 (cortex_a9_neon_fp_vmla_ddd): Likewise.
2641 (cortex_a9_neon_fp_vmla_qqq): Likewise.
2642 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
2643 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
2644 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2645 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2646 (cortex_a9_neon_bp_simple): Likewise.
2647 (cortex_a9_neon_bp_2cycle): Likewise.
2648 (cortex_a9_neon_bp_3cycle): Likewise.
2649 (cortex_a9_neon_ldr): Likewise.
2650 (cortex_a9_neon_str): Likewise.
2651 (cortex_a9_neon_vld1_1_2_regs): Likewise.
2652 (cortex_a9_neon_vld1_3_4_regs): Likewise.
2653 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2654 (cortex_a9_neon_vld2_4_regs): Likewise.
2655 (cortex_a9_neon_vld3_vld4): Likewise.
2656 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2657 (cortex_a9_neon_vst1_3_4_regs): Likewise.
2658 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
2659 (cortex_a9_neon_vst3_vst4): Likewise.
2660 (cortex_a9_neon_vld1_vld2_lane): Likewise.
2661 (cortex_a9_neon_vld3_vld4_lane): Likewise.
2662 (cortex_a9_neon_vst1_vst2_lane): Likewise.
2663 (cortex_a9_neon_vst3_vst4_lane): Likewise.
2664 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
2665 (cortex_a9_neon_mcr): Likewise.
2666 (cortex_a9_neon_mcr_2_mcrr): Likewise.
2667 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
2668 (cortex_a9_fps): Likewise.
2669 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
2671 (cortex_m4_fmuls): Likewise.
2672 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
2674 (cortex_r4_mrc): Likewise.
2675 * config/arm/iterators.md: Update comment referring to neon_type.
2676 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
2677 (iwmmxt_movsi_insn): Likewise.
2678 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
2680 (pj4_core_to_vfp): Likewise.
2681 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
2683 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
2684 (vfp_fstore): Likewise.
2685 * doc/md.texi: Change references to neon_type to refer to type.
2687 2013-09-04 Dodji Seketeli <dodji@redhat.com>
2689 * tree.h (DECL_BUILT_IN): Fix typo in comment.
2691 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
2693 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
2696 2013-09-04 Easwaran Raman <eraman@google.com>
2699 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
2700 (build_and_add_sum): Use it.
2701 (appears_later_in_bb): Simplify code.
2703 2013-09-04 Teresa Johnson <tejohnson@google.com>
2705 * dumpfile.c (dump_finish): Don't close stderr/stdout.
2707 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
2709 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
2711 2013-09-04 Jan Hubicka <jh@suse.cz>
2713 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
2714 * ipa-devirt.c: Include diganostic.h
2715 (odr_type_d): Add types and types_set.
2716 (hash_type_name): Work for types with vtables during LTO.
2717 (odr_hasher::remove): Fix comment; destroy types_set.
2718 (add_type_duplicate): New function,
2719 (get_odr_type): Use it.
2720 (dump_type_inheritance_graph): Dump type duplicates.
2721 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
2723 * tree.c (types_same_for_odr): Give exact answers on types with
2726 2013-09-04 Dodji Seketeli <dodji@redhat.com>
2728 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
2729 explaining their differences.
2731 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
2733 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
2735 2013-09-03 Jeff Law <law@redhat.com>
2737 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
2738 when not threading through a joiner block. Pass joiner/no joiner
2739 state to register_jump_thread.
2740 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
2741 state from argument rather than implying on path length.
2742 Dump the entire jump thread path into debugging dump.
2743 * tree-flow.h (register_jump_thread): Update prototype.
2745 2013-08-29 Xinliang David Li <davidxl@google.com>
2747 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
2748 Remove a trivial gcc_assert.
2750 2013-08-29 Xinliang David Li <davidxl@google.com>
2752 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
2753 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
2754 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
2755 Delay base decl alignment adjustment.
2756 * tree-vectorizer.c (vect_destroy_datarefs): New function.
2757 * tree-vectorizer.h: New data structure.
2758 (set_dr_misalignment): New function.
2759 (dr_misalignment): Ditto.
2760 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
2761 (vectorizable_load): Ditto.
2762 (ensure_base_align): New function.
2763 (vectorize_loops): Add dbg_cnt support.
2764 (execute_vect_slp): Ditto.
2765 * dbgcnt.def: New debug counter.
2766 * Makefile: New dependency.
2768 2013-09-03 Meador Inge <meadori@codesourcery.com>
2772 2013-08-30 Meador Inge <meadori@codesourcery.com>
2774 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2776 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
2778 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
2779 function descriptor.
2781 2013-09-03 Richard Biener <rguenther@suse.de>
2783 * tree-affine.c (add_elt_to_tree): Fix association issue,
2784 avoid useless converts and make sure to always return a
2785 properly typed result.
2787 2013-09-03 Richard Biener <rguenther@suse.de>
2790 * fold-const.c (negate_expr_p): Fix division case.
2791 (negate_expr): Likewise.
2793 2013-09-03 Richard Biener <rguenther@suse.de>
2796 * tree-streamer-out.c: Include tm.h.
2797 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
2799 2013-09-03 Jan Hubicka <jh@suse.cz>
2801 * tree-profile.c (tree_profiling): Cleanup CFG when done.
2803 2013-09-03 Alan Modra <amodra@gmail.com>
2805 * config.gcc (powerpc*-*-linux*): Add support for little-endian
2806 multilibs to big-endian target and vice versa.
2807 * config/rs6000/t-linux64: Use := assignment on all vars.
2808 (MULTILIB_EXTRA_OPTS): Remove fPIC.
2809 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
2810 * config/rs6000/t-linux64le: New file.
2811 * config/rs6000/t-linux64bele: New file.
2812 * config/rs6000/t-linux64lebe: New file.
2814 2013-09-02 Jan Hubicka <jh@suse.cz>
2816 * ipa-inline-transform.c (inline_transform): Do not
2817 optimize_inline_calls when not optimizing.
2819 2013-09-02 Jan Hubicka <jh@suse.cz>
2821 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
2822 duplicated nodes for assembler names.
2823 * symtab.c (symtab_unregister_node): Do not attempt to unlink
2824 hard registers from assembler name hash.
2826 2013-09-02 Jan Hubicka <jh@suse.cz>
2828 * ipa-split.c (execute_split_functions): Split externally visible
2829 functions called once.
2831 2013-09-02 Martin Jambor <mjambor@suse.cz>
2834 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
2835 linked list. When finding the correct duplicate, also consider also
2836 the caller in additon to its inlined_to node.
2838 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
2840 * config/aarch64/aarch64-simd-builtins.def
2841 (dup_lane_scalar): Remove.
2842 * config/aarch64/aarch64-simd.md
2843 (aarch64_simd_dup): Add 'w->w' alternative.
2844 (aarch64_dup_lane<mode>): Allow for VALL.
2845 (aarch64_dup_lane_scalar<mode>): Remove.
2846 (aarch64_dup_lane_<vswap_width_name><mode>): New.
2847 (aarch64_get_lane_signed<mode>): Add w->w altenative.
2848 (aarch64_get_lane_unsigned<mode>): Likewise.
2849 (aarch64_get_lane<mode>): Likewise.
2850 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
2851 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
2852 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
2853 (VCON): Change container of V2SF.
2854 (vswap_width_name): Likewise.
2855 * config/aarch64/arm_neon.h
2856 (__aarch64_vdup_lane_any): New.
2857 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2858 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
2859 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2861 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
2864 * expr.c (emit_move_complex): Do not move complex FP values as parts if
2865 the source or the destination is a single hard register.
2867 2013-09-02 Richard Biener <rguenther@suse.de>
2870 * tree-scalar-evolution.c (instantiate_scev_name): Allow
2873 2013-09-02 Richard Biener <rguenther@suse.de>
2875 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
2876 arithmetic to sizetype.
2878 2013-09-02 Bin Cheng <bin.cheng@arm.com>
2880 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
2881 Find auto-increment use both before and after candidate.
2883 2013-09-02 Marek Polacek <polacek@redhat.com>
2885 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
2887 2013-09-01 Jan Hubicka <jh@suse.cz>
2889 * Makefile.in: Add ipa-profile.o
2890 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
2891 * cgraph.c (struct cgraph_propagate_frequency_data,
2892 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
2893 ipa-profile.c; replace cgraph_ by ipa_ prefix.
2894 * cgraph.h (cgraph_propagate_frequency): Remove.
2895 * ipa-inline-analysis.c: Include ipa-utils.h;
2896 drop duplicated cfgloop.h.
2897 (inline_update_callee_summaries): Update.
2898 * ipa-profile.c: New file.
2899 * ipa-utils.h (ipa_propagate_frequency): Declare.
2900 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
2901 data-streamer.h, value-prof.h.
2902 (symtab_remove_unreachable_nodes): Update profile.
2903 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
2904 account_time_size, cmp_counts, dump_histogram,
2905 ipa_profile_generate_summary, ipa_profile_write_summary,
2906 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
2907 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
2908 Move to ipa-profile.c.
2910 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
2912 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
2914 2013-09-01 Jan Hubicka <jh@suse.cz>
2916 * common.opt (fdevirtualize-speculatively): New function.
2917 * invoke.texi (fdevirtualize-speculatively): Document.
2918 * ipa-devirt.c: Include ipa-inline.h
2919 (likely_target_p): New function.
2920 (ipa_devirt): New function.
2921 (gate_ipa_devirt): New function.
2922 (pass_data_ipa_devirt): New static var.
2923 (pass_ipa_devirt): Likewise.
2924 (make_pass_ipa_devirt): New function.
2925 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
2926 (common_handle_option): Disable devirtualization when
2927 value range profiling is available.
2928 * passes.def (pass_ipa_devirt): Add.
2929 * timever.def (TV_IPA_DEVIRT): New timevar.
2930 * tree-pass.h (make_pass_ipa_devirt):
2932 2013-09-01 Iain Sandoe <iain@codesourcery.com>
2934 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
2935 include sanitize(undefined).
2937 2013-08-31 Diego Novillo <dnovillo@google.com>
2939 * Makefile.in (TREE_CORE_H): Define.
2941 (GTFILES): Add tree-core.h.
2942 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
2944 * tree-core.h: New file.
2945 Move all data structures, enum, typedefs, global
2946 declarations and constants from ...
2949 2013-08-31 Jan Hubicka <jh@suse.cz>
2951 * bulitins.c (expand_builtin): Do not early exit for gcov
2952 instrumented functions.
2954 2013-08-31 Marek Polacek <polacek@redhat.com>
2956 * ubsan.c: Include tm_p.h.
2958 2013-08-31 Jan Hubicka <jh@suse.cz>
2960 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
2963 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
2964 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
2965 * tree-cfg.c (verify_gimple_label): ... here.
2966 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
2967 (ipa_merge_profiles): New function.
2968 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
2969 (lto_input_function_body): Likewise.
2970 * ipa-utils.h (ipa_merge_profiles): Declare.
2971 * lto-streamer.h (lto_input_function_body): Update prototype.
2972 (emit_label_in_global_context_p): Remove.
2973 * lto-symtab.c: Include ipa-utils.h
2974 (lto_cgraph_replace_node): Use ipa_merge_profiles.
2976 2013-08-31 Jan Hubicka <jh@suse.cz>
2978 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
2980 2013-08-31 Jan Hubicka <jh@suse.cz>
2982 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
2984 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
2986 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
2987 "cmp" RTX before signed_comparison_operator check to account
2990 2013-08-30 Jan Hubicka <jh@suse.cz>
2992 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
2993 (detect_type_change_1): Rename to ...
2994 (detect_type_change): ... this one; early return on non-polymorphic
2996 (detect_type_change_ssa): Add comp_type parameter; update
2997 use of detect_type_change.
2998 (compute_complex_assign_jump_func): Add param_type parameter;
2999 update use of detect_type_change_ssa.
3000 (compute_complex_ancestor_jump_func): Likewise.
3001 (ipa_get_callee_param_type): New function.
3002 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3003 update calls to the jump function computation functions.
3005 2013-08-30 Teresa Johnson <tejohnson@google.com>
3006 Steven Bosscher <steven@gcc.gnu.org>
3008 * cfgrtl.c (fixup_new_cold_bb): New routine.
3009 (commit_edge_insertions): Invoke fixup_partitions.
3010 (find_partition_fixes): New routine.
3011 (fixup_partitions): Ditto.
3012 (verify_hot_cold_block_grouping): Update comments.
3013 (rtl_verify_edges): Invoke find_partition_fixes.
3014 (rtl_verify_bb_pointers): Update comments.
3015 (rtl_verify_bb_layout): Ditto.
3016 * basic-block.h (probably_never_executed_edge_p): Declare.
3017 (fixup_partitions): Ditto.
3018 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3019 * bb-reorder.c (sanitize_hot_paths): New function.
3020 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3022 * predict.c (probably_never_executed_edge_p): New routine.
3023 * cfg.c (check_bb_profile): Add partition insanity warnings.
3025 2013-08-30 Meador Inge <meadori@codesourcery.com>
3027 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3029 2013-08-30 Marek Polacek <polacek@redhat.com>
3031 * Makefile.in (ubsan.o): Add.
3032 (c-family/c-ubsan.o): Add.
3033 (builtins.o): Add ubsan.h dependency.
3034 * ubsan.h: New file.
3035 * ubsan.c: New file.
3036 * common.opt: Add -fsanitize=undefined option.
3037 (flag_sanitize): Add variable.
3038 (fsanitize=): Add option. Add Driver.
3039 (fsanitize=thread): Remove option.
3040 (fsanitize=address): Likewise.
3041 (static-libubsan): New option.
3042 * doc/invoke.texi: Document the new flag and -static-libubsan.
3043 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3044 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3045 * builtin-attrs.def (ATTR_COLD): Define.
3046 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3047 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3048 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3049 * flag-types.h (sanitize_code): New enum.
3050 * opts.c (common_handle_option): Parse command line arguments
3051 of -fsanitize=. Add -fsanitize=unreachable option.
3052 * varasm.c (get_variable_section): Adjust.
3053 (assemble_noswitch_variable): Likewise.
3054 (assemble_variable): Likewise.
3055 (output_constant_def_contents): Likewise.
3056 (categorize_decl_for_section): Likewise.
3057 (place_block_symbol): Likewise.
3058 (output_object_block): Likewise.
3059 * builtins.def: Likewise.
3060 * toplev.c (compile_file): Likewise.
3061 (process_options): Likewise.
3062 * cppbuiltin.c: Likewise.
3063 * tsan.c (tsan_pass): Likewise.
3064 (tsan_gate): Likewise.
3065 (tsan_gate_O0): Likewise.
3066 * cfgexpand.c (partition_stack_vars): Likewise.
3067 (expand_stack_vars): Likewise.
3068 (defer_stack_allocation): Likewise.
3069 (expand_used_vars): Likewise.
3070 * cfgcleanup.c (old_insns_match_p): Likewise.
3071 * asan.c (asan_finish_file): Likewise.
3072 (asan_instrument): Likewise.
3073 (gate_asan): Likewise.
3074 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
3075 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3076 (asan_global_struct): Use pointer_sized_int_node instead
3077 calling build_nonstandard_integer_type.
3078 (initialize_sanitizer_builtins): Likewise.
3079 (asan_finish_file): Likewise.
3080 * gcc.c: Document %{%:function(args):X}.
3081 (static_spec_functions): Add sanitize.
3082 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
3083 store funcval != NULL there.
3084 (do_spec_1): Adjust handle_spec_function caller.
3085 (handle_braces): Allow %:function(args) as condition.
3086 (sanitize_spec_function): New function.
3087 (ADD_STATIC_LIBUBSAN_LIBS): Define.
3088 (LIBUBSAN_SPEC): Likewise.
3089 (LIBUBSAN_EARLY_SPEC): Likewise.
3090 (SANITIZER_SPEC): Handle libubsan.
3091 (SANITIZER_EARLY_SPEC): Likewise.
3092 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
3093 instead of fsanitize=address.
3094 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
3095 instead of fsanitize=address*.
3096 * builtins.c: Include ubsan.h.
3097 (fold_builtin_0): Instrument __builtin_unreachable.
3098 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
3099 instead of flag_asan.
3100 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
3101 (pointer_sized_int_node): Define.
3102 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
3104 2013-08-30 Mike Stump <mikestump@comcast.net>
3106 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
3109 2013-08-29 Jan Hubicka <jh@suse.cz>
3111 * cgraph.c (cgraph_function_body_availability): Handle weakref
3113 * passes.def: Remove pass_fixup_cfg.
3114 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
3115 track when we need to remove functions.
3116 (gate_ipa_inline): Execute inlining always; add comment why.
3117 (pass_data_ipa_inline): Remove TODO_remove_functions.
3118 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
3119 do not produce summaries.
3120 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
3121 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
3122 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
3123 constant pool and vtable.
3125 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
3127 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
3128 New arm_neon.h's internal macros to specify 64-bit constants.
3129 Avoid using stdint.h's macros.
3131 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
3133 * recog.c (verify_changes): Verify that changes[i].old is non-zero
3134 before applying REG_P.
3136 2013-08-30 Jakub Jelinek <jakub@redhat.com>
3138 PR tree-optimization/58277
3139 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
3140 after seeing too many stmts with vdef in between dombb and current
3141 bb, invalidate everything.
3143 2013-08-30 Richard Biener <rguenther@suse.de>
3145 * fold-const.c (fold_single_bit_test): Fix overflow test.
3147 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
3149 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
3150 and which can live in a register, always retrieve the value on entry.
3151 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
3152 passed by invisible reference specially.
3153 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
3154 (vt_add_function_parameter): Correctly deal with a parameter passed by
3155 invisible reference.
3157 2013-08-30 Jan Hubicka <jh@suse.cz>
3159 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
3161 2013-08-30 Richard Biener <rguenther@suse.de>
3163 PR tree-optimization/58228
3164 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
3165 allow invariant loads in nested loop vectorization.
3167 2013-08-30 Richard Biener <rguenther@suse.de>
3169 PR tree-optimization/58223
3170 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
3171 (has_anti_or_output_dependence): ... this and adjust to also
3172 look for output dependences.
3173 (mark_nodes_having_upstream_mem_writes): Adjust.
3174 (rdg_flag_uses): Likewise.
3176 2013-08-30 Richard Biener <rguenther@suse.de>
3178 PR tree-optimization/58010
3179 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
3180 assert that we have a loop-closed PHI.
3182 2013-08-29 Jan Hubicka <jh@suse.cz>
3184 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
3185 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
3186 * lto-section-in.c (lto_free_function_in_decl_state): New function.
3187 (lto_free_function_in_decl_state_for_node): New function.
3189 2013-08-29 Xinliang David Li <davidxl@google.com>
3191 * loop-unroll.c (report_unroll_peel): Minor message change.
3192 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
3193 Emit alignment peeling message with default -fopt-info.
3194 (vect_loop_versioning): Emit loop version info message.
3195 * tree-vectorizer.c (vectorize_loops): Minor message change.
3196 (execute_vect_slp): Ditto.
3198 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
3200 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
3201 of the clone from the DECL_NAME of the original function.
3203 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
3205 * passes.c (register_pass): Add overload.
3206 * tree-pass.h (register_pass): Forward declare it. Add comment.
3208 2013-08-29 Jan Hubicka <jh@suse.cz>
3210 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
3211 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
3212 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
3213 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3214 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
3215 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
3216 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3217 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
3219 2013-08-29 Teresa Johnson <tejohnson@google.com>
3221 * dumpfile.c (dump_loc): Output column number.
3222 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
3223 * doc/invoke.texi: Document optall -fopt-info flag.
3224 * profile.c (read_profile_edge_counts): Use new dump framework.
3225 (compute_branch_probabilities): Ditto.
3226 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
3227 when pass not in any opt group.
3228 * pass_manager.h (pass_manager::get_pass_profile): New method.
3229 * value-prof.c (check_counter): Use new dump framework.
3230 (check_ic_target): Ditto.
3231 * coverage.c (get_coverage_counts): Ditto.
3232 (coverage_init): Setup new dump framework.
3234 2013-08-29 Richard Biener <rguenther@suse.de>
3236 PR tree-optimization/58246
3237 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
3238 handle the dominance check inside a basic-block.
3240 2013-08-29 Richard Biener <rguenther@suse.de>
3243 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
3244 of default defs that appear in abnormal PHI nodes.
3246 2013-08-29 Richard Biener <rguenther@suse.de>
3248 PR tree-optimization/57685
3249 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
3250 single-use operands to avoid exponential complexity.
3252 2013-08-28 Dehao Chen <dehao@google.com>
3254 * ipa-inline.c (edge_badness): Fix integer underflow.
3256 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
3258 * gtm-builtins.def (_ITM_free): Declare leaf.
3260 2013-08-28 Jakub Jelinek <jakub@redhat.com>
3263 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
3264 (*tls_local_dynamic_base_64_largepic): Likewise.
3265 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
3266 Remove predicate from call operand.
3267 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
3268 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
3270 2013-08-28 Jeff Law <law@redhat.com>
3272 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
3273 checks for the number of predecessors and successors allowed.
3274 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
3275 which require copying a joiner block if there is a request which
3276 is a subpath that requires no joiner block copying.
3278 2013-08-28 Jan Hubicka <jh@suse.cz>
3280 * lto-streamer-out.c (DFS_write_tree_body): Drop
3281 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
3282 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
3283 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
3284 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
3285 Do not read DECL_ERROR_ISSUED.
3286 (unpack_ts_decl_with_vis_value_fields): Do not read
3288 (lto_input_ts_binfo_tree_pointers): Do not read
3289 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
3290 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
3291 write DECL_ERROR_ISSUED..
3292 (pack_ts_decl_with_vis_value_fields): Do not write
3294 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
3295 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
3296 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
3297 * tree.h (tree_decl_common): Update comment.
3298 (DECL_ERROR_ISSUED): Remove.
3300 2013-08-28 Jakub Jelinek <jakub@redhat.com>
3303 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
3305 2013-08-28 Jan Hubicka <jh@suse.cz>
3307 * builtins.def (free): Declare leaf.
3309 2013-08-27 David Malcolm <dmalcolm@redhat.com>
3312 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
3313 * configure: Regenerate.
3315 2013-08-27 Martin Jambor <mjambor@suse.cz>
3317 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
3318 (ipa_ancestor_jf_data): Likewise.
3319 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
3320 (ipa_get_jf_pass_through_type_preserved): New function.
3321 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
3322 (ipa_get_jf_ancestor_type_preserved): New function.
3323 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
3324 (ipa_get_jf_ancestor_result): Likewise.
3325 (propagate_vals_accross_pass_through): Use
3326 ipa_get_jf_pass_through_result to do all the value mappings.
3327 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
3328 type_preserved flag.
3329 (ipa_set_jf_cst_copy): New function.
3330 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
3331 (ipa_set_jf_arith_pass_through): Likewise.
3332 (ipa_set_ancestor_jf): Likewise.
3333 (compute_complex_assign_jump_func): Set type_preserved instead of
3335 (ipa_compute_jump_functions_for_edge): Likewise.
3336 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
3337 (update_jump_functions_after_inlining): Update type_preserved.
3338 Explicitely create jump functions when combining one with pass_through.
3339 (ipa_write_jump_function): Stream the type_preserved flags.
3340 (ipa_read_jump_function): Likewise.
3342 2013-08-27 Jakub Jelinek <jakub@redhat.com>
3343 Aldy Hernandez <aldyh@redhat.com>
3345 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
3346 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
3347 * function.h (struct function): Add has_force_vect_loops and
3349 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
3350 * gimple.c (gimple_build_omp_critical): Add KIND argument and
3352 * gimple.def: Update CLAUSES comments.
3353 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
3354 (gimple_build_omp_for): Add argument to prototype.
3355 (gimple_omp_for_kind): New.
3356 (gimple_omp_for_set_kind): New.
3357 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
3358 GOVD_DATA_SHARE_CLASS.
3359 (enum omp_region_type): Add ORT_SIMD.
3360 (gimple_add_tmp_var): Handle ORT_SIMD.
3361 (gimplify_var_or_parm_decl): Same.
3362 (is_gimple_stmt): Same.
3363 (omp_firstprivatize_variable): Same.
3364 (omp_add_variable): Only use splay_tree_insert if lookup failed.
3365 (omp_notice_variable): Handle ORT_SIMD.
3366 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
3367 (omp_check_private): Handle ORT_SIMD.
3368 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
3370 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
3371 Handle OMP_CLAUSE_LASTPRIVATE.
3372 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
3374 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
3375 (gimplify_expr): Handle OMP_SIMD.
3376 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
3377 (expand_GOMP_SIMD_VF): New.
3378 (expand_GOMP_SIMD_LAST_LANE): New.
3379 * internal-fn.def (GOMP_SIMD_LANE): New.
3380 (GOMP_SIMD_VF): New.
3381 (GOMP_SIMD_LAST_LANE): New.
3382 * omp-low.c: Include target.h.
3383 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
3385 (check_omp_nesting_restrictions): Same.
3387 (lower_rec_simd_input_clauses): New.
3388 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
3390 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
3391 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
3392 (expand_omp_build_assign): New.
3393 (expand_omp_for_init_counts): New.
3394 (expand_omp_for_init_vars): New.
3395 (extract_omp_for_update_vars): New.
3396 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
3397 and rewrite accordingly.
3398 (expand_omp_simd): New.
3399 (expand_omp_for): Use expand_omp_simd.
3400 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
3401 (lower_omp_for): Do not lower the body.
3402 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
3404 * tree-flow.h (find_omp_clause): Remove prototype.
3405 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
3406 forcing vectorization of the loop, or if flag_tree_vectorize.
3407 (gate_tree_if_conversion): Similarly.
3408 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
3409 gimple_build_omp_for.
3410 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
3411 * tree-parloops (create_parallel_loop): Pass kind argument to
3412 gimple_build_omp_for.
3413 * tree-pretty-print.c (dump_omp_clause): Add cases for
3414 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
3415 OMP_CLAUSE__SIMDUID_.
3416 (dump_generic_node): Handle OMP_SIMD.
3417 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
3418 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
3419 unroll OMP_SIMD loops here.
3420 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
3421 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
3423 (vect_analyze_data_refs): Handle simd loops.
3424 * tree-vect-loop.c (vectorizable_live_operation): Handle
3426 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
3427 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
3428 (vectorizable_load): Same.
3429 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
3430 (struct simduid_to_vf): New.
3431 (simduid_to_vf::hash): New.
3432 (simduid_to-vf::equal): New.
3433 (struct simd_array_to_simduid): New.
3434 (simd_array_to_simduid::hash): New.
3435 (simd_array_to_simduid::equal): New.
3436 (adjust_simduid_builtins): New.
3437 (struct note_simd_array_uses_struct): New.
3438 (note_simd_array_uses_cb): New.
3439 (note_simd_array_uses): New.
3440 (vectorize_loops): Handle simd hints and adjust simd builtins
3442 * tree-vectorizer.h (struct _stmt_vec_info): Add
3443 simd_lane_access_p field.
3444 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
3445 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
3446 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
3447 (omp_clause_code_name): Same.
3448 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
3449 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
3450 * tree.def (OMP_SIMD): New entry.
3451 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
3452 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
3453 (OMP_CLAUSE_DECL): Adjust range for new clauses.
3454 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
3455 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
3456 (OMP_CLAUSE_LINEAR_STEP): New.
3457 (OMP_CLAUSE_SAFELEN_EXPR): New.
3458 (OMP_CLAUSE__SIMDUID__DECL): New.
3459 (find_omp_clause): New prototype.
3461 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
3463 * config/i386/driver-i386.c (host_detect_local_cpu): Update
3464 Haswell processor detection.
3466 2013-08-27 Christian Widmer <shadow@umbrox.de>
3469 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
3470 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
3471 AVX2 capable processors.
3473 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
3475 * config/aarch64/arm_neon.h: Replace all inline asm implementations
3476 of vget_low_* with implementations in terms of other intrinsics.
3478 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
3481 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
3482 values to -1, 0 and 1.
3484 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
3486 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
3487 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
3488 (<optab><mode>3_insn): Remove.
3489 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
3490 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
3491 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
3492 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
3493 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
3494 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
3495 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
3496 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
3497 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
3498 (ror<mode>3_insn): Likewise.
3499 * config/aarch64/predicates.md (aarch64_simd_register): New.
3501 2013-08-27 Richard Biener <rguenther@suse.de>
3503 PR tree-optimization/57521
3504 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
3505 one edge is non-critical.
3506 (find_phi_replacement_condition): Make sure to use a non-critical
3507 edge. Cleanup and remove old bug workarounds.
3508 (bb_postdominates_preds): Remove.
3509 (if_convertible_loop_p_1): Do not compute post-dominators.
3510 (combine_blocks): Do not free post-dominators.
3511 (main_tree_if_conversion): Likewise.
3512 (pass_data_if_conversion): Add TODO_verify_ssa.
3514 2013-08-27 DJ Delorie <dj@redhat.com>
3516 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
3518 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
3520 * function.c (assign_parm_find_data_types): Set passed_mode and
3521 nominal_mode to the TYPE_MODE of nominal_type for the built
3522 pointer type in case of the struct-pass-by-reference.
3524 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3526 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
3527 (UINT16_TYPE): Change default to "unsigned int".
3529 * config/avr/avr.opt (mfract-convert-truncate): New option.
3530 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
3531 is set, round negative fractional integers according to n1169
3532 when converting to integer types.
3534 2013-08-26 Jan Hubicka <jh@suse.cz>
3536 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
3537 methods can not be called indirectly when their address is not taken.
3539 2013-08-26 Jan Hubicka <jh@suse.cz>
3541 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
3544 2013-08-26 Jan Hubicka <jh@suse.cz>
3546 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
3549 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3551 * reload.c (find_valid_class): Allow classes that do not include
3552 FIRST_PSEUDO_REGISTER - 1.
3554 2013-08-26 Jan Hubicka <jh@suse.cz>
3556 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
3557 fix edge count/frequency when speculation failed; fix type check
3558 for the direct call.
3560 2013-08-26 Jan Hubicka <jh@suse.cz>
3562 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
3564 2013-08-26 Jan Hubicka <jh@suse.cz>
3566 * ipa-inline-transform.c (inline_transform): Be ready for basic block
3567 to be changed by edge redirection.
3569 2013-08-26 Jan Hubicka <jh@suse.cz>
3571 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
3572 formating; add sanity check.
3573 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
3574 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
3575 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
3576 (resolve_noninline_speculation): Update callee keys, too.
3578 2013-08-26 Jan Hubicka <jh@suse.cz>
3580 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
3581 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
3583 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3585 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
3588 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
3590 * config/i386/i386.c (ix86_debug_options): Remove prototype.
3591 (x86_64_elf_select_section): Ditto.
3592 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
3594 (ix86_pass_by_reference): Ditto.
3595 (output_set_got): Ditto.
3596 (ix86_unary_operator_ok): Ditto.
3597 (ix86_expand_builtin): Ditto.
3599 2013-08-23 Jan Hubicka <jh@suse.cz>
3601 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
3603 2013-08-23 Jan Hubicka <jh@suse.cz>
3605 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
3606 (tree_decl_with_vis): Add FINAL field.
3608 2013-08-23 Jeff Law <law@redhat.com>
3610 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
3611 the debugging dump when the expression is fully redundant.
3613 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
3615 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
3616 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
3617 * pretty-print.c (pp_formatted_text_data): Likewise.
3618 (pp_write_text_to_stream): Likewise.
3619 (pp_write_text_as_dot_label_to_stream): Likewise.
3620 (pp_append_r): Likewise.
3621 (pp_format): Likewise.
3622 (pp_flush): Likewise.
3623 (pp_clear_output_area): Likewise.
3624 (pp_append_text): Likewise.
3625 (pp_formatted_text): Likewise.
3626 (pp_remaining_character_count_for_line): Likewise.
3627 (pp_newline): Likewise.
3628 (pp_character): Likewise.
3629 (output_buffer::~output_buffer): Define.
3630 (pretty_printer::~pretty_printer): Destruct output buffer.
3631 * pretty-print.h (output_buffer::~output_buffer): Declare.
3632 (pretty_printer::~pretty_printer): Declare virtual.
3634 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
3637 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
3638 HOST_WIDE_INT_M1U): New macros.
3639 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
3640 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
3641 unsigned -1 for lshift.
3642 * cse.c (cse_insn): Likewise.
3643 * double-int.c (rshift_double, lshift_double): Likewise.
3644 * builtins.c (fold_builtin_bitop): Likewise.
3645 * combine.c (force_to_mode): Likewise.
3646 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
3647 * simplify-rtx.c (simplify_const_unary_operation,
3648 simplify_const_binary_operation): Likewise.
3649 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
3650 check_va_list_escapes): Likewise.
3651 * rtlanal.c (nonzero_bits1): Likewise.
3652 * expmed.c (expand_smod_pow2): Likewise.
3653 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
3655 2013-08-23 Jan Hubicka <jh@suse.cz>
3657 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
3658 as having address taken.
3660 2013-08-23 Jan Hubicka <jh@suse.cz>
3662 * ipa-utils.h (method_class_type): Declare.
3663 * ipa-devirt.c (method_class_type): Export.
3665 * cgraphunit.c (analyze_functions): Do basic devirtualization;
3666 do not walk base classes of anonymous types.
3668 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
3670 PR rtl-optimization/58220
3672 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
3673 handle SEQUENCE insns properly.
3675 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
3677 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
3679 (pp_newline_and_indent): Likewise.
3680 (pp_separate_with): Likewise.
3681 * pretty-print.c (pp_newline_and_flush): Define.
3682 (pp_newline_and_indent): Likewise.
3683 (pp_separate_with): Likewise.
3685 2013-08-23 Jakub Jelinek <jakub@redhat.com>
3688 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
3689 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
3691 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
3693 * config/i386/predicates.md (ext_sse_reg_operand): New.
3694 * config/i386/i386.md (*movti_internal): Use
3695 predicate to determine if EVEX is needed.
3696 (*movsi_internal): Ditto.
3697 (*movdf_internal): Ditto.
3698 (*movsf_internal): Ditto.
3699 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
3701 2013-08-23 Jakub Jelinek <jakub@redhat.com>
3703 PR tree-optimization/58209
3704 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
3705 (find_tail_calls): Give up for pointer result types if m is non-NULL.
3706 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
3707 emit POINTER_PLUS_EXPR.
3708 (create_tailcall_accumulator): For pointer result type accumulate in
3711 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
3713 * configure.ac: Add backslashes missing from the last change.
3714 * configure: Regenerate.
3716 2013-08-22 Jan Hubicka <jh@susue.cz>
3718 * ipa.c (function_and_variable_visibility): First remember function
3719 was global and then make it local.
3721 2013-08-22 Julian Brown <julian@codesourcery.com>
3723 * configure.ac: Add aarch64 to list of arches which use "nop" in
3725 * configure: Regenerate.
3727 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3729 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
3730 gnu_libc_has_function.
3731 * config/s390/tpf.h: Likewise.
3733 2013-08-22 Jan Hubicka <jh@susue.cz>
3735 * timevar.c (validate_phases): Add cast.
3737 2013-08-22 Jan Hubicka <jh@susue.cz>
3739 * timevar.c (validate_phases): Use size_t for memory.
3740 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
3742 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
3744 * pretty-print.h (output_buffer::output_buffer): Declare.
3745 (pretty_printer::pretty_printer): Likewise.
3746 (pp_construct): Remove.
3747 * pretty-print.c (output_buffer::output_buffer): Define.
3748 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
3749 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
3750 (print_gimple_expr): Likewise.
3751 (print_gimple_seq): Likewise.
3752 (gimple_dump_bb): Likewise.
3753 * sched-vis.c (dump_value_slim): Likewise.
3754 (dump_insn_slim): Likewise.
3755 (dump_rtl_slim): Likewise.
3756 (str_pattern_slim): Likewise.
3757 * tree-mudflap.c (mf_varname_tree): Likewise.
3758 * graph.c (print_graph_cfg): Likewise.
3759 (start_graph_dump): Likewise.
3760 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
3762 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
3763 pretty printer initialization.
3764 * coretypes.h (diagnostic_context): Remove superflous type alias
3766 (pretty_printer): Likewise. Declare directly as a class.
3767 (pretty_print_info): Remove declaration as class.
3768 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
3769 and pp_clear_output_area.
3770 (asan_add_global): Likewise.
3772 2013-08-22 Jan Hubicka <jh@suse.cz>
3774 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
3775 * ipa-utils.h (update_type_inheritance_graph): Declare.
3776 (possible_polymorphic_call_target_p): Declare.
3777 (possible_polymorphic_call_target_p): New.
3778 * ipa-devirt.c: Update toplevel comments.
3779 (cached_polymorphic_call_targets): Move up.
3780 (odr_type_d): Move ID down.
3781 (polymorphic_type_binfo_p): Update comment.
3782 (odr_hasher::remove): Likewise;
3783 (get_odr_type): Set anonymous_namespace.
3784 (dump_odr_type): Dump it.
3785 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
3786 (maybe_record_node): Record node in cached_polymorphic_call_targets.
3787 (record_binfo): Add comment.
3788 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
3790 (devirt_node_removal_hook): Do not iCE when cache is freed.
3791 (possible_polymorphic_call_target_p): New predicate.
3792 (update_type_inheritance_graph): New function.
3794 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
3795 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3796 Sergey Lega <sergey.s.lega@intel.com>
3797 Anna Tikhonova <anna.tikhonova@intel.com>
3798 Ilya Tocar <ilya.tocar@intel.com>
3799 Andrey Turetskiy <andrey.turetskiy@intel.com>
3800 Ilya Verbin <ilya.verbin@intel.com>
3801 Kirill Yukhin <kirill.yukhin@intel.com>
3802 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3804 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
3805 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
3806 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
3807 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
3808 (OPTION_MASK_ISA_AVX2_UNSET): Update.
3809 (OPTION_MASK_ISA_AVX512F_UNSET): New.
3810 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
3811 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
3812 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
3813 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
3814 OPT_mavx512pf, OPT_mavx512er cases.
3815 * config/i386/constraints.md (v): New constraint.
3816 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
3817 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
3818 (bit_AVX512CD): New.
3819 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
3820 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
3821 * config/i386/i386-c.c (ix86_target_macros_internal):
3822 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
3824 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
3825 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
3826 * config/i386/i386.c (regclass_map, dbx_register_map)
3827 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
3828 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
3829 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
3831 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
3832 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
3833 -mavx512pf options. Fix formatting.
3834 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
3835 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
3836 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
3837 -mavx512cd, -mavx512pf options.
3838 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
3839 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
3840 (ix86_preferred_output_reload_class): Replace SSE_REGS with
3842 (ix86_hard_regno_mode_ok): Support 512-bit registers.
3843 (ix86_set_reg_reg_cost): Ditto.
3844 (x86_order_regs_for_local_alloc): Ditto.
3845 (MAX_VECT_LEN): Extend to 64-byte.
3846 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
3847 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
3848 (TARGET_AVX512ER, TARGET_AVX512CD): New.
3849 (BIGGEST_ALIGNMENT): Extend to 512-bits.
3850 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
3851 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
3852 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
3853 (SSE_REG_MODE_P): Support new modes.
3854 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
3855 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
3856 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
3857 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
3858 (REG_CLASS_CONTENTS): Add new registers.
3859 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
3860 (EXT_REX_SSE_REGNO_P): New.
3861 (HI_REGISTER_NAMES): Add new registers.
3862 * config/i386/i386.md: Define constants for new registers.
3863 (mode): Add new 512-bit modes.
3864 (prefix): Support evex prefix.
3865 (isa): Support avx512f, noavx512f, fma_avx512f.
3866 (ssemodesuffix): Add new 512-bit modes.
3868 (*movxi_internal_avx512f): Ditto.
3869 (*movdi_internal): Replace constraint "x" with the new constraint "v".
3871 (*movsi_internal): Likewise.
3872 (*movdf_internal): Likewise.
3873 (*movsf_internal): Likewise.
3874 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
3875 (<code><mode>3): Likewise.
3876 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
3878 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
3879 with the new constraint "v".
3880 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
3882 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
3883 with the new constraint "v".
3884 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
3885 (<sse2>_storedqu<avxsizesuffix>): Likewise.
3886 (*<plusminus_insn><mode>3): Likewise.
3887 (<sse>_vm<plusminus_insn><mode>3): Likewise.
3888 (*mul<mode>3): Likewise.
3889 (<sse>_vmmul<mode>3): Likewise.
3890 (<sse>_div<mode>3): Likewise.
3891 (<sse>_vmdiv<mode>3): Likewise.
3892 (<sse>_sqrt<mode>2): Likewise.
3893 (<sse>_vmsqrt<mode>2): Likewise.
3894 (*<code><mode>3_finite): Likewise.
3895 (*<code><mode>3) <smaxmin>: Likewise.
3896 (<sse>_vm<code><mode>3): Likewise.
3897 (*<code><mode>3) <any_logic>: Likewise.
3898 (*fma_fmadd_<mode>): Likewise.
3899 (*fma_fmsub_<mode>): Likewise.
3900 (*fma_fnmadd_<mode>): Likewise.
3901 (*fma_fnmsub_<mode>): Likewise.
3902 (*fma_fmaddsub_<mode>): Likewise.
3903 (*fma_fmsubadd_<mode>): Likewise.
3904 (*fmai_fmadd_<mode>): Likewise.
3905 (*fmai_fmsub_<mode>): Likewise.
3906 (*fmai_fnmadd_<mode>): Likewise.
3907 (*fmai_fnmsub_<mode>): Likewise.
3908 (sse_cvtsi2ss): Likewise.
3909 (sse_cvtsi2ssq): Likewise.
3910 (sse_cvtss2si): Likewise.
3911 (sse_cvtss2si_2): Likewise.
3912 (sse_cvtss2siq): Likewise.
3913 (sse_cvtss2siq_2): Likewise.
3914 (sse_cvttss2si): Likewise.
3915 (sse_cvtss2siq_2): Likewise.
3916 (float<sseintvecmodelower><mode>2): Likewise.
3917 (sse2_cvtsd2si_2): Likewise.
3918 (sse2_cvtsd2siq_2): Likewise.
3919 (*<plusminus_insn><mode>3): Likewise.
3920 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
3921 (*<sse4_1_avx2>_mul<mode>3): Likewise.
3922 (ashr<mode>3): Likewise.
3923 (<shift_insn><mode>3): Likewise.
3924 (avx2_<code><mode>3): Likewise.
3925 (*avx2_<code><mode>3): Likewise.
3926 (*andnot<mode>3): Likewise.
3927 (*<code><mode>3) <any_logic>: Likewise.
3928 (abs<mode>2): Likewise.
3929 (avx2_permvar<mode>): Likewise.
3930 (avx2_perm<mode>_1): Likewise.
3931 (*avx_vpermilp<mode>): Likewise.
3932 (avx_vpermilvar<mode>3): Likewise.
3933 (avx2_ashrv<mode>): Likewise.
3934 (avx2_<shift_insn>v<mode>): Likewise.
3935 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
3937 * doc/rtl.texi: Document XImode.
3939 2013-08-21 Jeff Law <law@redhat.com>
3941 * tree-flow.h (register_jump_thread): Pass vector of edges
3942 instead of each important edge.
3943 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
3944 thread path into a vector and pass that to register_jump_thread.
3945 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
3946 passed in edge vector to the current 3-edge form.
3949 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3951 * dce.c (fini_dce): Call df_analyze again just in case
3952 delete_unmarked_insns removed anything.
3954 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
3956 * reload.h (struct reg_equivs): Rename to ..
3957 (struct reg_equivs_s): .. this.
3959 2013-08-20 Martin Liska <marxin.liska@gmail.com>
3961 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
3963 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3965 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
3967 2013-08-21 Jeff Law <law@redhat.com>
3969 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
3970 simplify assignments too. If the RHS collapses to a singleton
3971 range, then return the value for the range.
3973 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
3975 * config/i386/sse.md (V16): Rename to...
3977 (mov<mode>): Update iterator name.
3978 (*mov<mode>_internal): Ditto.
3979 (push<mode>1): Ditto.
3980 (movmisalign<mode>): Ditto.
3982 2013-08-20 Jan Hubicka <jh@suse.cz>
3985 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
3986 entry for direct edges.
3987 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
3989 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3991 Revert my last two changes, r201865 and r201864:
3994 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3996 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3997 instances can own GC refs.
3999 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4000 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4001 (gcc::context::gt_pch_nx): Likewise.
4002 (gcc::context::gt_pch_nx): Likewise.
4003 * ggc.h (gt_ggc_mx <T>): New.
4004 (gt_pch_nx_with_op <T>): New.
4005 (gt_pch_nx <T>): New.
4006 * passes.c (opt_pass::gt_ggc_mx): New.
4007 (opt_pass::gt_pch_nx): New.
4008 (opt_pass::gt_pch_nx_with_op): New.
4009 (pass_manager::gt_ggc_mx): New.
4010 (pass_manager::gt_pch_nx): New.
4011 (pass_manager::gt_pch_nx_with_op): New.
4012 (pass_manager::operator new): Use
4013 ggc_internal_cleared_alloc_stat rather than xcalloc.
4014 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4015 (pass_manager::gt_ggc_mx): New.
4016 (pass_manager::gt_pch_nx): New.
4017 (pass_manager::gt_pch_nx_with_op): New.
4018 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4019 (opt_pass::operator new): New.
4020 (opt_pass::gt_ggc_mx): New.
4021 (opt_pass::gt_pch_nx): New.
4022 (opt_pass::gt_pch_nx_with_op): New.
4025 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4027 * Makefile.in (GTFILES): Add context.h.
4028 * context.c (gcc::context::operator new): New.
4029 (gcc::context::gt_ggc_mx): New.
4030 (gcc::context::gt_pch_nx): New.
4031 (gcc::context::gt_pch_nx): New.
4032 * context.h (gcc::context): Add GTY((user)) marking.
4033 (gcc::context::operator new): New.
4034 (gcc::context::gt_ggc_mx): New.
4035 (gcc::context::gt_pch_nx): New.
4036 (gcc::context::gt_pch_nx): New.
4037 (g): Add GTY marking.
4038 (gt_ggc_mx (gcc::context *)): New.
4039 (gt_pch_nx (gcc::context *)): New.
4040 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4041 void *cookie)): New.
4042 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4044 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4046 * dce.c (fini_dce): Call df_analyze again just in case
4047 delete_unmarked_insns removed anything.
4049 2013-08-20 Teresa Johnson <tejohnson@google.com>
4051 PR rtl-optimizations/57451
4052 * final.c (reemit_insn_block_notes): Prevent lexical blocks
4053 from crossing split section boundaries.
4055 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
4057 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
4058 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
4059 with MULTLIB_DEFAULTS.
4061 2013-08-20 Nick Clifton <nickc@redhat.com>
4063 * target.def (narrow_volatile_bitfield): Note that the default
4064 value is false, not !TARGET_STRICT_ALIGN.
4065 * doc/tm.texi: Regenerate.
4067 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
4069 Fix LIB_SPEC for systems without libpthread.
4071 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
4072 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
4074 * config/i386/linux-common.h: Likewise.
4075 * config/mips/linux-common.h: Likewise.
4077 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
4079 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
4082 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4084 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4085 instances can own GC refs.
4087 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4088 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4089 (gcc::context::gt_pch_nx): Likewise.
4090 (gcc::context::gt_pch_nx): Likewise.
4091 * ggc.h (gt_ggc_mx <T>): New.
4092 (gt_pch_nx_with_op <T>): New.
4093 (gt_pch_nx <T>): New.
4094 * passes.c (opt_pass::gt_ggc_mx): New.
4095 (opt_pass::gt_pch_nx): New.
4096 (opt_pass::gt_pch_nx_with_op): New.
4097 (pass_manager::gt_ggc_mx): New.
4098 (pass_manager::gt_pch_nx): New.
4099 (pass_manager::gt_pch_nx_with_op): New.
4100 (pass_manager::operator new): Use
4101 ggc_internal_cleared_alloc_stat rather than xcalloc.
4102 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4103 (pass_manager::gt_ggc_mx): New.
4104 (pass_manager::gt_pch_nx): New.
4105 (pass_manager::gt_pch_nx_with_op): New.
4106 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4107 (opt_pass::operator new): New.
4108 (opt_pass::gt_ggc_mx): New.
4109 (opt_pass::gt_pch_nx): New.
4110 (opt_pass::gt_pch_nx_with_op): New.
4112 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4114 * Makefile.in (GTFILES): Add context.h.
4115 * context.c (gcc::context::operator new): New.
4116 (gcc::context::gt_ggc_mx): New.
4117 (gcc::context::gt_pch_nx): New.
4118 (gcc::context::gt_pch_nx): New.
4119 * context.h (gcc::context): Add GTY((user)) marking.
4120 (gcc::context::operator new): New.
4121 (gcc::context::gt_ggc_mx): New.
4122 (gcc::context::gt_pch_nx): New.
4123 (gcc::context::gt_pch_nx): New.
4124 (g): Add GTY marking.
4125 (gt_ggc_mx (gcc::context *)): New.
4126 (gt_pch_nx (gcc::context *)): New.
4127 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4128 void *cookie)): New.
4129 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4131 2013-08-20 Alan Modra <amodra@gmail.com>
4134 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
4135 (rs6000_emit_epilogue): Likewise.
4137 2013-08-19 Dehao Chen <dehao@google.com>
4139 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
4141 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
4142 Jakub Jelinek <jakub@redhat.com>
4144 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
4145 (BUILT_IN_FABSD64): Likewise.
4146 (BUILT_IN_FABSD128): Likewise.
4147 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
4148 (fold_builtin_1): Likewise.
4149 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
4150 destination and source operands.
4151 (*abstd2_fpr): Likewise.
4152 (*nabstd2_fpr): Likewise.
4154 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
4156 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
4159 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
4161 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
4164 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
4166 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
4167 * builtins.c (default_libc_has_function): New.
4168 (gnu_libc_has_function): Ditto.
4169 (no_c99_libc_has_function): Ditto.
4170 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
4171 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
4172 (fold_builtin_sincos): Likewise.
4173 (fold_builtin_cexp): Likewise.
4174 * builtins.def (DEF_C94_BUILTIN): Likewise.
4175 (DEF_C99_BUILTIN): Likewise.
4176 (DEF_C99_C90RES_BUILTIN): Likewise.
4177 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
4178 definitions to using this define.
4179 * config/darwin-protos.h (darwin_libc_has_function): New.
4180 * config/darwin.c (darwin_libc_has_function): Ditto.
4181 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
4182 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
4183 * config/darwin.h: Ditto.
4184 * config/elfos.h: Ditto.
4185 * config/freebsd.h: Ditto.
4186 * config/i386/cygming.h: Ditto.
4187 * config/i386/djgpp.h: Ditto.
4188 * config/i386/i386-interix.h: Ditto.
4189 * config/microblaze/microblaze.h: Ditto.
4190 * config/mmix/mmix.h: Ditto.
4191 * config/gnu-user.h: Ditto.
4192 * config/ia64/hpux.h: Ditto.
4193 * config/pa/pa-hpux.h: Ditto.
4194 * config/pdp11/pdp11.h: Ditto.
4195 * config/picochip/picochip.h: Ditto.
4196 * config/linux.h: Ditto.
4197 * config/netbsd.h: Ditto.
4198 * config/openbsd.h: Ditto.
4199 * config/rs6000/aix43.h: Ditto.
4200 * config/rs6000/aix51.h: Ditto.
4201 * config/rs6000/aix52.h: Ditto.
4202 * config/rs6000/aix53.h: Ditto.
4203 * config/rs6000/aix61.h: Ditto.
4204 * config/rs6000/darwin.h: Ditto.
4205 * config/rs6000/linux.h: Ditto.
4206 * config/rs6000/linux64.h: Ditto.
4207 * config/s390/tpf.h: Ditto.
4208 * config/sol2-10.h: Ditto.
4209 * config/sol2.h: Ditto.
4210 * config/vms/vms.h: Ditto.
4211 * config/vxworks.h: Ditto.
4212 * config/linux-android.c (linux_android_libc_has_function):
4213 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
4214 * config/linux-protos.h (linux_android_libc_has_function):
4216 * config/i386/i386.c (ix86_libc_has_function): New.
4217 * config/i386/i386-protos.h
4218 (ix86_libc_has_function): New declaration.
4219 * config/i386/i386.md
4220 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
4221 ("isinf<mode>2): Likewise.
4222 * convert.c (convert_to_integer): Using new target hook
4223 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
4224 TARGET_C99_FUNCTIONS.
4225 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
4226 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
4227 * coretypes.h (function_class): New enum for different
4228 classes of functions.
4229 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
4230 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
4231 (TARGET_HAS_SINCOS): Likewise.
4232 (TARGET_LIBC_HAS_FUNCTION): New.
4233 * doc/tm.texi: Regenerated.
4234 * targhooks.h (default_libc_has_function): New declaration.
4235 (no_c99_libc_has_function): Ditto.
4236 (gnu_libc_has_function): Ditto.
4237 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
4238 and TARGET_HAS_SINCOS.
4240 2013-08-18 Jan Hubicka <jh@suse.cz>
4242 * Makeifle-in (ipa-devirt.o): New.
4243 (GTFILES): Add ipa-utils.h and ipa-devirt.c
4244 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
4245 (analyze_functions): Look into possible targets of polymorphic call.
4246 * dumpfile.c (dump_files): Add type-inheritance dump.
4247 * dumpfile.h (TDI_inheritance): New.
4248 * ipa-devirt.c: New file.
4249 * ipa-utils.h (odr_type_d): Forward declare.
4250 (odr_type): New type.
4251 (build_type_inheritance_graph): Declare.
4252 (possible_polymorphic_call_targets): Declare and introduce inline
4253 variant when only edge is pased.
4254 (dump_possible_polymorphic_call_targets): Likewise.
4255 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
4256 * tree.c (type_in_anonymous_namespace_p): Break out from ...
4257 (types_same_for_odr): ... here.
4258 * tree.h (type_in_anonymous_namespace_p): Declare.
4260 2013-08-18 Jakub Jelinek <jakub@redhat.com>
4262 PR tree-optimization/58006
4263 * tree-parloops.c (take_address_of): Don't ICE if get_name
4265 (eliminate_local_variables_stmt): Remove clobber stmts.
4267 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
4269 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
4270 error message is issued for an alias to undefined symbol.
4272 2013-08-18 Jan Hubicka <jh@suse.cz>
4274 * cgraph.c (cgraph_create_indirect_edge): Discover
4275 polymorphic calls and record basic info into indirect_info.
4276 * gimple-fold.c (gimple_fold_call): When doing BINFO based
4277 devirtualization, ignore objc function calls.
4278 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
4279 call with no parm index info.
4280 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
4281 * tree.c (virtual_method_call_p): New function.
4282 * tree.h (virtual_method_call_p): Declare.
4284 2013-08-16 Jan Hubicka <jh@suse.cz>
4287 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
4289 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
4291 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
4294 2013-08-16 David Malcolm <dmalcolm@redhat.com>
4296 * gengtype.c (type_for_name): Add special-case support for
4297 locating types within the "gcc::" namespace.
4298 (open_base_files): Emit a "using namespace gcc" directive.
4300 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
4303 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
4304 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
4306 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
4307 array instead of each individual operand as a separate argument.
4308 (emit_fusion_gpr_load): Likewise.
4309 (expand_fusion_gpr_load): Add new function declaration.
4311 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
4312 signature to have the operands passed as an array, instead of as
4313 separate arguments. Allow ZERO_EXTEND to be in the memory
4314 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
4315 depend on the register live/dead flags when peepholes are run.
4316 (expand_fusion_gpr_load): New function to be called from the
4317 peephole2 pass, to change the register that addis sets to be the
4319 (emit_fusion_gpr_load): Change the calling signature to have the
4320 operands passed as an array, instead of as separate arguments.
4321 Allow ZERO_EXTEND to be in the memory address, and also
4322 SIGN_EXTEND if -mpower8-fusion-sign.
4324 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
4326 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
4327 adjust the register addis loads up in the peephole2 pass. Do not
4328 depend on the register live/dead state when the peephole pass is done.
4330 2013-08-16 David Malcolm <dmalcolm@redhat.com>
4332 * gengtype.c (create_user_defined_type): Ensure that the kind
4333 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
4334 declaration is seen before the GTY((user)) marking.
4336 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
4339 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
4341 2013-08-16 Jan Hubicka <jh@suse.cz>
4343 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
4344 arugment expected_type.
4345 (gimple_fold_call): Use it.
4346 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
4347 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
4348 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
4349 (try_make_edge_direct_virtual_call): Likewise.
4350 * tree.c (obj_type_ref_class): New.
4351 * tree.h (obj_type_ref_class): Use it.
4353 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
4355 * sched-vis.c (rtl_slim_pp_initialized): Remove.
4356 (rtl_slim_pp): Likewise.
4357 (init_rtl_slim_pretty_print): Likewise.
4358 (dump_value_slim): Don't call it. Use local pretty printer.
4359 (dump_insn_slim): Likewise.
4360 (dump_rtl_slim): Likewise.
4361 (str_pattern_slim): Likewise.
4362 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
4365 2013-08-16 Jakub Jelinek <jakub@redhat.com>
4367 PR tree-optimization/58164
4368 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
4369 walk gimple_goto_dest of GIMPLE_GOTO.
4371 PR tree-optimization/58165
4372 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
4373 bi_call must be the last stmt in a bb, don't split_block, instead
4374 use fallthru edge from it and give up if there is none.
4375 Release conds vector when returning early.
4377 2013-08-14 Xinliang David Li <davidxl@google.com>
4379 * config/i386/i386.c (ix86_option_override_internal):
4380 Remove unused variable and field.
4382 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4385 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
4386 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
4387 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
4388 and Linux, correct BLKmode alignment when 128-bit alignment is
4389 required and compatibility flag is not set.
4390 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
4391 for zero-size arguments when compatibility flag is not set.
4393 2013-08-14 Jakub Jelinek <jakub@redhat.com>
4395 PR tree-optimization/58145
4396 * tree-sra.c (build_ref_for_offset): If prev_base has
4397 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
4399 2013-08-14 Xinliang David Li <davidxl@google.com>
4401 * config/i386/i386.c (ix86_option_override_internal):
4402 Fix uninitialized variable error.
4404 2013-08-14 Xinliang David Li <davidxl@google.com>
4406 * config/i386/i386.opt: Define two new options.
4407 * config/i386/x86-tune.def: Add arch selector field in macros.
4408 * config/i386/i386.h: Adjust macro definition.
4409 * config/i386/i386.c (ix86_option_override_internal):
4411 (parse_mtune_ctrl_str): New function.
4412 (set_ix86_tune_features): New function.
4413 (ix86_function_specific_restore): Call the new helper function.
4415 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
4417 PR rtl-optimization/57662
4418 * sel-sched.c (code_motion_process_successors): When the current insn
4419 is removed after the recursive traversal, break from the loop.
4420 Add comments and debug printouts.
4422 2013-08-14 Jakub Jelinek <jakub@redhat.com>
4423 Alexandre Oliva <aoliva@redhat.com>
4426 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
4427 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
4428 there also UNSPEC_PLTOFF.
4430 2013-08-14 Marek Polacek <polacek@redhat.com>
4432 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
4433 NUM_CONDITIONS bit positions.
4435 2013-08-13 Cary Coutant <ccoutant@google.com>
4437 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
4438 (attr_checksum): Hash vector contents instead of pointer.
4439 (attr_checksum_ordered): Likewise.
4441 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
4443 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
4444 when Pmode != word_mode. Add length_address attribute.
4445 (sse3_monitor_<mode>): Merge from sse3_monitor and
4446 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
4447 Pmode != word_mode. Update insn length attribute.
4448 * config/i386/i386.c (ix86_option_override_internal): Update
4449 ix86_gen_monitor selection for merged sse3_monitor insn.
4451 2013-08-13 Julian Brown <julian@codesourcery.com>
4453 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
4454 perform invalid legitimization on greater-than-word-size modes for
4457 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
4459 * ira.c (setup_class_translate_array): Use aclass instead of cl
4460 for classes not fully covered by allocno classes.
4462 2013-08-13 Jakub Jelinek <jakub@redhat.com>
4464 PR tree-optimization/57661
4465 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
4466 * tree-inline.c (tree_function_versioning): Initialize it.
4467 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
4468 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
4469 that is not being copied.
4472 * asan.c (instrument_strlen_call): Fix typo in comment.
4473 Use char * type even for the lhs of POINTER_PLUS_EXPR.
4475 2013-08-13 Steve Ellcey <sellcey@mips.com>
4477 * config/mips/mips.md (prefetch): Use lw instead of ld on
4478 loongson in 32bit mode.
4480 2013-08-13 Nick Clifton <nickc@redhat.com>
4482 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
4484 2013-08-13 Jan Hubicka <jh@suse.cz>
4486 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
4487 introduced edge; fix typo in sanity check.
4488 (cgraph_resolve_speculation): Export; improve diagnostic.
4489 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
4490 speculation at type mismatch.
4491 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
4492 (cgraph_resolve_speculation): Declare.
4493 (symtab_can_be_discarded): New function.
4494 * value-prof.c (gimple_ic_transform): Remove actual transform code.
4495 * ipa-inline-transform.c (speculation_removed): New global var.
4496 (clone_inlined_nodes): See if speculation can be removed.
4497 (inline_call): If speculations was removed, we growths may not match.
4498 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
4499 (speculation_useful_p): New function.
4500 (resolve_noninline_speculation): New function.
4501 (inline_small_functions): Resolve useless speculations.
4502 * ipa-inline.h (speculation_useful_p): Declare
4503 * ipa.c (can_replace_by_local_alias): Simplify.
4504 (ipa_profile): Produce speculative calls in non-lto, too;
4505 add simple cost model; produce local aliases.
4507 2013-08-13 David Malcolm <dmalcolm@redhat.com>
4509 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
4512 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
4514 * config/i386/i386.c (ix86_function_versions): Use error + inform.
4516 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
4518 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
4519 iterator instead of X87MODEF.
4521 2013-08-12 Perez Read <netfirewall@gmail.com>
4524 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
4525 operand 0 for intel asm alternative.
4526 (*movabs<mode>_2): Ditto for operand 1.
4528 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
4530 * config/aarch64/arm_none.h
4531 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
4533 2013-08-12 Nick Clifton <nickc@redhat.com>
4535 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
4537 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
4539 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
4540 expand for QI/HImode operand to produce more effictive code for
4541 unsigned char(short) --> float(double) conversion.
4543 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
4545 * doc/invoke.texi: Mention that -ftls-model does not force the final
4548 2013-08-12 Marek Polacek <polacek@redhat.com>
4549 Marc Glisse <marc.glisse@inria.fr>
4551 PR tree-optimization/57980
4552 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
4553 when creating -1 constant.
4555 2013-08-10 Jan Hubicka <jh@suse.cz>
4557 Workaround binutils PR14342.
4558 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
4559 (gimple_init_edge_profiler): Likewise.
4560 (gimple_gen_ic_func_profiler): Likewise.
4562 2013-08-09 Jan Hubicka <jh@suse.cz>
4564 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
4566 2013-08-09 Xinliang David Li <davidxl@google.com>
4568 * config/i386/stringop.def: New file.
4569 * config/i386/stringop.opt: New file.
4570 * config/i386/i386-opts.h: Include stringopt.def.
4571 * config/i386/i386.opt: Include stringopt.opt.
4572 * config/i386/i386.c (ix86_option_override_internal):
4573 Override default size based stringop inline strategies with options.
4574 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
4577 2013-08-09 Jan Hubicka <jh@suse.cz>
4579 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
4581 2013-08-09 Jan Hubicka <jh@suse.cz>
4583 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
4585 (dump_cgraph_node): Dump profile-id.
4586 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
4587 and common_target_probability.
4588 * lto-cgraph.c (lto_output_edge): Stream common targets.
4589 (lto_output_node): Stream profile ids.
4590 (input_node): Stream profile ids.
4591 (input_edge): Stream common targets.
4592 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
4593 * ipa.c: Include value-prof.h
4594 (ipa_profile_generate_summary): Turn indirect call statement histograms
4595 into common targets.
4596 (ipa_profile): Turn common targets into speculative edges.
4598 2013-08-09 Jan Hubicka <jh@suse.cz>
4600 * cgraph.h (cgraph_node): Add profile_id.
4601 * value-prof.c (cgraph_node_map): Turn into pointer_map.
4602 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
4603 (del_node_map): Update.
4604 (find_func_by_funcdef_no): Replace by ...
4605 (find_func_by_profile_id): ... this one.
4606 (gimple_ic_transform): Do not remove useful histograms when
4607 speculation is not done; dump info when indirect call removal
4609 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
4610 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
4611 (__gcov_indirect_call_profiler_v2): .. this one.
4612 * profile.h (init_node_map): Update.
4613 * coverage.c (coverage_compute_profile_id): New function.
4614 * coverage.h (coverage_compute_profile_id): Declare.
4615 * tree-profile.c (init_ic_make_global_vars): Make
4616 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
4617 (gimple_init_edge_profiler): Update prototype of
4618 __gcov_indirect_call_profiler.
4619 (gimple_gen_ic_func_profiler): Simplify.
4620 (tree_profiling): Use init_node_map
4622 2013-08-09 Jan Hubicka <jh@suse.cz>
4624 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
4625 non-speculative refs.
4626 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
4627 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
4628 (cgraph_set_call_stmt): Likewise.
4629 (cgraph_create_edge_1): Fix release checking compilatoin;
4631 (cgraph_free_edge): Free indirect info.
4632 (cgraph_turn_edge_to_speculative): New function.
4633 (cgraph_speculative_call_info): New function.
4634 (cgraph_make_edge_direct): Return direct edge; handle speculation.
4635 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
4636 (dump_cgraph_node): Dump speculation.
4637 (verify_edge_count_and_frequency): Accept speculative edges.
4638 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
4639 (verify_cgraph_node): Handle speculation.
4640 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
4641 (cgraph_set_call_stmt): Update prototype.
4642 (cgraph_make_edge_direct): Update prototype.
4643 (cgraph_speculative_call_info): Declare.
4644 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
4645 to change; update call of ipa_find_references.
4646 * ipa-ref.c (ipa_record_reference): Fix return value; clear
4647 lto_stmt_uid and speculative flags.
4648 (ipa_dump_references): Dump speculation.
4649 (ipa_clone_references): Clone speculative flag.
4650 (ipa_clone_referring): Likewise.
4651 (ipa_clone_ref): New function.
4652 (ipa_find_reference): Look into lto_stmt_uids
4653 (ipa_clear_stmts_in_references): Do not clear speculative calls.
4654 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
4655 (ipa_find_reference): Update declaration.
4656 (ipa_clone_ref): Declare.
4657 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
4658 stream speculative flag.
4659 (lto_output_ref): Stream statements uids and speculation.
4660 (input_ref): Likewise.
4661 (input_edge): Stream speuclation.
4662 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
4663 (cgraph_set_call_stmt_including_clones): Handle speculation.
4664 * ipa-inline.c (heap_edge_removal_hook): New function.
4665 (inline_small_functions): Register it.
4666 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
4667 also initialize refs.
4668 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
4670 (try_make_edge_direct_simple_call): Likewise.
4671 (try_make_edge_direct_simple_call): Likewise.
4672 (update_indirect_edges_after_inlining): Likewise.
4673 (remove_described_reference): Look proper lto_stmt_uid.
4674 (propagate_controlled_uses): Likewise.
4675 (propagate_controlled_uses): Liekwise.
4676 * tree-inline.c (copy_bb): Copy speculative edges.
4677 (redirect_all_calls): New function.
4678 (copy_cfg_body): Do redirection after loop info is updated.
4679 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
4681 2013-08-09 Jan Hubicka <jh@suse.cz>
4683 * lto-streamer-out.c (output_function): Renumber PHIs.
4684 * lto-streamer-in.c (input_function): Likewise.
4686 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
4688 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
4689 (get_lane_unsigned): Likewise.
4690 (dup_lane_scalar): Likewise.
4691 (get_lane): enable for VALL.
4692 * config/aarch64/aarch64-simd.md
4693 (aarch64_dup_lane_scalar<mode>): Remove.
4694 (aarch64_get_lane_signed<mode>): Likewise.
4695 (aarch64_get_lane_unsigned<mode>): Likewise.
4696 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
4697 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
4698 (aarch64_get_lane<mode>): Enable for all vector modes.
4699 (aarch64_get_lanedi): Remove misleading constraints.
4700 * config/aarch64/arm_neon.h
4701 (__aarch64_vget_lane_any): Define.
4702 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
4703 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
4704 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
4705 * config/aarch64/iterators.md (VDQQH): New.
4709 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
4711 * configure.ac: Add GAS check for LEON instructions on SPARC.
4712 * configure: Regenerate.
4713 * config.in: Likewise.
4714 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
4716 * config/sparc/sparc.opt (LEON, LEON3): New masks.
4717 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
4719 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
4720 (AS_LEON_FLAG): New macro.
4721 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
4722 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
4723 Deal with LEON and LEON3 for the memory model.
4724 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
4725 (atomic_compare_and_swap<mode>_1): Likewise.
4726 (*atomic_compare_and_swap<mode>_1): Likewise.
4728 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4730 * config/arm/neon.md (vcond): Fix floating-point vector
4731 comparisons against 0.
4733 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
4735 * lra-constraints.c (emit_spill_move): Remove assert.
4736 (process_alt_operands): Add more debugging
4737 output. Increase reject for spilling into memory. Decrease
4738 reject for reloading scratch.
4739 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
4741 2013-08-08 Steve Ellcey <sellcey@mips.com>
4743 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
4744 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
4745 micromips incompatible. Add nan2008.
4746 (MULTILIB_DIRNAMES): Add nan2008.
4747 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
4748 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
4749 and micromips incompatible. Add nan2008.
4750 (MULTILIB_DIRNAMES): Add nan2008.
4751 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
4753 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
4755 PR rtl-optimization/58079
4756 * combine.c (combine_simplify_rtx): Avoid using SUBST if
4757 simplify_comparison has widened a comparison with an integer.
4759 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4761 * config/arm/neon.md (movmisalign<mode>): Disable when we
4762 don't allow unaligned accesses.
4763 (*movmisalign<mode>_neon_store): Likewise.
4764 (*movmisalign<mode>_neon_load): Likewise.
4765 (*movmisalign<mode>_neon_store): Likewise.
4766 (*movmisalign<mode>_neon_load): Likewise.
4768 2013-08-08 Jan Hubicka <jh@suse.cz>
4770 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
4771 (make_pass_rebuild_cgraph_edges): Also clear references.
4772 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
4773 * ipa-inline-transform.c (inline_transform): Remove all references
4775 * cgraphunit.c (expand_function): Remove all references after
4777 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
4778 (ipa_find_reference): Rewrite to iterator.
4779 (remove_stmt_references): Likewise.
4780 (ipa_clear_stmts_in_references): New function.
4781 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
4782 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
4784 * ipa-split.c (split_function): Remove references in split function.
4786 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4789 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
4790 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
4792 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4795 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
4796 suggested mode for the assignment isn't compatible with the
4799 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
4802 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
4804 2013-08-07 Xinliang David Li <davidxl@google.com>
4806 * config/i386/i386.opt: New option -mtune-ctrl=.
4807 * config/i386/x86-tune.def: New file.
4808 * config/i386/i386.h: include x86-tune.def.
4809 * config/i386/i386.c (ix86_option_override_internal):
4810 Parsing -mtune-ctrl= option and set tune features.
4812 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
4815 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
4816 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
4819 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4821 * rtl.h (update_alignments): Declare.
4822 * final.c (grow_label_align): New function extracted from...
4823 (shorten_branches): ...here. Call it.
4824 (update_alignments): New function.
4825 * reorg.c (sibling_labels): New variable.
4826 (get_label_before): Add SIBLING parameter. If it is non-zero, push
4827 the new label along with it onto the sibling_labels vector.
4828 (fill_simple_delay_slots): Adjust call to get_label_before.
4829 (fill_slots_from_thread): Likewise.
4830 (relax_delay_slots): Likewise.
4831 (make_return_insns): Likewise.
4832 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
4834 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4836 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
4837 document its semantics.
4838 (diagnostic_report_diagnostic): Adjust accordingly.
4840 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4842 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
4843 (sparc_option_override): ...and port to new C++ pass API.
4844 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
4846 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
4848 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
4850 2013-08-06 Caroline Tice <cmtice@google.com>
4852 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
4853 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
4854 * tree-pass.h: Add pass_vtable_verify.
4855 * varasm.c (assemble_variable): Add code to properly set the comdat
4856 section and name for the .vtable_map_vars section.
4857 (assemble_vtyv_preinit_initializer): New function.
4858 (default_sectin_type_flags): Make sure .vtable_map_vars section has
4860 * output.h: Add function decl for assemble_vtv_preinit_initializer.
4861 * vtable-verify.c: New file.
4862 * vtable-verify.h: New file.
4863 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
4864 initialiation levels.
4865 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
4866 * passes.def: Insert pass_vtable_verify.
4867 * aclocal.m4: Reorder includes.
4868 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
4869 -fvtv-counts options.
4870 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
4871 as appropriate, if -fvtable-verify=... is used.
4872 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
4873 -fvtable-verify=... is used.
4874 * Makefile.in (OBJS): Add vtable-verify.o to list.
4875 (vtable-verify.o): Add new build rule.
4876 (GTFILES): Add vtable-verify.c to list.
4877 * common.opt (fvtable-verify=): New flag.
4878 (vtv_priority): Values for fvtable-verify= flag.
4879 (fvtv-counts): New flag.
4880 (fvtv-debug): New flag.
4881 * tree.h (save_vtable_map_decl): New extern function decl.
4883 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4885 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
4886 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
4887 (pass_data_rl78_devirt): ...new pass_data instance and...
4888 (make_pass_rl78_devirt): ...new function.
4889 (rl78_asm_file_start): Port pass registration to new C++ API.
4891 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4893 * coretypes.h (rtl_opt_pass): Add.
4894 (gcc::context): Add.
4895 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
4896 (epiphany_init): Port to new C++ pass API.
4897 (epiphany_optimize_mode_switching): Likewise.
4898 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
4899 (pass_manager::get_pass_mode_switching): New.
4900 (pass_manager::get_pass_peephole2): New.
4901 * mode-switching.c (pass_mode_switching): Add clone method.
4902 * recog.c (pass_peephole2): Add clone method.
4903 (pass_split_all_insns): Add clone method.
4905 2013-08-06 David Malcolm <dmalcolm@redhat.com>
4907 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
4908 (mips_option_override): ...here, porting to new C++ API for passes.
4910 2013-08-06 Jan Hubicka <jh@suse.cz>
4912 * cgraph.c (cgraph_get_body): New function based on lto.c
4914 * cgraph.h (cgraph_get_body): Declare.
4915 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
4917 * cgraphunit.c (expand_function): Get body prior expanding.
4918 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
4919 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
4921 * passes.c (do_per_function_toporder): Get body.
4922 * tree-inline.c (expand_call_inline): Get body prior inlining it.
4923 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
4925 2013-08-06 Martin Jambor <mjambor@suse.cz>
4928 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
4929 is not re-finalized. Rename second parameter to no_collect.
4931 2013-08-06 Martin Jambor <mjambor@suse.cz>
4934 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
4935 MEM_REF has proper alignment information.
4937 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
4940 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
4942 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
4943 rtx (*) (rtx, ...) with insn_gen_fn.
4944 * genoutput.c (output_insn_data): Cast gen_? function pointers to
4945 insn_gen_fn::stored_funcptr. Add initializer braces.
4947 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4949 Rewrite how instances of passes are cloned to remove assumptions
4950 about their sizes (thus allowing pass subclasses to have
4951 additional data fields, albeit non-GC-managed ones at this point).
4953 * passes.c (make_pass_instance): Now that passes have clone
4954 methods, rewrite this function to eliminate XNEW and memcpy
4955 calls that used hardcoded sizes. Since this function no longer
4956 creates pass instances, rename it to...
4957 (add_pass_instance): ...this. Document the old way that passes were
4958 numbered and flagged, and rework this function to continue using it.
4959 (next_pass_1): Add an initial_pass argument for use by
4961 (position_pass): When adding multiple instances of a pass, use
4962 the pass's clone method, rather than relying on the XNEW/memcpy
4963 within the former make_pass_instance (now add_pass_instance).
4964 (pass_manager::pass_manager): When invoking next_pass_1, also supply
4965 the initial instance of the current pass within the pass manager.
4967 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4969 This is the automated part of the conversion of passes from C
4970 structs to C++ classes.
4972 Patch autogenerated by refactor_passes.py from
4973 https://github.com/davidmalcolm/gcc-refactoring-scripts
4974 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
4976 * asan.c (pass_asan): Convert from a global struct to a subclass of
4977 gimple_opt_pass along with...
4978 (pass_data_asan): ...new pass_data instance and...
4979 (make_pass_asan): ...new function.
4980 (pass_asan_O0): Convert from a global struct to a subclass of
4981 gimple_opt_pass along with...
4982 (pass_data_asan_O0): ...new pass_data instance and...
4983 (make_pass_asan_O0): ...new function.
4984 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
4985 subclass of rtl_opt_pass along with...
4986 (pass_data_inc_dec): ...new pass_data instance and...
4987 (make_pass_inc_dec): ...new function.
4988 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
4989 a subclass of rtl_opt_pass along with...
4990 (pass_data_reorder_blocks): ...new pass_data instance and...
4991 (make_pass_reorder_blocks): ...new function.
4992 (pass_duplicate_computed_gotos): Convert from a global struct to a
4993 subclass of rtl_opt_pass along with...
4994 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
4995 (make_pass_duplicate_computed_gotos): ...new function.
4996 (pass_partition_blocks): Convert from a global struct to a subclass of
4997 rtl_opt_pass along with...
4998 (pass_data_partition_blocks): ...new pass_data instance and...
4999 (make_pass_partition_blocks): ...new function.
5000 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5001 struct to a subclass of rtl_opt_pass along with...
5002 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5004 (make_pass_branch_target_load_optimize1): ...new function.
5005 (pass_branch_target_load_optimize2): Convert from a global struct to a
5006 subclass of rtl_opt_pass along with...
5007 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5009 (make_pass_branch_target_load_optimize2): ...new function.
5010 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5011 of rtl_opt_pass along with...
5012 (pass_data_jump): ...new pass_data instance and...
5013 (make_pass_jump): ...new function.
5014 (pass_jump2): Convert from a global struct to a subclass of
5015 rtl_opt_pass along with...
5016 (pass_data_jump2): ...new pass_data instance and...
5017 (make_pass_jump2): ...new function.
5018 * cfgexpand.c (pass_expand): Convert from a global struct to a
5019 subclass of rtl_opt_pass along with...
5020 (pass_data_expand): ...new pass_data instance and...
5021 (make_pass_expand): ...new function.
5022 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5023 of rtl_opt_pass along with...
5024 (pass_data_free_cfg): ...new pass_data instance and...
5025 (make_pass_free_cfg): ...new function.
5026 (pass_into_cfg_layout_mode): Convert from a global struct to a
5027 subclass of rtl_opt_pass along with...
5028 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5029 (make_pass_into_cfg_layout_mode): ...new function.
5030 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5031 subclass of rtl_opt_pass along with...
5032 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5033 (make_pass_outof_cfg_layout_mode): ...new function.
5034 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5035 struct to a subclass of gimple_opt_pass along with...
5036 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5037 (make_pass_build_cgraph_edges): ...new function.
5038 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5039 subclass of gimple_opt_pass along with...
5040 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5041 (make_pass_rebuild_cgraph_edges): ...new function.
5042 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5043 subclass of gimple_opt_pass along with...
5044 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5046 (make_pass_remove_cgraph_callee_edges): ...new function.
5047 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5048 struct to a subclass of rtl_opt_pass along with...
5049 (pass_data_stack_adjustments): ...new pass_data instance and...
5050 (make_pass_stack_adjustments): ...new function.
5051 * combine.c (pass_combine): Convert from a global struct to a subclass
5052 of rtl_opt_pass along with...
5053 (pass_data_combine): ...new pass_data instance and...
5054 (make_pass_combine): ...new function.
5055 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
5056 global struct to a subclass of rtl_opt_pass along with...
5057 (pass_data_compare_elim_after_reload): ...new pass_data instance
5059 (make_pass_compare_elim_after_reload): ...new function.
5060 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
5061 of rtl_opt_pass along with...
5062 (pass_data_rtl_cprop): ...new pass_data instance and...
5063 (make_pass_rtl_cprop): ...new function.
5064 * cse.c (pass_cse): Convert from a global struct to a subclass of
5065 rtl_opt_pass along with...
5066 (pass_data_cse): ...new pass_data instance and...
5067 (make_pass_cse): ...new function.
5068 (pass_cse2): Convert from a global struct to a subclass of
5069 rtl_opt_pass along with...
5070 (pass_data_cse2): ...new pass_data instance and...
5071 (make_pass_cse2): ...new function.
5072 (pass_cse_after_global_opts): Convert from a global struct to a
5073 subclass of rtl_opt_pass along with...
5074 (pass_data_cse_after_global_opts): ...new pass_data instance and...
5075 (make_pass_cse_after_global_opts): ...new function.
5076 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
5077 of rtl_opt_pass along with...
5078 (pass_data_ud_rtl_dce): ...new pass_data instance and...
5079 (make_pass_ud_rtl_dce): ...new function.
5080 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
5081 rtl_opt_pass along with...
5082 (pass_data_fast_rtl_dce): ...new pass_data instance and...
5083 (make_pass_fast_rtl_dce): ...new function.
5084 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
5085 a subclass of rtl_opt_pass along with...
5086 (pass_data_df_initialize_opt): ...new pass_data instance and...
5087 (make_pass_df_initialize_opt): ...new function.
5088 (pass_df_initialize_no_opt): Convert from a global struct to a
5089 subclass of rtl_opt_pass along with...
5090 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
5091 (make_pass_df_initialize_no_opt): ...new function.
5092 (pass_df_finish): Convert from a global struct to a subclass of
5093 rtl_opt_pass along with...
5094 (pass_data_df_finish): ...new pass_data instance and...
5095 (make_pass_df_finish): ...new function.
5096 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
5097 rtl_opt_pass along with...
5098 (pass_data_rtl_dse1): ...new pass_data instance and...
5099 (make_pass_rtl_dse1): ...new function.
5100 (pass_rtl_dse2): Convert from a global struct to a subclass of
5101 rtl_opt_pass along with...
5102 (pass_data_rtl_dse2): ...new pass_data instance and...
5103 (make_pass_rtl_dse2): ...new function.
5104 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
5105 subclass of rtl_opt_pass along with...
5106 (pass_data_dwarf2_frame): ...new pass_data instance and...
5107 (make_pass_dwarf2_frame): ...new function.
5108 * except.c (pass_set_nothrow_function_flags): Convert from a global
5109 struct to a subclass of rtl_opt_pass along with...
5110 (pass_data_set_nothrow_function_flags): ...new pass_data instance
5112 (make_pass_set_nothrow_function_flags): ...new function.
5113 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
5114 subclass of rtl_opt_pass along with...
5115 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
5117 (make_pass_convert_to_eh_region_ranges): ...new function.
5118 * final.c (pass_compute_alignments): Convert from a global struct to a
5119 subclass of rtl_opt_pass along with...
5120 (pass_data_compute_alignments): ...new pass_data instance and...
5121 (make_pass_compute_alignments): ...new function.
5122 (pass_final): Convert from a global struct to a subclass of
5123 rtl_opt_pass along with...
5124 (pass_data_final): ...new pass_data instance and...
5125 (make_pass_final): ...new function.
5126 (pass_shorten_branches): Convert from a global struct to a subclass of
5127 rtl_opt_pass along with...
5128 (pass_data_shorten_branches): ...new pass_data instance and...
5129 (make_pass_shorten_branches): ...new function.
5130 (pass_clean_state): Convert from a global struct to a subclass of
5131 rtl_opt_pass along with...
5132 (pass_data_clean_state): ...new pass_data instance and...
5133 (make_pass_clean_state): ...new function.
5134 * function.c (pass_instantiate_virtual_regs): Convert from a global
5135 struct to a subclass of rtl_opt_pass along with...
5136 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
5137 (make_pass_instantiate_virtual_regs): ...new function.
5138 (pass_leaf_regs): Convert from a global struct to a subclass of
5139 rtl_opt_pass along with...
5140 (pass_data_leaf_regs): ...new pass_data instance and...
5141 (make_pass_leaf_regs): ...new function.
5142 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
5143 subclass of rtl_opt_pass along with...
5144 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
5146 (make_pass_thread_prologue_and_epilogue): ...new function.
5147 (pass_match_asm_constraints): Convert from a global struct to a
5148 subclass of rtl_opt_pass along with...
5149 (pass_data_match_asm_constraints): ...new pass_data instance and...
5150 (make_pass_match_asm_constraints): ...new function.
5151 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
5152 subclass of rtl_opt_pass along with...
5153 (pass_data_rtl_fwprop): ...new pass_data instance and...
5154 (make_pass_rtl_fwprop): ...new function.
5155 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
5156 rtl_opt_pass along with...
5157 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
5158 (make_pass_rtl_fwprop_addr): ...new function.
5159 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
5160 rtl_opt_pass along with...
5161 (pass_data_rtl_pre): ...new pass_data instance and...
5162 (make_pass_rtl_pre): ...new function.
5163 (pass_rtl_hoist): Convert from a global struct to a subclass of
5164 rtl_opt_pass along with...
5165 (pass_data_rtl_hoist): ...new pass_data instance and...
5166 (make_pass_rtl_hoist): ...new function.
5167 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
5168 subclass of gimple_opt_pass along with...
5169 (pass_data_lower_cf): ...new pass_data instance and...
5170 (make_pass_lower_cf): ...new function.
5171 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
5172 from a global struct to a subclass of gimple_opt_pass along with...
5173 (pass_data_strength_reduction): ...new pass_data instance and...
5174 (make_pass_strength_reduction): ...new function.
5175 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
5176 of rtl_opt_pass along with...
5177 (pass_data_rtl_ifcvt): ...new pass_data instance and...
5178 (make_pass_rtl_ifcvt): ...new function.
5179 (pass_if_after_combine): Convert from a global struct to a subclass of
5180 rtl_opt_pass along with...
5181 (pass_data_if_after_combine): ...new pass_data instance and...
5182 (make_pass_if_after_combine): ...new function.
5183 (pass_if_after_reload): Convert from a global struct to a subclass of
5184 rtl_opt_pass along with...
5185 (pass_data_if_after_reload): ...new pass_data instance and...
5186 (make_pass_if_after_reload): ...new function.
5187 * init-regs.c (pass_initialize_regs): Convert from a global struct to
5188 a subclass of rtl_opt_pass along with...
5189 (pass_data_initialize_regs): ...new pass_data instance and...
5190 (make_pass_initialize_regs): ...new function.
5191 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
5192 of ipa_opt_pass_d along with...
5193 (pass_data_ipa_cp): ...new pass_data instance and...
5194 (make_pass_ipa_cp): ...new function.
5195 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
5196 global struct to a subclass of gimple_opt_pass along with...
5197 (pass_data_inline_parameters): ...new pass_data instance and...
5198 (make_pass_inline_parameters): ...new function.
5199 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
5200 subclass of gimple_opt_pass along with...
5201 (pass_data_early_inline): ...new pass_data instance and...
5202 (make_pass_early_inline): ...new function.
5203 (pass_ipa_inline): Convert from a global struct to a subclass of
5204 ipa_opt_pass_d along with...
5205 (pass_data_ipa_inline): ...new pass_data instance and...
5206 (make_pass_ipa_inline): ...new function.
5207 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
5208 struct to a subclass of gimple_opt_pass along with...
5209 (pass_data_local_pure_const): ...new pass_data instance and...
5210 (make_pass_local_pure_const): ...new function.
5211 (pass_ipa_pure_const): Convert from a global struct to a subclass of
5212 ipa_opt_pass_d along with...
5213 (pass_data_ipa_pure_const): ...new pass_data instance and...
5214 (make_pass_ipa_pure_const): ...new function.
5215 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
5216 to a subclass of ipa_opt_pass_d along with...
5217 (pass_data_ipa_reference): ...new pass_data instance and...
5218 (make_pass_ipa_reference): ...new function.
5219 * ipa-split.c (pass_split_functions): Convert from a global struct to
5220 a subclass of gimple_opt_pass along with...
5221 (pass_data_split_functions): ...new pass_data instance and...
5222 (make_pass_split_functions): ...new function.
5223 (pass_feedback_split_functions): Convert from a global struct to a
5224 subclass of gimple_opt_pass along with...
5225 (pass_data_feedback_split_functions): ...new pass_data instance and...
5226 (make_pass_feedback_split_functions): ...new function.
5227 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
5228 global struct to a subclass of simple_ipa_opt_pass along with...
5229 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
5231 (make_pass_ipa_function_and_variable_visibility): ...new function.
5232 (pass_ipa_free_inline_summary): Convert from a global struct to a
5233 subclass of simple_ipa_opt_pass along with...
5234 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
5235 (make_pass_ipa_free_inline_summary): ...new function.
5236 (pass_ipa_whole_program_visibility): Convert from a global struct to a
5237 subclass of ipa_opt_pass_d along with...
5238 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
5240 (make_pass_ipa_whole_program_visibility): ...new function.
5241 (pass_ipa_profile): Convert from a global struct to a subclass of
5242 ipa_opt_pass_d along with...
5243 (pass_data_ipa_profile): ...new pass_data instance and...
5244 (make_pass_ipa_profile): ...new function.
5245 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
5246 ipa_opt_pass_d along with...
5247 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
5248 (make_pass_ipa_cdtor_merge): ...new function.
5249 * ira.c (pass_ira): Convert from a global struct to a subclass of
5250 rtl_opt_pass along with...
5251 (pass_data_ira): ...new pass_data instance and...
5252 (make_pass_ira): ...new function.
5253 (pass_reload): Convert from a global struct to a subclass of
5254 rtl_opt_pass along with...
5255 (pass_data_reload): ...new pass_data instance and...
5256 (make_pass_reload): ...new function.
5257 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
5258 subclass of rtl_opt_pass along with...
5259 (pass_data_cleanup_barriers): ...new pass_data instance and...
5260 (make_pass_cleanup_barriers): ...new function.
5261 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
5262 of rtl_opt_pass along with...
5263 (pass_data_loop2): ...new pass_data instance and...
5264 (make_pass_loop2): ...new function.
5265 (pass_rtl_loop_init): Convert from a global struct to a subclass of
5266 rtl_opt_pass along with...
5267 (pass_data_rtl_loop_init): ...new pass_data instance and...
5268 (make_pass_rtl_loop_init): ...new function.
5269 (pass_rtl_loop_done): Convert from a global struct to a subclass of
5270 rtl_opt_pass along with...
5271 (pass_data_rtl_loop_done): ...new pass_data instance and...
5272 (make_pass_rtl_loop_done): ...new function.
5273 (pass_rtl_move_loop_invariants): Convert from a global struct to a
5274 subclass of rtl_opt_pass along with...
5275 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
5276 (make_pass_rtl_move_loop_invariants): ...new function.
5277 (pass_rtl_unswitch): Convert from a global struct to a subclass of
5278 rtl_opt_pass along with...
5279 (pass_data_rtl_unswitch): ...new pass_data instance and...
5280 (make_pass_rtl_unswitch): ...new function.
5281 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
5282 subclass of rtl_opt_pass along with...
5283 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
5285 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5286 (pass_rtl_doloop): Convert from a global struct to a subclass of
5287 rtl_opt_pass along with...
5288 (pass_data_rtl_doloop): ...new pass_data instance and...
5289 (make_pass_rtl_doloop): ...new function.
5290 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
5291 a subclass of rtl_opt_pass along with...
5292 (pass_data_lower_subreg): ...new pass_data instance and...
5293 (make_pass_lower_subreg): ...new function.
5294 (pass_lower_subreg2): Convert from a global struct to a subclass of
5295 rtl_opt_pass along with...
5296 (pass_data_lower_subreg2): ...new pass_data instance and...
5297 (make_pass_lower_subreg2): ...new function.
5298 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
5299 struct to a subclass of ipa_opt_pass_d along with...
5300 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
5301 (make_pass_ipa_lto_gimple_out): ...new function.
5302 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
5303 of ipa_opt_pass_d along with...
5304 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
5305 (make_pass_ipa_lto_finish_out): ...new function.
5306 * mode-switching.c (pass_mode_switching): Convert from a global struct
5307 to a subclass of rtl_opt_pass along with...
5308 (pass_data_mode_switching): ...new pass_data instance and...
5309 (make_pass_mode_switching): ...new function.
5310 * modulo-sched.c (pass_sms): Convert from a global struct to a
5311 subclass of rtl_opt_pass along with...
5312 (pass_data_sms): ...new pass_data instance and...
5313 (make_pass_sms): ...new function.
5314 * omp-low.c (pass_expand_omp): Convert from a global struct to a
5315 subclass of gimple_opt_pass along with...
5316 (pass_data_expand_omp): ...new pass_data instance and...
5317 (make_pass_expand_omp): ...new function.
5318 (pass_lower_omp): Convert from a global struct to a subclass of
5319 gimple_opt_pass along with...
5320 (pass_data_lower_omp): ...new pass_data instance and...
5321 (make_pass_lower_omp): ...new function.
5322 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
5323 of gimple_opt_pass along with...
5324 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
5325 (make_pass_diagnose_omp_blocks): ...new function.
5326 * passes.c (pass_early_local_passes): Convert from a global struct to
5327 a subclass of simple_ipa_opt_pass along with...
5328 (pass_data_early_local_passes): ...new pass_data instance and...
5329 (make_pass_early_local_passes): ...new function.
5330 (pass_all_early_optimizations): Convert from a global struct to a
5331 subclass of gimple_opt_pass along with...
5332 (pass_data_all_early_optimizations): ...new pass_data instance and...
5333 (make_pass_all_early_optimizations): ...new function.
5334 (pass_all_optimizations): Convert from a global struct to a subclass
5335 of gimple_opt_pass along with...
5336 (pass_data_all_optimizations): ...new pass_data instance and...
5337 (make_pass_all_optimizations): ...new function.
5338 (pass_all_optimizations_g): Convert from a global struct to a subclass
5339 of gimple_opt_pass along with...
5340 (pass_data_all_optimizations_g): ...new pass_data instance and...
5341 (make_pass_all_optimizations_g): ...new function.
5342 (pass_rest_of_compilation): Convert from a global struct to a subclass
5343 of rtl_opt_pass along with...
5344 (pass_data_rest_of_compilation): ...new pass_data instance and...
5345 (make_pass_rest_of_compilation): ...new function.
5346 (pass_postreload): Convert from a global struct to a subclass of
5347 rtl_opt_pass along with...
5348 (pass_data_postreload): ...new pass_data instance and...
5349 (make_pass_postreload): ...new function.
5350 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
5351 subclass of rtl_opt_pass along with...
5352 (pass_data_gcse2): ...new pass_data instance and...
5353 (make_pass_gcse2): ...new function.
5354 * postreload.c (pass_postreload_cse): Convert from a global struct to
5355 a subclass of rtl_opt_pass along with...
5356 (pass_data_postreload_cse): ...new pass_data instance and...
5357 (make_pass_postreload_cse): ...new function.
5358 * predict.c (pass_profile): Convert from a global struct to a subclass
5359 of gimple_opt_pass along with...
5360 (pass_data_profile): ...new pass_data instance and...
5361 (make_pass_profile): ...new function.
5362 (pass_strip_predict_hints): Convert from a global struct to a subclass
5363 of gimple_opt_pass along with...
5364 (pass_data_strip_predict_hints): ...new pass_data instance and...
5365 (make_pass_strip_predict_hints): ...new function.
5366 * recog.c (pass_peephole2): Convert from a global struct to a subclass
5367 of rtl_opt_pass along with...
5368 (pass_data_peephole2): ...new pass_data instance and...
5369 (make_pass_peephole2): ...new function.
5370 (pass_split_all_insns): Convert from a global struct to a subclass of
5371 rtl_opt_pass along with...
5372 (pass_data_split_all_insns): ...new pass_data instance and...
5373 (make_pass_split_all_insns): ...new function.
5374 (pass_split_after_reload): Convert from a global struct to a subclass
5375 of rtl_opt_pass along with...
5376 (pass_data_split_after_reload): ...new pass_data instance and...
5377 (make_pass_split_after_reload): ...new function.
5378 (pass_split_before_regstack): Convert from a global struct to a
5379 subclass of rtl_opt_pass along with...
5380 (pass_data_split_before_regstack): ...new pass_data instance and...
5381 (make_pass_split_before_regstack): ...new function.
5382 (pass_split_before_sched2): Convert from a global struct to a subclass
5383 of rtl_opt_pass along with...
5384 (pass_data_split_before_sched2): ...new pass_data instance and...
5385 (make_pass_split_before_sched2): ...new function.
5386 (pass_split_for_shorten_branches): Convert from a global struct to a
5387 subclass of rtl_opt_pass along with...
5388 (pass_data_split_for_shorten_branches): ...new pass_data instance
5390 (make_pass_split_for_shorten_branches): ...new function.
5391 * ree.c (pass_ree): Convert from a global struct to a subclass of
5392 rtl_opt_pass along with...
5393 (pass_data_ree): ...new pass_data instance and...
5394 (make_pass_ree): ...new function.
5395 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
5396 subclass of rtl_opt_pass along with...
5397 (pass_data_stack_regs): ...new pass_data instance and...
5398 (make_pass_stack_regs): ...new function.
5399 (pass_stack_regs_run): Convert from a global struct to a subclass of
5400 rtl_opt_pass along with...
5401 (pass_data_stack_regs_run): ...new pass_data instance and...
5402 (make_pass_stack_regs_run): ...new function.
5403 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
5404 subclass of rtl_opt_pass along with...
5405 (pass_data_cprop_hardreg): ...new pass_data instance and...
5406 (make_pass_cprop_hardreg): ...new function.
5407 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
5408 subclass of rtl_opt_pass along with...
5409 (pass_data_reginfo_init): ...new pass_data instance and...
5410 (make_pass_reginfo_init): ...new function.
5411 * regmove.c (pass_regmove): Convert from a global struct to a subclass
5412 of rtl_opt_pass along with...
5413 (pass_data_regmove): ...new pass_data instance and...
5414 (make_pass_regmove): ...new function.
5415 * regrename.c (pass_regrename): Convert from a global struct to a
5416 subclass of rtl_opt_pass along with...
5417 (pass_data_regrename): ...new pass_data instance and...
5418 (make_pass_regrename): ...new function.
5419 * reorg.c (pass_delay_slots): Convert from a global struct to a
5420 subclass of rtl_opt_pass along with...
5421 (pass_data_delay_slots): ...new pass_data instance and...
5422 (make_pass_delay_slots): ...new function.
5423 (pass_machine_reorg): Convert from a global struct to a subclass of
5424 rtl_opt_pass along with...
5425 (pass_data_machine_reorg): ...new pass_data instance and...
5426 (make_pass_machine_reorg): ...new function.
5427 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
5428 of rtl_opt_pass along with...
5429 (pass_data_sched): ...new pass_data instance and...
5430 (make_pass_sched): ...new function.
5431 (pass_sched2): Convert from a global struct to a subclass of
5432 rtl_opt_pass along with...
5433 (pass_data_sched2): ...new pass_data instance and...
5434 (make_pass_sched2): ...new function.
5435 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
5436 to a subclass of rtl_opt_pass along with...
5437 (pass_data_stack_ptr_mod): ...new pass_data instance and...
5438 (make_pass_stack_ptr_mod): ...new function.
5439 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
5440 to a subclass of rtl_opt_pass along with...
5441 (pass_data_rtl_store_motion): ...new pass_data instance and...
5442 (make_pass_rtl_store_motion): ...new function.
5443 * tracer.c (pass_tracer): Convert from a global struct to a subclass
5444 of gimple_opt_pass along with...
5445 (pass_data_tracer): ...new pass_data instance and...
5446 (make_pass_tracer): ...new function.
5447 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
5448 to a subclass of gimple_opt_pass along with...
5449 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
5450 (make_pass_diagnose_tm_blocks): ...new function.
5451 (pass_lower_tm): Convert from a global struct to a subclass of
5452 gimple_opt_pass along with...
5453 (pass_data_lower_tm): ...new pass_data instance and...
5454 (make_pass_lower_tm): ...new function.
5455 (pass_tm_init): Convert from a global struct to a subclass of
5456 gimple_opt_pass along with...
5457 (pass_data_tm_init): ...new pass_data instance and...
5458 (make_pass_tm_init): ...new function.
5459 (pass_tm_mark): Convert from a global struct to a subclass of
5460 gimple_opt_pass along with...
5461 (pass_data_tm_mark): ...new pass_data instance and...
5462 (make_pass_tm_mark): ...new function.
5463 (pass_tm_edges): Convert from a global struct to a subclass of
5464 gimple_opt_pass along with...
5465 (pass_data_tm_edges): ...new pass_data instance and...
5466 (make_pass_tm_edges): ...new function.
5467 (pass_tm_memopt): Convert from a global struct to a subclass of
5468 gimple_opt_pass along with...
5469 (pass_data_tm_memopt): ...new pass_data instance and...
5470 (make_pass_tm_memopt): ...new function.
5471 (pass_ipa_tm): Convert from a global struct to a subclass of
5472 simple_ipa_opt_pass along with...
5473 (pass_data_ipa_tm): ...new pass_data instance and...
5474 (make_pass_ipa_tm): ...new function.
5475 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
5476 subclass of gimple_opt_pass along with...
5477 (pass_data_call_cdce): ...new pass_data instance and...
5478 (make_pass_call_cdce): ...new function.
5479 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
5480 subclass of gimple_opt_pass along with...
5481 (pass_data_build_cfg): ...new pass_data instance and...
5482 (make_pass_build_cfg): ...new function.
5483 (pass_split_crit_edges): Convert from a global struct to a subclass of
5484 gimple_opt_pass along with...
5485 (pass_data_split_crit_edges): ...new pass_data instance and...
5486 (make_pass_split_crit_edges): ...new function.
5487 (pass_warn_function_return): Convert from a global struct to a
5488 subclass of gimple_opt_pass along with...
5489 (pass_data_warn_function_return): ...new pass_data instance and...
5490 (make_pass_warn_function_return): ...new function.
5491 (pass_warn_function_noreturn): Convert from a global struct to a
5492 subclass of gimple_opt_pass along with...
5493 (pass_data_warn_function_noreturn): ...new pass_data instance and...
5494 (make_pass_warn_function_noreturn): ...new function.
5495 (pass_warn_unused_result): Convert from a global struct to a subclass
5496 of gimple_opt_pass along with...
5497 (pass_data_warn_unused_result): ...new pass_data instance and...
5498 (make_pass_warn_unused_result): ...new function.
5499 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
5500 a subclass of gimple_opt_pass along with...
5501 (pass_data_merge_phi): ...new pass_data instance and...
5502 (make_pass_merge_phi): ...new function.
5503 * tree-complex.c (pass_lower_complex): Convert from a global struct to
5504 a subclass of gimple_opt_pass along with...
5505 (pass_data_lower_complex): ...new pass_data instance and...
5506 (make_pass_lower_complex): ...new function.
5507 (pass_lower_complex_O0): Convert from a global struct to a subclass of
5508 gimple_opt_pass along with...
5509 (pass_data_lower_complex_O0): ...new pass_data instance and...
5510 (make_pass_lower_complex_O0): ...new function.
5511 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
5512 subclass of gimple_opt_pass along with...
5513 (pass_data_lower_eh): ...new pass_data instance and...
5514 (make_pass_lower_eh): ...new function.
5515 (pass_refactor_eh): Convert from a global struct to a subclass of
5516 gimple_opt_pass along with...
5517 (pass_data_refactor_eh): ...new pass_data instance and...
5518 (make_pass_refactor_eh): ...new function.
5519 (pass_lower_resx): Convert from a global struct to a subclass of
5520 gimple_opt_pass along with...
5521 (pass_data_lower_resx): ...new pass_data instance and...
5522 (make_pass_lower_resx): ...new function.
5523 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
5524 of gimple_opt_pass along with...
5525 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
5526 (make_pass_lower_eh_dispatch): ...new function.
5527 (pass_cleanup_eh): Convert from a global struct to a subclass of
5528 gimple_opt_pass along with...
5529 (pass_data_cleanup_eh): ...new pass_data instance and...
5530 (make_pass_cleanup_eh): ...new function.
5531 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
5532 to a subclass of simple_ipa_opt_pass along with...
5533 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
5534 (make_pass_ipa_lower_emutls): ...new function.
5535 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
5536 a subclass of gimple_opt_pass along with...
5537 (pass_data_if_conversion): ...new pass_data instance and...
5538 (make_pass_if_conversion): ...new function.
5539 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
5540 subclass of gimple_opt_pass along with...
5541 (pass_data_build_ssa): ...new pass_data instance and...
5542 (make_pass_build_ssa): ...new function.
5543 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
5544 global struct to a subclass of gimple_opt_pass along with...
5545 (pass_data_loop_distribution): ...new pass_data instance and...
5546 (make_pass_loop_distribution): ...new function.
5547 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
5548 subclass of gimple_opt_pass along with...
5549 (pass_data_mudflap_1): ...new pass_data instance and...
5550 (make_pass_mudflap_1): ...new function.
5551 (pass_mudflap_2): Convert from a global struct to a subclass of
5552 gimple_opt_pass along with...
5553 (pass_data_mudflap_2): ...new pass_data instance and...
5554 (make_pass_mudflap_2): ...new function.
5555 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
5556 subclass of gimple_opt_pass along with...
5557 (pass_data_mudflap_1): ...new pass_data instance and...
5558 (make_pass_mudflap_1): ...new function.
5559 (pass_mudflap_2): Convert from a global struct to a subclass of
5560 gimple_opt_pass along with...
5561 (pass_data_mudflap_2): ...new pass_data instance and...
5562 (make_pass_mudflap_2): ...new function.
5563 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
5564 gimple_opt_pass along with...
5565 (pass_data_nrv): ...new pass_data instance and...
5566 (make_pass_nrv): ...new function.
5567 (pass_return_slot): Convert from a global struct to a subclass of
5568 gimple_opt_pass along with...
5569 (pass_data_return_slot): ...new pass_data instance and...
5570 (make_pass_return_slot): ...new function.
5571 * tree-object-size.c (pass_object_sizes): Convert from a global struct
5572 to a subclass of gimple_opt_pass along with...
5573 (pass_data_object_sizes): ...new pass_data instance and...
5574 (make_pass_object_sizes): ...new function.
5575 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
5576 global struct to a subclass of gimple_opt_pass along with...
5577 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
5579 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5580 (pass_fixup_cfg): Convert from a global struct to a subclass of
5581 gimple_opt_pass along with...
5582 (pass_data_fixup_cfg): ...new pass_data instance and...
5583 (make_pass_fixup_cfg): ...new function.
5584 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
5585 (make_pass_mudflap_1): ...new function.
5586 (pass_mudflap_2): Replace declaration with that of...
5587 (make_pass_mudflap_2): ...new function.
5588 (pass_asan): Replace declaration with that of...
5589 (make_pass_asan): ...new function.
5590 (pass_asan_O0): Replace declaration with that of...
5591 (make_pass_asan_O0): ...new function.
5592 (pass_tsan): Replace declaration with that of...
5593 (make_pass_tsan): ...new function.
5594 (pass_tsan_O0): Replace declaration with that of...
5595 (make_pass_tsan_O0): ...new function.
5596 (pass_lower_cf): Replace declaration with that of...
5597 (make_pass_lower_cf): ...new function.
5598 (pass_refactor_eh): Replace declaration with that of...
5599 (make_pass_refactor_eh): ...new function.
5600 (pass_lower_eh): Replace declaration with that of...
5601 (make_pass_lower_eh): ...new function.
5602 (pass_lower_eh_dispatch): Replace declaration with that of...
5603 (make_pass_lower_eh_dispatch): ...new function.
5604 (pass_lower_resx): Replace declaration with that of...
5605 (make_pass_lower_resx): ...new function.
5606 (pass_build_cfg): Replace declaration with that of...
5607 (make_pass_build_cfg): ...new function.
5608 (pass_early_tree_profile): Replace declaration with that of...
5609 (make_pass_early_tree_profile): ...new function.
5610 (pass_cleanup_eh): Replace declaration with that of...
5611 (make_pass_cleanup_eh): ...new function.
5612 (pass_sra): Replace declaration with that of...
5613 (make_pass_sra): ...new function.
5614 (pass_sra_early): Replace declaration with that of...
5615 (make_pass_sra_early): ...new function.
5616 (pass_early_ipa_sra): Replace declaration with that of...
5617 (make_pass_early_ipa_sra): ...new function.
5618 (pass_tail_recursion): Replace declaration with that of...
5619 (make_pass_tail_recursion): ...new function.
5620 (pass_tail_calls): Replace declaration with that of...
5621 (make_pass_tail_calls): ...new function.
5622 (pass_tree_loop): Replace declaration with that of...
5623 (make_pass_tree_loop): ...new function.
5624 (pass_tree_loop_init): Replace declaration with that of...
5625 (make_pass_tree_loop_init): ...new function.
5626 (pass_lim): Replace declaration with that of...
5627 (make_pass_lim): ...new function.
5628 (pass_tree_unswitch): Replace declaration with that of...
5629 (make_pass_tree_unswitch): ...new function.
5630 (pass_predcom): Replace declaration with that of...
5631 (make_pass_predcom): ...new function.
5632 (pass_iv_canon): Replace declaration with that of...
5633 (make_pass_iv_canon): ...new function.
5634 (pass_scev_cprop): Replace declaration with that of...
5635 (make_pass_scev_cprop): ...new function.
5636 (pass_empty_loop): Replace declaration with that of...
5637 (make_pass_empty_loop): ...new function.
5638 (pass_record_bounds): Replace declaration with that of...
5639 (make_pass_record_bounds): ...new function.
5640 (pass_graphite): Replace declaration with that of...
5641 (make_pass_graphite): ...new function.
5642 (pass_graphite_transforms): Replace declaration with that of...
5643 (make_pass_graphite_transforms): ...new function.
5644 (pass_if_conversion): Replace declaration with that of...
5645 (make_pass_if_conversion): ...new function.
5646 (pass_loop_distribution): Replace declaration with that of...
5647 (make_pass_loop_distribution): ...new function.
5648 (pass_vectorize): Replace declaration with that of...
5649 (make_pass_vectorize): ...new function.
5650 (pass_slp_vectorize): Replace declaration with that of...
5651 (make_pass_slp_vectorize): ...new function.
5652 (pass_complete_unroll): Replace declaration with that of...
5653 (make_pass_complete_unroll): ...new function.
5654 (pass_complete_unrolli): Replace declaration with that of...
5655 (make_pass_complete_unrolli): ...new function.
5656 (pass_parallelize_loops): Replace declaration with that of...
5657 (make_pass_parallelize_loops): ...new function.
5658 (pass_loop_prefetch): Replace declaration with that of...
5659 (make_pass_loop_prefetch): ...new function.
5660 (pass_iv_optimize): Replace declaration with that of...
5661 (make_pass_iv_optimize): ...new function.
5662 (pass_tree_loop_done): Replace declaration with that of...
5663 (make_pass_tree_loop_done): ...new function.
5664 (pass_ch): Replace declaration with that of...
5665 (make_pass_ch): ...new function.
5666 (pass_ccp): Replace declaration with that of...
5667 (make_pass_ccp): ...new function.
5668 (pass_phi_only_cprop): Replace declaration with that of...
5669 (make_pass_phi_only_cprop): ...new function.
5670 (pass_build_ssa): Replace declaration with that of...
5671 (make_pass_build_ssa): ...new function.
5672 (pass_build_alias): Replace declaration with that of...
5673 (make_pass_build_alias): ...new function.
5674 (pass_build_ealias): Replace declaration with that of...
5675 (make_pass_build_ealias): ...new function.
5676 (pass_dominator): Replace declaration with that of...
5677 (make_pass_dominator): ...new function.
5678 (pass_dce): Replace declaration with that of...
5679 (make_pass_dce): ...new function.
5680 (pass_dce_loop): Replace declaration with that of...
5681 (make_pass_dce_loop): ...new function.
5682 (pass_cd_dce): Replace declaration with that of...
5683 (make_pass_cd_dce): ...new function.
5684 (pass_call_cdce): Replace declaration with that of...
5685 (make_pass_call_cdce): ...new function.
5686 (pass_merge_phi): Replace declaration with that of...
5687 (make_pass_merge_phi): ...new function.
5688 (pass_split_crit_edges): Replace declaration with that of...
5689 (make_pass_split_crit_edges): ...new function.
5690 (pass_pre): Replace declaration with that of...
5691 (make_pass_pre): ...new function.
5692 (pass_profile): Replace declaration with that of...
5693 (make_pass_profile): ...new function.
5694 (pass_strip_predict_hints): Replace declaration with that of...
5695 (make_pass_strip_predict_hints): ...new function.
5696 (pass_lower_complex_O0): Replace declaration with that of...
5697 (make_pass_lower_complex_O0): ...new function.
5698 (pass_lower_complex): Replace declaration with that of...
5699 (make_pass_lower_complex): ...new function.
5700 (pass_lower_vector): Replace declaration with that of...
5701 (make_pass_lower_vector): ...new function.
5702 (pass_lower_vector_ssa): Replace declaration with that of...
5703 (make_pass_lower_vector_ssa): ...new function.
5704 (pass_lower_omp): Replace declaration with that of...
5705 (make_pass_lower_omp): ...new function.
5706 (pass_diagnose_omp_blocks): Replace declaration with that of...
5707 (make_pass_diagnose_omp_blocks): ...new function.
5708 (pass_expand_omp): Replace declaration with that of...
5709 (make_pass_expand_omp): ...new function.
5710 (pass_expand_omp_ssa): Replace declaration with that of...
5711 (make_pass_expand_omp_ssa): ...new function.
5712 (pass_object_sizes): Replace declaration with that of...
5713 (make_pass_object_sizes): ...new function.
5714 (pass_strlen): Replace declaration with that of...
5715 (make_pass_strlen): ...new function.
5716 (pass_fold_builtins): Replace declaration with that of...
5717 (make_pass_fold_builtins): ...new function.
5718 (pass_stdarg): Replace declaration with that of...
5719 (make_pass_stdarg): ...new function.
5720 (pass_early_warn_uninitialized): Replace declaration with that of...
5721 (make_pass_early_warn_uninitialized): ...new function.
5722 (pass_late_warn_uninitialized): Replace declaration with that of...
5723 (make_pass_late_warn_uninitialized): ...new function.
5724 (pass_cse_reciprocals): Replace declaration with that of...
5725 (make_pass_cse_reciprocals): ...new function.
5726 (pass_cse_sincos): Replace declaration with that of...
5727 (make_pass_cse_sincos): ...new function.
5728 (pass_optimize_bswap): Replace declaration with that of...
5729 (make_pass_optimize_bswap): ...new function.
5730 (pass_optimize_widening_mul): Replace declaration with that of...
5731 (make_pass_optimize_widening_mul): ...new function.
5732 (pass_warn_function_return): Replace declaration with that of...
5733 (make_pass_warn_function_return): ...new function.
5734 (pass_warn_function_noreturn): Replace declaration with that of...
5735 (make_pass_warn_function_noreturn): ...new function.
5736 (pass_cselim): Replace declaration with that of...
5737 (make_pass_cselim): ...new function.
5738 (pass_phiopt): Replace declaration with that of...
5739 (make_pass_phiopt): ...new function.
5740 (pass_forwprop): Replace declaration with that of...
5741 (make_pass_forwprop): ...new function.
5742 (pass_phiprop): Replace declaration with that of...
5743 (make_pass_phiprop): ...new function.
5744 (pass_tree_ifcombine): Replace declaration with that of...
5745 (make_pass_tree_ifcombine): ...new function.
5746 (pass_dse): Replace declaration with that of...
5747 (make_pass_dse): ...new function.
5748 (pass_nrv): Replace declaration with that of...
5749 (make_pass_nrv): ...new function.
5750 (pass_rename_ssa_copies): Replace declaration with that of...
5751 (make_pass_rename_ssa_copies): ...new function.
5752 (pass_sink_code): Replace declaration with that of...
5753 (make_pass_sink_code): ...new function.
5754 (pass_fre): Replace declaration with that of...
5755 (make_pass_fre): ...new function.
5756 (pass_check_data_deps): Replace declaration with that of...
5757 (make_pass_check_data_deps): ...new function.
5758 (pass_copy_prop): Replace declaration with that of...
5759 (make_pass_copy_prop): ...new function.
5760 (pass_vrp): Replace declaration with that of...
5761 (make_pass_vrp): ...new function.
5762 (pass_uncprop): Replace declaration with that of...
5763 (make_pass_uncprop): ...new function.
5764 (pass_return_slot): Replace declaration with that of...
5765 (make_pass_return_slot): ...new function.
5766 (pass_reassoc): Replace declaration with that of...
5767 (make_pass_reassoc): ...new function.
5768 (pass_rebuild_cgraph_edges): Replace declaration with that of...
5769 (make_pass_rebuild_cgraph_edges): ...new function.
5770 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
5771 (make_pass_remove_cgraph_callee_edges): ...new function.
5772 (pass_build_cgraph_edges): Replace declaration with that of...
5773 (make_pass_build_cgraph_edges): ...new function.
5774 (pass_local_pure_const): Replace declaration with that of...
5775 (make_pass_local_pure_const): ...new function.
5776 (pass_tracer): Replace declaration with that of...
5777 (make_pass_tracer): ...new function.
5778 (pass_warn_unused_result): Replace declaration with that of...
5779 (make_pass_warn_unused_result): ...new function.
5780 (pass_diagnose_tm_blocks): Replace declaration with that of...
5781 (make_pass_diagnose_tm_blocks): ...new function.
5782 (pass_lower_tm): Replace declaration with that of...
5783 (make_pass_lower_tm): ...new function.
5784 (pass_tm_init): Replace declaration with that of...
5785 (make_pass_tm_init): ...new function.
5786 (pass_tm_mark): Replace declaration with that of...
5787 (make_pass_tm_mark): ...new function.
5788 (pass_tm_memopt): Replace declaration with that of...
5789 (make_pass_tm_memopt): ...new function.
5790 (pass_tm_edges): Replace declaration with that of...
5791 (make_pass_tm_edges): ...new function.
5792 (pass_split_functions): Replace declaration with that of...
5793 (make_pass_split_functions): ...new function.
5794 (pass_feedback_split_functions): Replace declaration with that of...
5795 (make_pass_feedback_split_functions): ...new function.
5796 (pass_strength_reduction): Replace declaration with that of...
5797 (make_pass_strength_reduction): ...new function.
5798 (pass_ipa_lower_emutls): Replace declaration with that of...
5799 (make_pass_ipa_lower_emutls): ...new function.
5800 (pass_ipa_function_and_variable_visibility): Replace declaration with
5802 (make_pass_ipa_function_and_variable_visibility): ...new function.
5803 (pass_ipa_tree_profile): Replace declaration with that of...
5804 (make_pass_ipa_tree_profile): ...new function.
5805 (pass_early_local_passes): Replace declaration with that of...
5806 (make_pass_early_local_passes): ...new function.
5807 (pass_ipa_whole_program_visibility): Replace declaration with that
5809 (make_pass_ipa_whole_program_visibility): ...new function.
5810 (pass_ipa_lto_gimple_out): Replace declaration with that of...
5811 (make_pass_ipa_lto_gimple_out): ...new function.
5812 (pass_ipa_increase_alignment): Replace declaration with that of...
5813 (make_pass_ipa_increase_alignment): ...new function.
5814 (pass_ipa_inline): Replace declaration with that of...
5815 (make_pass_ipa_inline): ...new function.
5816 (pass_ipa_free_lang_data): Replace declaration with that of...
5817 (make_pass_ipa_free_lang_data): ...new function.
5818 (pass_ipa_free_inline_summary): Replace declaration with that of...
5819 (make_pass_ipa_free_inline_summary): ...new function.
5820 (pass_ipa_cp): Replace declaration with that of...
5821 (make_pass_ipa_cp): ...new function.
5822 (pass_ipa_reference): Replace declaration with that of...
5823 (make_pass_ipa_reference): ...new function.
5824 (pass_ipa_pure_const): Replace declaration with that of...
5825 (make_pass_ipa_pure_const): ...new function.
5826 (pass_ipa_pta): Replace declaration with that of...
5827 (make_pass_ipa_pta): ...new function.
5828 (pass_ipa_lto_finish_out): Replace declaration with that of...
5829 (make_pass_ipa_lto_finish_out): ...new function.
5830 (pass_ipa_tm): Replace declaration with that of...
5831 (make_pass_ipa_tm): ...new function.
5832 (pass_ipa_profile): Replace declaration with that of...
5833 (make_pass_ipa_profile): ...new function.
5834 (pass_ipa_cdtor_merge): Replace declaration with that of...
5835 (make_pass_ipa_cdtor_merge): ...new function.
5836 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
5838 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5839 (pass_init_datastructures): Replace declaration with that of...
5840 (make_pass_init_datastructures): ...new function.
5841 (pass_fixup_cfg): Replace declaration with that of...
5842 (make_pass_fixup_cfg): ...new function.
5843 (pass_expand): Replace declaration with that of...
5844 (make_pass_expand): ...new function.
5845 (pass_instantiate_virtual_regs): Replace declaration with that of...
5846 (make_pass_instantiate_virtual_regs): ...new function.
5847 (pass_rtl_fwprop): Replace declaration with that of...
5848 (make_pass_rtl_fwprop): ...new function.
5849 (pass_rtl_fwprop_addr): Replace declaration with that of...
5850 (make_pass_rtl_fwprop_addr): ...new function.
5851 (pass_jump): Replace declaration with that of...
5852 (make_pass_jump): ...new function.
5853 (pass_jump2): Replace declaration with that of...
5854 (make_pass_jump2): ...new function.
5855 (pass_lower_subreg): Replace declaration with that of...
5856 (make_pass_lower_subreg): ...new function.
5857 (pass_cse): Replace declaration with that of...
5858 (make_pass_cse): ...new function.
5859 (pass_fast_rtl_dce): Replace declaration with that of...
5860 (make_pass_fast_rtl_dce): ...new function.
5861 (pass_ud_rtl_dce): Replace declaration with that of...
5862 (make_pass_ud_rtl_dce): ...new function.
5863 (pass_rtl_dce): Replace declaration with that of...
5864 (make_pass_rtl_dce): ...new function.
5865 (pass_rtl_dse1): Replace declaration with that of...
5866 (make_pass_rtl_dse1): ...new function.
5867 (pass_rtl_dse2): Replace declaration with that of...
5868 (make_pass_rtl_dse2): ...new function.
5869 (pass_rtl_dse3): Replace declaration with that of...
5870 (make_pass_rtl_dse3): ...new function.
5871 (pass_rtl_cprop): Replace declaration with that of...
5872 (make_pass_rtl_cprop): ...new function.
5873 (pass_rtl_pre): Replace declaration with that of...
5874 (make_pass_rtl_pre): ...new function.
5875 (pass_rtl_hoist): Replace declaration with that of...
5876 (make_pass_rtl_hoist): ...new function.
5877 (pass_rtl_store_motion): Replace declaration with that of...
5878 (make_pass_rtl_store_motion): ...new function.
5879 (pass_cse_after_global_opts): Replace declaration with that of...
5880 (make_pass_cse_after_global_opts): ...new function.
5881 (pass_rtl_ifcvt): Replace declaration with that of...
5882 (make_pass_rtl_ifcvt): ...new function.
5883 (pass_into_cfg_layout_mode): Replace declaration with that of...
5884 (make_pass_into_cfg_layout_mode): ...new function.
5885 (pass_outof_cfg_layout_mode): Replace declaration with that of...
5886 (make_pass_outof_cfg_layout_mode): ...new function.
5887 (pass_loop2): Replace declaration with that of...
5888 (make_pass_loop2): ...new function.
5889 (pass_rtl_loop_init): Replace declaration with that of...
5890 (make_pass_rtl_loop_init): ...new function.
5891 (pass_rtl_move_loop_invariants): Replace declaration with that of...
5892 (make_pass_rtl_move_loop_invariants): ...new function.
5893 (pass_rtl_unswitch): Replace declaration with that of...
5894 (make_pass_rtl_unswitch): ...new function.
5895 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
5896 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5897 (pass_rtl_doloop): Replace declaration with that of...
5898 (make_pass_rtl_doloop): ...new function.
5899 (pass_rtl_loop_done): Replace declaration with that of...
5900 (make_pass_rtl_loop_done): ...new function.
5901 (pass_web): Replace declaration with that of...
5902 (make_pass_web): ...new function.
5903 (pass_cse2): Replace declaration with that of...
5904 (make_pass_cse2): ...new function.
5905 (pass_df_initialize_opt): Replace declaration with that of...
5906 (make_pass_df_initialize_opt): ...new function.
5907 (pass_df_initialize_no_opt): Replace declaration with that of...
5908 (make_pass_df_initialize_no_opt): ...new function.
5909 (pass_reginfo_init): Replace declaration with that of...
5910 (make_pass_reginfo_init): ...new function.
5911 (pass_inc_dec): Replace declaration with that of...
5912 (make_pass_inc_dec): ...new function.
5913 (pass_stack_ptr_mod): Replace declaration with that of...
5914 (make_pass_stack_ptr_mod): ...new function.
5915 (pass_initialize_regs): Replace declaration with that of...
5916 (make_pass_initialize_regs): ...new function.
5917 (pass_combine): Replace declaration with that of...
5918 (make_pass_combine): ...new function.
5919 (pass_if_after_combine): Replace declaration with that of...
5920 (make_pass_if_after_combine): ...new function.
5921 (pass_ree): Replace declaration with that of...
5922 (make_pass_ree): ...new function.
5923 (pass_partition_blocks): Replace declaration with that of...
5924 (make_pass_partition_blocks): ...new function.
5925 (pass_match_asm_constraints): Replace declaration with that of...
5926 (make_pass_match_asm_constraints): ...new function.
5927 (pass_regmove): Replace declaration with that of...
5928 (make_pass_regmove): ...new function.
5929 (pass_split_all_insns): Replace declaration with that of...
5930 (make_pass_split_all_insns): ...new function.
5931 (pass_fast_rtl_byte_dce): Replace declaration with that of...
5932 (make_pass_fast_rtl_byte_dce): ...new function.
5933 (pass_lower_subreg2): Replace declaration with that of...
5934 (make_pass_lower_subreg2): ...new function.
5935 (pass_mode_switching): Replace declaration with that of...
5936 (make_pass_mode_switching): ...new function.
5937 (pass_sms): Replace declaration with that of...
5938 (make_pass_sms): ...new function.
5939 (pass_sched): Replace declaration with that of...
5940 (make_pass_sched): ...new function.
5941 (pass_ira): Replace declaration with that of...
5942 (make_pass_ira): ...new function.
5943 (pass_reload): Replace declaration with that of...
5944 (make_pass_reload): ...new function.
5945 (pass_clean_state): Replace declaration with that of...
5946 (make_pass_clean_state): ...new function.
5947 (pass_branch_prob): Replace declaration with that of...
5948 (make_pass_branch_prob): ...new function.
5949 (pass_value_profile_transformations): Replace declaration with that
5951 (make_pass_value_profile_transformations): ...new function.
5952 (pass_postreload_cse): Replace declaration with that of...
5953 (make_pass_postreload_cse): ...new function.
5954 (pass_gcse2): Replace declaration with that of...
5955 (make_pass_gcse2): ...new function.
5956 (pass_split_after_reload): Replace declaration with that of...
5957 (make_pass_split_after_reload): ...new function.
5958 (pass_branch_target_load_optimize1): Replace declaration with that
5960 (make_pass_branch_target_load_optimize1): ...new function.
5961 (pass_thread_prologue_and_epilogue): Replace declaration with that
5963 (make_pass_thread_prologue_and_epilogue): ...new function.
5964 (pass_stack_adjustments): Replace declaration with that of...
5965 (make_pass_stack_adjustments): ...new function.
5966 (pass_peephole2): Replace declaration with that of...
5967 (make_pass_peephole2): ...new function.
5968 (pass_if_after_reload): Replace declaration with that of...
5969 (make_pass_if_after_reload): ...new function.
5970 (pass_regrename): Replace declaration with that of...
5971 (make_pass_regrename): ...new function.
5972 (pass_cprop_hardreg): Replace declaration with that of...
5973 (make_pass_cprop_hardreg): ...new function.
5974 (pass_reorder_blocks): Replace declaration with that of...
5975 (make_pass_reorder_blocks): ...new function.
5976 (pass_branch_target_load_optimize2): Replace declaration with that
5978 (make_pass_branch_target_load_optimize2): ...new function.
5979 (pass_leaf_regs): Replace declaration with that of...
5980 (make_pass_leaf_regs): ...new function.
5981 (pass_split_before_sched2): Replace declaration with that of...
5982 (make_pass_split_before_sched2): ...new function.
5983 (pass_compare_elim_after_reload): Replace declaration with that of...
5984 (make_pass_compare_elim_after_reload): ...new function.
5985 (pass_sched2): Replace declaration with that of...
5986 (make_pass_sched2): ...new function.
5987 (pass_stack_regs): Replace declaration with that of...
5988 (make_pass_stack_regs): ...new function.
5989 (pass_stack_regs_run): Replace declaration with that of...
5990 (make_pass_stack_regs_run): ...new function.
5991 (pass_df_finish): Replace declaration with that of...
5992 (make_pass_df_finish): ...new function.
5993 (pass_compute_alignments): Replace declaration with that of...
5994 (make_pass_compute_alignments): ...new function.
5995 (pass_duplicate_computed_gotos): Replace declaration with that of...
5996 (make_pass_duplicate_computed_gotos): ...new function.
5997 (pass_variable_tracking): Replace declaration with that of...
5998 (make_pass_variable_tracking): ...new function.
5999 (pass_free_cfg): Replace declaration with that of...
6000 (make_pass_free_cfg): ...new function.
6001 (pass_machine_reorg): Replace declaration with that of...
6002 (make_pass_machine_reorg): ...new function.
6003 (pass_cleanup_barriers): Replace declaration with that of...
6004 (make_pass_cleanup_barriers): ...new function.
6005 (pass_delay_slots): Replace declaration with that of...
6006 (make_pass_delay_slots): ...new function.
6007 (pass_split_for_shorten_branches): Replace declaration with that of...
6008 (make_pass_split_for_shorten_branches): ...new function.
6009 (pass_split_before_regstack): Replace declaration with that of...
6010 (make_pass_split_before_regstack): ...new function.
6011 (pass_convert_to_eh_region_ranges): Replace declaration with that
6013 (make_pass_convert_to_eh_region_ranges): ...new function.
6014 (pass_shorten_branches): Replace declaration with that of...
6015 (make_pass_shorten_branches): ...new function.
6016 (pass_set_nothrow_function_flags): Replace declaration with that of...
6017 (make_pass_set_nothrow_function_flags): ...new function.
6018 (pass_dwarf2_frame): Replace declaration with that of...
6019 (make_pass_dwarf2_frame): ...new function.
6020 (pass_final): Replace declaration with that of...
6021 (make_pass_final): ...new function.
6022 (pass_rtl_seqabstr): Replace declaration with that of...
6023 (make_pass_rtl_seqabstr): ...new function.
6024 (pass_release_ssa_names): Replace declaration with that of...
6025 (make_pass_release_ssa_names): ...new function.
6026 (pass_early_inline): Replace declaration with that of...
6027 (make_pass_early_inline): ...new function.
6028 (pass_inline_parameters): Replace declaration with that of...
6029 (make_pass_inline_parameters): ...new function.
6030 (pass_update_address_taken): Replace declaration with that of...
6031 (make_pass_update_address_taken): ...new function.
6032 (pass_convert_switch): Replace declaration with that of...
6033 (make_pass_convert_switch): ...new function.
6034 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6035 to a subclass of simple_ipa_opt_pass along with...
6036 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6037 (make_pass_ipa_tree_profile): ...new function.
6038 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6039 subclass of gimple_opt_pass along with...
6040 (pass_data_sra_early): ...new pass_data instance and...
6041 (make_pass_sra_early): ...new function.
6042 (pass_sra): Convert from a global struct to a subclass of
6043 gimple_opt_pass along with...
6044 (pass_data_sra): ...new pass_data instance and...
6045 (make_pass_sra): ...new function.
6046 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6047 gimple_opt_pass along with...
6048 (pass_data_early_ipa_sra): ...new pass_data instance and...
6049 (make_pass_early_ipa_sra): ...new function.
6050 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6051 subclass of gimple_opt_pass along with...
6052 (pass_data_ccp): ...new pass_data instance and...
6053 (make_pass_ccp): ...new function.
6054 (pass_fold_builtins): Convert from a global struct to a subclass of
6055 gimple_opt_pass along with...
6056 (pass_data_fold_builtins): ...new pass_data instance and...
6057 (make_pass_fold_builtins): ...new function.
6058 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
6059 subclass of gimple_opt_pass along with...
6060 (pass_data_copy_prop): ...new pass_data instance and...
6061 (make_pass_copy_prop): ...new function.
6062 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
6063 global struct to a subclass of gimple_opt_pass along with...
6064 (pass_data_rename_ssa_copies): ...new pass_data instance and...
6065 (make_pass_rename_ssa_copies): ...new function.
6066 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
6067 subclass of gimple_opt_pass along with...
6068 (pass_data_dce): ...new pass_data instance and...
6069 (make_pass_dce): ...new function.
6070 (pass_dce_loop): Convert from a global struct to a subclass of
6071 gimple_opt_pass along with...
6072 (pass_data_dce_loop): ...new pass_data instance and...
6073 (make_pass_dce_loop): ...new function.
6074 (pass_cd_dce): Convert from a global struct to a subclass of
6075 gimple_opt_pass along with...
6076 (pass_data_cd_dce): ...new pass_data instance and...
6077 (make_pass_cd_dce): ...new function.
6078 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
6079 subclass of gimple_opt_pass along with...
6080 (pass_data_dominator): ...new pass_data instance and...
6081 (make_pass_dominator): ...new function.
6082 (pass_phi_only_cprop): Convert from a global struct to a subclass of
6083 gimple_opt_pass along with...
6084 (pass_data_phi_only_cprop): ...new pass_data instance and...
6085 (make_pass_phi_only_cprop): ...new function.
6086 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
6087 subclass of gimple_opt_pass along with...
6088 (pass_data_dse): ...new pass_data instance and...
6089 (make_pass_dse): ...new function.
6090 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
6091 a subclass of gimple_opt_pass along with...
6092 (pass_data_forwprop): ...new pass_data instance and...
6093 (make_pass_forwprop): ...new function.
6094 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
6095 struct to a subclass of gimple_opt_pass along with...
6096 (pass_data_tree_ifcombine): ...new pass_data instance and...
6097 (make_pass_tree_ifcombine): ...new function.
6098 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
6099 subclass of gimple_opt_pass along with...
6100 (pass_data_ch): ...new pass_data instance and...
6101 (make_pass_ch): ...new function.
6102 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
6103 subclass of gimple_opt_pass along with...
6104 (pass_data_tree_loop): ...new pass_data instance and...
6105 (make_pass_tree_loop): ...new function.
6106 (pass_tree_loop_init): Convert from a global struct to a subclass of
6107 gimple_opt_pass along with...
6108 (pass_data_tree_loop_init): ...new pass_data instance and...
6109 (make_pass_tree_loop_init): ...new function.
6110 (pass_lim): Convert from a global struct to a subclass of
6111 gimple_opt_pass along with...
6112 (pass_data_lim): ...new pass_data instance and...
6113 (make_pass_lim): ...new function.
6114 (pass_tree_unswitch): Convert from a global struct to a subclass of
6115 gimple_opt_pass along with...
6116 (pass_data_tree_unswitch): ...new pass_data instance and...
6117 (make_pass_tree_unswitch): ...new function.
6118 (pass_predcom): Convert from a global struct to a subclass of
6119 gimple_opt_pass along with...
6120 (pass_data_predcom): ...new pass_data instance and...
6121 (make_pass_predcom): ...new function.
6122 (pass_vectorize): Convert from a global struct to a subclass of
6123 gimple_opt_pass along with...
6124 (pass_data_vectorize): ...new pass_data instance and...
6125 (make_pass_vectorize): ...new function.
6126 (pass_graphite): Convert from a global struct to a subclass of
6127 gimple_opt_pass along with...
6128 (pass_data_graphite): ...new pass_data instance and...
6129 (make_pass_graphite): ...new function.
6130 (pass_graphite_transforms): Convert from a global struct to a subclass
6131 of gimple_opt_pass along with...
6132 (pass_data_graphite_transforms): ...new pass_data instance and...
6133 (make_pass_graphite_transforms): ...new function.
6134 (pass_check_data_deps): Convert from a global struct to a subclass of
6135 gimple_opt_pass along with...
6136 (pass_data_check_data_deps): ...new pass_data instance and...
6137 (make_pass_check_data_deps): ...new function.
6138 (pass_iv_canon): Convert from a global struct to a subclass of
6139 gimple_opt_pass along with...
6140 (pass_data_iv_canon): ...new pass_data instance and...
6141 (make_pass_iv_canon): ...new function.
6142 (pass_scev_cprop): Convert from a global struct to a subclass of
6143 gimple_opt_pass along with...
6144 (pass_data_scev_cprop): ...new pass_data instance and...
6145 (make_pass_scev_cprop): ...new function.
6146 (pass_record_bounds): Convert from a global struct to a subclass of
6147 gimple_opt_pass along with...
6148 (pass_data_record_bounds): ...new pass_data instance and...
6149 (make_pass_record_bounds): ...new function.
6150 (pass_complete_unroll): Convert from a global struct to a subclass of
6151 gimple_opt_pass along with...
6152 (pass_data_complete_unroll): ...new pass_data instance and...
6153 (make_pass_complete_unroll): ...new function.
6154 (pass_complete_unrolli): Convert from a global struct to a subclass of
6155 gimple_opt_pass along with...
6156 (pass_data_complete_unrolli): ...new pass_data instance and...
6157 (make_pass_complete_unrolli): ...new function.
6158 (pass_parallelize_loops): Convert from a global struct to a subclass
6159 of gimple_opt_pass along with...
6160 (pass_data_parallelize_loops): ...new pass_data instance and...
6161 (make_pass_parallelize_loops): ...new function.
6162 (pass_loop_prefetch): Convert from a global struct to a subclass of
6163 gimple_opt_pass along with...
6164 (pass_data_loop_prefetch): ...new pass_data instance and...
6165 (make_pass_loop_prefetch): ...new function.
6166 (pass_iv_optimize): Convert from a global struct to a subclass of
6167 gimple_opt_pass along with...
6168 (pass_data_iv_optimize): ...new pass_data instance and...
6169 (make_pass_iv_optimize): ...new function.
6170 (pass_tree_loop_done): Convert from a global struct to a subclass of
6171 gimple_opt_pass along with...
6172 (pass_data_tree_loop_done): ...new pass_data instance and...
6173 (make_pass_tree_loop_done): ...new function.
6174 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
6175 struct to a subclass of gimple_opt_pass along with...
6176 (pass_data_cse_reciprocals): ...new pass_data instance and...
6177 (make_pass_cse_reciprocals): ...new function.
6178 (pass_cse_sincos): Convert from a global struct to a subclass of
6179 gimple_opt_pass along with...
6180 (pass_data_cse_sincos): ...new pass_data instance and...
6181 (make_pass_cse_sincos): ...new function.
6182 (pass_optimize_bswap): Convert from a global struct to a subclass of
6183 gimple_opt_pass along with...
6184 (pass_data_optimize_bswap): ...new pass_data instance and...
6185 (make_pass_optimize_bswap): ...new function.
6186 (pass_optimize_widening_mul): Convert from a global struct to a
6187 subclass of gimple_opt_pass along with...
6188 (pass_data_optimize_widening_mul): ...new pass_data instance and...
6189 (make_pass_optimize_widening_mul): ...new function.
6190 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
6191 subclass of gimple_opt_pass along with...
6192 (pass_data_phiopt): ...new pass_data instance and...
6193 (make_pass_phiopt): ...new function.
6194 (pass_cselim): Convert from a global struct to a subclass of
6195 gimple_opt_pass along with...
6196 (pass_data_cselim): ...new pass_data instance and...
6197 (make_pass_cselim): ...new function.
6198 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
6199 subclass of gimple_opt_pass along with...
6200 (pass_data_phiprop): ...new pass_data instance and...
6201 (make_pass_phiprop): ...new function.
6202 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
6203 subclass of gimple_opt_pass along with...
6204 (pass_data_pre): ...new pass_data instance and...
6205 (make_pass_pre): ...new function.
6206 (pass_fre): Convert from a global struct to a subclass of
6207 gimple_opt_pass along with...
6208 (pass_data_fre): ...new pass_data instance and...
6209 (make_pass_fre): ...new function.
6210 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
6211 subclass of gimple_opt_pass along with...
6212 (pass_data_reassoc): ...new pass_data instance and...
6213 (make_pass_reassoc): ...new function.
6214 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
6215 subclass of gimple_opt_pass along with...
6216 (pass_data_sink_code): ...new pass_data instance and...
6217 (make_pass_sink_code): ...new function.
6218 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
6219 subclass of gimple_opt_pass along with...
6220 (pass_data_strlen): ...new pass_data instance and...
6221 (make_pass_strlen): ...new function.
6222 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
6223 struct to a subclass of gimple_opt_pass along with...
6224 (pass_data_build_alias): ...new pass_data instance and...
6225 (make_pass_build_alias): ...new function.
6226 (pass_build_ealias): Convert from a global struct to a subclass of
6227 gimple_opt_pass along with...
6228 (pass_data_build_ealias): ...new pass_data instance and...
6229 (make_pass_build_ealias): ...new function.
6230 (pass_ipa_pta): Convert from a global struct to a subclass of
6231 simple_ipa_opt_pass along with...
6232 (pass_data_ipa_pta): ...new pass_data instance and...
6233 (make_pass_ipa_pta): ...new function.
6234 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
6235 subclass of gimple_opt_pass along with...
6236 (pass_data_uncprop): ...new pass_data instance and...
6237 (make_pass_uncprop): ...new function.
6238 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
6239 global struct to a subclass of gimple_opt_pass along with...
6240 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
6241 (make_pass_late_warn_uninitialized): ...new function.
6242 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
6243 to a subclass of gimple_opt_pass along with...
6244 (pass_data_init_datastructures): ...new pass_data instance and...
6245 (make_pass_init_datastructures): ...new function.
6246 (pass_early_warn_uninitialized): Convert from a global struct to a
6247 subclass of gimple_opt_pass along with...
6248 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
6249 (make_pass_early_warn_uninitialized): ...new function.
6250 (pass_update_address_taken): Convert from a global struct to a
6251 subclass of gimple_opt_pass along with...
6252 (pass_data_update_address_taken): ...new pass_data instance and...
6253 (make_pass_update_address_taken): ...new function.
6254 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
6255 struct to a subclass of gimple_opt_pass along with...
6256 (pass_data_release_ssa_names): ...new pass_data instance and...
6257 (make_pass_release_ssa_names): ...new function.
6258 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
6259 subclass of gimple_opt_pass along with...
6260 (pass_data_stdarg): ...new pass_data instance and...
6261 (make_pass_stdarg): ...new function.
6262 * tree-switch-conversion.c (pass_convert_switch): Convert from a
6263 global struct to a subclass of gimple_opt_pass along with...
6264 (pass_data_convert_switch): ...new pass_data instance and...
6265 (make_pass_convert_switch): ...new function.
6266 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
6267 to a subclass of gimple_opt_pass along with...
6268 (pass_data_tail_recursion): ...new pass_data instance and...
6269 (make_pass_tail_recursion): ...new function.
6270 (pass_tail_calls): Convert from a global struct to a subclass of
6271 gimple_opt_pass along with...
6272 (pass_data_tail_calls): ...new pass_data instance and...
6273 (make_pass_tail_calls): ...new function.
6274 * tree-vect-generic.c (pass_lower_vector): Convert from a global
6275 struct to a subclass of gimple_opt_pass along with...
6276 (pass_data_lower_vector): ...new pass_data instance and...
6277 (make_pass_lower_vector): ...new function.
6278 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
6279 gimple_opt_pass along with...
6280 (pass_data_lower_vector_ssa): ...new pass_data instance and...
6281 (make_pass_lower_vector_ssa): ...new function.
6282 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
6283 to a subclass of gimple_opt_pass along with...
6284 (pass_data_slp_vectorize): ...new pass_data instance and...
6285 (make_pass_slp_vectorize): ...new function.
6286 (pass_ipa_increase_alignment): Convert from a global struct to a
6287 subclass of simple_ipa_opt_pass along with...
6288 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
6289 (make_pass_ipa_increase_alignment): ...new function.
6290 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
6291 gimple_opt_pass along with...
6292 (pass_data_vrp): ...new pass_data instance and...
6293 (make_pass_vrp): ...new function.
6294 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
6295 subclass of simple_ipa_opt_pass along with...
6296 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
6297 (make_pass_ipa_free_lang_data): ...new function.
6298 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
6299 gimple_opt_pass along with...
6300 (pass_data_tsan): ...new pass_data instance and...
6301 (make_pass_tsan): ...new function.
6302 (pass_tsan_O0): Convert from a global struct to a subclass of
6303 gimple_opt_pass along with...
6304 (pass_data_tsan_O0): ...new pass_data instance and...
6305 (make_pass_tsan_O0): ...new function.
6306 * var-tracking.c (pass_variable_tracking): Convert from a global
6307 struct to a subclass of rtl_opt_pass along with...
6308 (pass_data_variable_tracking): ...new pass_data instance and...
6309 (make_pass_variable_tracking): ...new function.
6310 * web.c (pass_web): Convert from a global struct to a subclass of
6311 rtl_opt_pass along with...
6312 (pass_data_web): ...new pass_data instance and...
6313 (make_pass_web): ...new function.
6314 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
6315 declaration with that of...
6316 (make_pass_mode_switch_use): ...new function.
6317 (pass_resolve_sw_modes): Replace declaration with that of...
6318 (make_pass_resolve_sw_modes): ...new function.
6319 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
6320 from a global struct to a subclass of rtl_opt_pass along with...
6321 (pass_data_mode_switch_use): ...new pass_data instance and...
6322 (make_pass_mode_switch_use): ...new function.
6323 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
6324 from a global struct to a subclass of rtl_opt_pass along with...
6325 (pass_data_resolve_sw_modes): ...new pass_data instance and...
6326 (make_pass_resolve_sw_modes): ...new function.
6327 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
6328 struct to a subclass of rtl_opt_pass along with...
6329 (pass_data_insert_vzeroupper): ...new pass_data instance and...
6330 (make_pass_insert_vzeroupper): ...new function.
6331 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
6332 global struct to a subclass of rtl_opt_pass along with...
6333 (pass_data_work_around_errata): ...new pass_data instance and...
6334 (make_pass_work_around_errata): ...new function.
6335 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
6336 struct to a subclass of rtl_opt_pass along with...
6337 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
6338 (make_pass_mips_machine_reorg2): ...new function.
6340 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6342 * passes.c (pass_manager::operator new): New.
6344 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6346 Handwritten part of conversion of passes to C++ classes.
6348 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
6349 (toplev.o): Add dep on PASS_MANAGER_H.
6350 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
6351 of early local pases to reflect this moving from a global to a
6352 member of gcc::pass_manager.
6353 (cgraph_add_new_function): Likewise.
6354 * lto-cgraph.c (lto_output_node): Update for conversion of
6355 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
6356 * passes.c (opt_pass::clone): New.
6357 (opt_pass::gate): New.
6358 (opt_pass::execute): New.
6359 (opt_pass::opt_pass): New.
6360 (pass_manager::execute_early_local_passes): New.
6361 (pass_manager::execute_pass_mode_switching): new.
6362 (finish_optimization_passes): Convert to...
6363 (pass_manager::finish_optimization_passes): ...this.
6364 (finish_optimization_passes): Update for conversion of passes to
6366 (register_dump_files_1): Use has_gate since we cannot portably
6367 check a vtable entry against NULL.
6368 (dump_one_pass): Likewise.
6369 (ipa_write_summaries_2): Likewise.
6370 (ipa_write_optimization_summaries_1): Likewise.
6371 (ipa_read_summaries_1): Likewise.
6372 (ipa_read_optimization_summaries_1): Likewise.
6373 (execute_ipa_stmt_fixups): Likewise.
6374 (pass_manager::pass_manager): Rewrite pass-creation, invoking
6375 pass-creation functions rather than wiring up globals, and
6376 storing the results in fields of pass_manager generated using
6378 (pass_manager::dump_profile_report): Update for conversion of
6379 passes to C++ classes.
6380 (pass_manager::execute_ipa_summary_passes): Likewise.
6381 (execute_one_ipa_transform_pass): Likewise.
6382 (execute_one_pass): Use has_gate and has_execute since we cannot
6383 portably check a vtable entry against NULL.
6384 * pass_manager.h (pass_manager::finish_optimization_passes): New.
6385 (pass_manager): Use pass-instances.def to add fields for the
6386 various pass instances.
6387 * toplev.c (finalize): Update for move of
6388 finish_optimization_passes to a method of gcc::pass_manager.
6389 * toplev.h (finish_optimization_passes): Move to method of class
6391 * tree-pass.h (struct pass_data): New.
6392 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
6393 (opt_pass::gate): Convert to virtual function.
6394 (opt_pass::~opt_pass): New.
6395 (opt_pass::clone): New.
6396 (opt_pass::execute): Convert to virtual function.
6397 (opt_pass::opt_pass): New.
6398 (opt_pass::ctxt_): new.
6399 (gimple_opt_pass): Convert to subclass of opt_pass.
6400 (gimple_opt_pass::gimple_opt_pass): New.
6401 (rtl_opt_pass): Convert to subclass of opt_pass.
6402 (rtl_opt_pass::rtl_opt_pass): New.
6403 (ipa_opt_pass_d): Convert to subclass of opt_pass.
6404 (ipa_opt_pass_d::ipa_opt_pass_d): New.
6405 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
6406 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
6407 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
6408 invocation of pass_mode_switching to reflect this moving from a
6409 global to a member of gcc::pass_manager.
6410 (ix86_option_override): Rework how pass_insert_vzeroupper is
6411 added to the pass_manager to reflect autogenerated changes.
6412 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
6414 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
6416 PR rtl-optimization/57708
6417 * recog.c (peep2_find_free_register): Validate all regs in a
6420 2013-08-05 Jan Hubicka <jh@suse.cz>
6423 * cgraph.c (verify_cgraph_node): Accept local flags from other
6425 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
6426 (function_and_variable_visibility): Likewise.
6427 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
6429 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6431 * graph.c (init_graph_slim_pretty_print): Remove.
6432 (print_graph_cfg): Do not call it. Use local pretty printer.
6433 (start_graph_dump): Likewise.
6435 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6437 * gimple-pretty-print.c (buffer): Remove.
6438 (initialized): Likewise.
6439 (maybe_init_pretty_print): Likewise.
6440 (print_gimple_stmt): Do not call it. Use non-static local
6441 pretty_printer variable.
6442 (print_gimple_expr): Likewise.
6443 (print_gimple_seq): Likewise.
6444 (gimple_dump_bb): Likewise.
6446 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6448 * asan.c (asan_pp): Remove.
6449 (asan_pp_initialized): Likewise.
6450 (asan_pp_initialize): Likewise.
6451 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
6452 (asan_emit_stack_protection): Tidy. Use local pretty printer.
6453 (asan_add_global): Likewise.
6455 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
6457 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
6458 * diagnostic.h (diagnostic_flush_buffer): Adjust.
6459 * pretty-print.c (pp_formatted_text_data): Likewise.
6460 (pp_indent): Rename from pp_base_indent.
6461 (pp_format): Rename from pp_base_format.
6462 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
6463 (pp_format_verbatim): Rename from pp_base_format_verbatim.
6464 (pp_flush): Rename from pp_base_flush.
6465 (pp_set_line_maximum_length): Rename from
6466 pp_base_set_line_maximum_length.
6467 (pp_clear_output_area): Rename from pp_base_clear_output_area.
6468 (pp_set_prefix): Rename from pp_base_set_prefix.
6469 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
6470 (pp_emit_prefix): Rename from pp_base_emit_prefix.
6471 (pp_append_text): Rename from pp_base_append_text.
6472 (pp_formatted_text): Rename from pp_base_formatted_text.
6473 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
6474 (pp_remaining_character_count_for_line): Rename from
6475 pp_base_remaining_character_count_for_line.
6476 (pp_newline): Rename from pp_base_newline.
6477 (pp_character): Rename from pp_base_character.
6478 (pp_string): Rename from pp_base_string.
6479 (pp_maybe_space): Rename from pp_base_maybe_space.
6480 * asan.c (asan_pp_string): Adjust.
6481 (asan_emit_stack_protection): Likewise.
6482 (asan_add_global): Likewise.
6483 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
6484 * tree-mudflap.c (mf_varname_tree): Likewise.
6485 * tree-pretty-print.c (pp_tree_identifier): Rename from
6486 pp_base_tree_identifier.
6487 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
6488 Declare as function.
6490 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
6492 * pretty-print.h (pp_bar_bar): New.
6493 (pp_ampersand_ampersand): Likewise.
6494 (pp_less_equal): Likewise.
6495 (pp_greater_equal): Likewise.
6496 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
6497 printer functions instead of pp_string or operators and punctuators.
6498 (dump_gimple_call): Likewise.
6499 (dump_gimple_omp_for): Likewise.
6500 (dump_gimple_transaction): Likewise.
6501 (dump_gimple_phi): Likewise.
6502 (pp_gimple_stmt_1): Likewise.
6503 * sched-vis.c (print_insn): Likewise.
6504 * tree-mudflap.c (mf_varname_tree): Likewise.
6505 * tree-pretty-print.c (dump_block_node): Likewise.
6506 (dump_generic_node): Likewise.
6508 2013-08-02 Jan Hubicka <jh@suse.cz>
6510 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
6512 * lto-streamer-out.c (tree_is_indexable): Results decls and
6513 parm decls are not indexable.
6514 (DFS_write_tree_body): Do not follow args and results.
6515 (hash_tree): Likewise.
6516 (output_functions): Rearrange so struct function is needed
6517 only when real body is output; be able to also ouptut abstract
6518 functions; output DECL_ARGUMENTS and DECL_RESULT.
6519 (lto_output): When not in WPA, ale store abstract functions.
6520 (write_symbol): Do not care about RESULT_DECL.
6521 (output_symbol_p): Handle correctly sbtract decls.
6522 * lto-streamer-in.c (input_function): Rearrange so struct
6523 function can be NULL at entry; allow streaming of
6524 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
6525 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
6526 sanity check during LTO.
6527 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
6528 RESULT_DECl and DECL_ARGUMENTS.
6529 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
6532 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
6534 * pretty-print.h (pp_underscore): New.
6536 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
6537 printer functions instead of pp_character.
6538 (dump_binary_rhs): Likewise.
6539 (dump_ternary_rhs): Likewise.
6540 (dump_gimple_call_args): Likewise.
6541 (pp_points_to_solution): Likewise.
6542 (dump_gimple_call): Likewise.
6543 (dump_gimple_switch): Likewise.
6544 (dump_gimple_cond): Likewise.
6545 (dump_gimple_bind): Likewise.
6546 (dump_gimple_try): Likewise.
6547 (dump_gimple_omp_for): Likewise.
6548 (dump_gimple_omp_continue): Likewise.
6549 (dump_gimple_omp_single): Likewise.
6550 (dump_gimple_omp_sections): Likewise.
6551 (dump_gimple_omp_block): Likewise.
6552 (dump_gimple_omp_critical): Likewise.
6553 (dump_gimple_transaction): Likewise.
6554 (dump_gimple_asm): Likewise.
6555 (dump_gimple_phi): Likewise.
6556 (dump_gimple_omp_parallel): Likewise.
6557 (dump_gimple_omp_task): Likewise.
6558 (dump_gimple_omp_atomic_load): Likewise.
6559 (dump_gimple_omp_atomic_store): Likewise.
6560 (dump_gimple_mem_ops): Likewise.
6561 (pp_gimple_stmt_1): Likewise.
6562 (pp_cfg_jump): Likewise.
6563 (dump_implicit_edges): Likewise.
6564 (gimple_dump_bb_for_graph): Likewise.
6565 * graph.c (draw_cfg_node): Likewise.
6566 * langhooks.c (lhd_print_error_function): Likewise.
6567 * sched-vis.c (print_exp): Likewise.
6568 (print_value): Likewise.
6569 (print_pattern): Likewise.
6570 (print_insn): Likewise.
6571 (rtl_dump_bb_for_graph): Likewise.
6572 * tree-pretty-print.c (dump_function_declaration): Likewise.
6573 (dump_array_domain): Likewise.
6574 (dump_omp_clause): Likewise.
6575 (dump_location): Likewise.
6576 (dump_generic_node): Likewise.
6577 (print_struct_decl): Likewise.
6578 * diagnostic.c (diagnostic_show_locus): Use pp_space.
6580 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6582 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
6583 candidate table when replacing a candidate statement.
6584 (replace_rhs_if_not_dup): Likewise.
6585 (replace_one_candidate): Likewise.
6587 2013-08-02 Jan Hubicka <jh@suse.cz>
6588 Martin Liska <marxin.liska@gmail.com>
6590 * cgraphunit.c (add_new_function): Fix logic when adding from
6592 (assemble_thunk): Rename to ...
6593 (expand_thunk); .. this one; export; get it working with
6594 general functions; make produced gimple valid.
6595 * cgraph.h (expand_thunk): Declare.
6597 2013-08-02 Jan Hubicka <jh@suse.cz>
6599 * ipa-cp.c (gather_context_independent_values): Use
6600 ipa_get_param_move_cost.
6601 (get_replacement_map): Remove PARAM; move parameter folding
6603 (create_specialized_node): Update.
6604 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
6605 assert that we have gimple body; update move_cost.
6606 (count_formal_params): Assert that we have gimple body.
6607 (ipa_dump_param): New function.
6608 (ipa_alloc_node_params): Break out from ...
6609 (ipa_initialize_node_params): ... here.
6610 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
6611 (ipa_write_node_info): Stream move costs.
6612 (ipa_read_node_info): Read move costs.
6613 (ipa_update_after_lto_read): Do not recompute node params.
6614 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
6615 (ipa_get_param): Check we are not in WPA.
6616 (ipa_get_param_move_cost): New.
6617 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
6618 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
6619 parm numbers to be present.
6621 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
6623 PR rtl-optimization/58048
6624 * lra-constraints.c (process_alt_operands): Don't check asm
6625 operand on register.
6627 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
6629 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
6630 the implied StoreLoad barrier for atomic operations if before.
6632 2013-08-02 Jan Hubicka <jh@suse.cz>
6633 Martin Liska <marxin.liska@gmail.com>
6635 * cgraph.c (cgraph_function_body_availability): Do not check
6637 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
6638 symtab_node_availability): Declare.
6639 * ipa.c (can_replace_by_local_alias): New.
6640 (function_and_variable_visibility): Use it.
6641 * symtab.c (symtab_for_node_and_aliases,
6642 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
6644 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
6646 PR rtl-optimization/57963
6647 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
6648 (lra_constraints): Use them.
6650 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
6652 * config/arm/types.md (define_attr "type"): Add "load_acq"
6654 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
6656 (cortex_a53_store1): Likewise.
6658 2013-08-01 Jan Hubicka <jh@suse.cz>
6660 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
6661 partitions are not needed.
6663 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
6665 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
6666 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
6667 MAYBE_NON_Q_CLASS_P where appropriate.
6669 2013-08-01 Jan Hubicka <jh@suse.cz>
6671 * cgraph.h (release_function_body): Declare.
6672 * tree.c (free_lang_data_in_decl): Free, parameters and return values
6673 of unused delcarations.
6675 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6677 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
6678 RTL form when subtracting a constant.
6680 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6682 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
6683 Generate canonical plus rtx with negated immediate instead of minus
6685 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
6687 2013-08-01 Jan Hubicka <jh@suse.cz>
6689 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
6690 (cgraph_release_function_body): Likewise.
6691 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
6692 * cgraph.h (cgrpah_node): Rename abstract_and_needed
6693 to used_as_abstract_origin.
6694 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
6695 symbols used as abstract origins.
6696 * cgraphunit.c (analyze_functions): Update.
6697 * ipa.c (symtab_remove_unreachable_nodes): Recompute
6698 used_as_abstract_origin.
6699 * tree-inline.c (tree_function_versioning): Update
6700 used_as_abstract_origin; be ready for DECL_RESULT and
6701 DECL_ARGUMENTS to be NULL.
6703 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
6704 for abstract functions.
6705 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
6708 2013-08-01 Jan Hubicka <jh@suse.cz>
6710 * profile.c (compute_value_histograms): Fix thinko.
6712 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
6714 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
6715 aarch-common-protos.h to extra_headers.
6716 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
6717 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
6718 * config/aarch64/t-aarch64 (aarch-common.o): Define.
6720 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
6722 * config/aarch64/aarch64.md (define_attr "type"): Delete.
6723 Include "../arm/types.md". Define "type" attribute for all patterns.
6724 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
6727 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
6729 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
6730 to support power8 load fusion.
6731 (fusion_gpr_mem_load): Likewise.
6733 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
6735 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
6736 declarations for power8 load fusion.
6737 (emit_fusion_gpr_load): Likewise.
6739 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
6740 tuning for power8, turn on fusion mode by default. Turn on sign
6741 extending fusion mode if normal fusion mode is on, and we are at
6743 (fusion_gpr_load_p): New function, return true if we can fuse an
6744 addis instruction with a dependent load to a GPR.
6745 (emit_fusion_gpr_load): Emit the instructions for power8 load
6748 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
6749 (VSX load fusion peepholes): New peepholes to fuse together an
6750 addi instruction with a VSX load instruction.
6752 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
6753 peepholes to fuse an addis instruction with a load to a GPR base
6754 register. If we are supporting sign extending fusions, convert
6755 sign extending loads to zero extending loads and add an explicit
6758 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6760 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
6761 aarch-common-protos.h to extra_headers.
6762 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
6763 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
6764 (arm_early_store_addr_dep): Likewise.
6765 (arm_no_early_alu_shift_dep): Likewise.
6766 (arm_no_early_alu_shift_value_dep): Likewise.
6767 (arm_no_early_mul_dep): Likewise.
6768 (arm_no_early_store_addr_dep): Likewise.
6769 (arm_mac_accumulator_is_mul_result): Likewise.
6770 (arm_mac_accumulator_is_result): Likewise.
6771 * config/arm/aarch-common.c: ... here. New file.
6772 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
6774 (arm_early_store_addr_dep): Likewise.
6775 (arm_no_early_alu_shift_dep): Likewise.
6776 (arm_no_early_alu_shift_value_dep): Likewise.
6777 (arm_no_early_mul_dep): Likewise.
6778 (arm_no_early_store_addr_dep): Likewise.
6779 (arm_mac_accumulator_is_mul_result): Likewise.
6780 (arm_mac_accumulator_is_result): Likewise.
6781 * config/arm/aarch-common-protos.h: ... here. New file.
6782 * config/arm/t-arm (aarch-common.o): Define.
6784 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6786 * config/arm/arm.md: Include new file "types.md".
6787 (define_attr "type"): Move from here to ...
6788 (define_attr "mul32"): Likewise.
6789 (define_attr "mul64"): Likewise.
6790 * config/arm/types.md: ... here. New file.
6792 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
6794 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
6795 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
6797 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6799 * gen-pass-instances.awk: Fix offset of substr().
6801 2013-07-31 David Malcolm <dmalcolm@redhat.com>
6803 * Makefile.in (pass-instances.def): New.
6804 (passes.o): Replace dependency on passes.def with one on
6807 * gen-pass-instances.awk: New.
6809 * passes.c (pass_manager::pass_manager): Use pass-instances.def
6810 rather than passes.def, updating local definition of NEXT_PASS
6811 macro to add an extra NUM parameter (currently unused).
6813 2013-07-30 David Malcolm <dmalcolm@redhat.com>
6815 * Makefile.in (PASS_MANAGER_H): New.
6816 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
6817 (passes.o): Likewise.
6818 (statistics.o): Likewise.
6819 (cgraphunit.o): Likewise.
6820 (context.o): Depend on PASS_MANAGER_H.
6822 * pass_manager.h: New.
6824 * cgraphunit.c (cgraph_add_new_function): Update for moves
6825 of globals to fields of pass_manager.
6826 (analyze_function): Likewise.
6827 (expand_function): Likewise.
6828 (ipa_passes): Likewise.
6829 (compile): Likewise.
6831 * context.c (context::context): New.
6832 * context.h (context::context): New.
6833 (context::get_passes): New.
6834 (context::passes_): New.
6836 * lto-cgraph.c (input_node): Update for moves of globals to
6837 fields of pass_manager.
6839 * passes.c (all_passes): Remove, in favor of a field of the
6840 same name within the new class pass_manager.
6841 (all_small_ipa_passes): Likewise.
6842 (all_lowering_passes): Likewise.
6843 (all_regular_ipa_passes): Likewise.
6844 (all_late_ipa_passes): Likewise.
6845 (all_lto_gen_passes): Likewise.
6846 (passes_by_id): Likewise.
6847 (passes_by_id_size): Likewise.
6848 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6849 the new class pass_manager.
6850 (set_pass_for_id): Convert to...
6851 (pass_manager::set_pass_for_id): ...method.
6852 (get_pass_for_id): Convert to...
6853 (pass_manager::get_pass_for_id): ...method.
6854 (register_one_dump_file): Move body of implementation into...
6855 (pass_manager::register_one_dump_file): ...here.
6856 (register_dump_files_1): Convert to...
6857 (pass_manager::register_dump_files_1): ...method.
6858 (register_dump_files): Convert to...
6859 (pass_manager::register_dump_files): ...method.
6860 (create_pass_tab): Update for moves of globals to fields of
6862 (dump_passes): Move body of implementation into...
6863 (pass_manager::dump_passes): ...here.
6864 (register_pass): Move body of implementation into...
6865 (pass_manager::register_pass): ...here.
6866 (init_optimization_passes): Convert into...
6867 (pass_manager::pass_manager): ...constructor for new
6868 pass_manager class, and initialize the pass_lists array.
6869 (check_profile_consistency): Update for moves of globals to
6870 fields of pass_manager.
6871 (dump_profile_report): Move body of implementation into...
6872 (pass_manager::dump_profile_report): ...here.
6873 (ipa_write_summaries_1): Update for moves of pass lists from
6874 being globals to fields of pass_manager.
6875 (ipa_write_optimization_summaries): Likewise.
6876 (ipa_read_summaries): Likewise.
6877 (ipa_read_optimization_summaries): Likewise.
6878 (execute_all_ipa_stmt_fixups): Likewise.
6880 * statistics.c (statistics_fini): Update for moves of globals to
6881 fields of pass_manager.
6883 * toplev.c (general_init): Replace call to
6884 init_optimization_passes with construction of the pass_manager
6887 * tree-pass.h (all_passes): Remove, in favor of a field of the
6888 same name within the new class pass_manager.
6889 (all_small_ipa_passes): Likewise.
6890 (all_lowering_passes): Likewise.
6891 (all_regular_ipa_passes): Likewise.
6892 (all_lto_gen_passes): Likewise.
6893 (all_late_ipa_passes): Likewise.
6894 (passes_by_id): Likewise.
6895 (passes_by_id_size): Likewise.
6896 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6897 the new class pass_manager.
6898 (get_pass_for_id): Remove.
6900 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6902 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
6905 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6907 * arm.md (mulhi3): New expand pattern.
6909 2013-07-30 Jan Hubicka <jh@suse.cz>
6910 Martin Liska <marxin.liska@gmail.com>
6912 * profile.c (compute_value_histograms): Do not ICE when
6913 there is mismatch only on some counters.
6915 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6917 PR rtl-optimization/57637
6918 * function.c (move_insn_for_shrink_wrap): Also check the
6919 GEN set of the LIVE problem for the liveness analysis
6920 if it exists, otherwise give up.
6922 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6924 PR tree-optimization/57993
6925 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
6926 replaced statement in the candidate table.
6927 (phi_add_costs): Return infinite cost when the hidden basis does
6928 not dominate all phis on which the candidate is dependent.
6929 (replace_one_candidate): Record replaced statement in the
6932 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
6934 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
6935 (ashlv2si3): New expander.
6936 (*ashlv2si3_i): New define_insn_and_split.
6937 * predicates.md (float_operation): Allow patterns with three
6940 PR rtl-optimization/58021
6941 * mode-switching.c (create_pre_exit): Always split off preceding
6942 insns if we are not at the basic block head.
6944 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
6946 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
6947 (UCLIBC_DYNAMIC_LINKER): New macro.
6948 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
6950 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
6951 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
6953 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
6954 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
6955 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
6956 for SF and DF modes. Use ieee_quad_format for TF mode.
6957 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
6958 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
6959 (mips_option_override): Handle `-mnan=legacy'.
6960 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
6961 `-mabs=2008' and `-mnan=2008'.
6962 (OPTION_DEFAULT_SPECS): Add "nan" default.
6963 (ASM_SPEC): Handle `-mnan='.
6964 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
6965 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
6966 comment accordingly.
6967 (neg<mode>2): Likewise.
6968 * config/mips/mips.opt (mabs, mnan): New options.
6969 * doc/install.texi (Configuration): Document `--with-nan=' option.
6970 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
6972 (MIPS Options): Document them.
6973 * config.gcc <mips*-*-*>: Handle `--with-nan='.
6974 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
6975 * configure: Regenerate.
6976 * config.in: Regenerate.
6978 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6980 * config/i386/i386.md (float post-reload splitters): Do not check
6981 for subregs of SSE registers.
6983 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6984 H.J. Lu <hongjiu.lu@intel.com>
6988 * config/i386/i386.md (post-reload splitter
6989 to avoid partial SSE reg dependency stalls): New pattern.
6991 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
6993 * config/s390/s390.md ("movcc"): Swap load and store instructions.
6995 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
6997 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
6998 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7000 2013-07-26 Cary Coutant <ccoutant@google.com>
7002 * dwarf2out.c (die_checksum_ordered): Don't include template
7003 instantiations in signature.
7004 (is_template_parameter): New function.
7005 (is_template_instantiation): New function.
7006 (generate_skeleton_bottom_up): Don't include template instantiations
7008 (generate_skeleton): Likewise.
7009 (break_out_comdat_types): Move recursive call to break out nested
7011 (prune_unused_types_mark_generic_parms_dies): Call
7012 is_template_parameter.
7014 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7016 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7017 uses vector registers.
7018 * config/aarch64/iterators.md: Add attributes rtn and vas.
7020 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7021 Richard Earnshaw <richard.earnshaw@arm.com>
7023 * combine.c (simplify_comparison): Re-canonicalize operands
7025 * config/arm/arm.md (movcond_addsi): New splitter.
7027 2013-07-25 Sterling Augustine <saugustine@google.com>
7029 * dwarf2out.c (size_of_pubnames): Move code to...
7030 (include_pubname_in_output): ...here. New.
7031 (want_pubnames): Rearrange.
7032 (output_pubnames): Call include_pubname_in_output. Move assertion.
7034 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7036 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7038 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7041 * doc/extend.texi: Remove obsolete builtins. Fix
7042 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7044 2013-07-25 Jan Hubicka <jh@suse.cz>
7046 * cgraph.c (release_function_body): Break out from ...
7047 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7049 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7050 old_tree in the map.
7051 (create_specialized_node): Update.
7052 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
7054 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
7055 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
7056 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
7057 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
7060 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7062 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
7063 addsi3_carryin_alt2_<optab>): Correct output template.
7065 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7067 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
7068 Adjust for arm_restrict_it.
7069 Remove trailing whitespace.
7071 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
7073 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
7074 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
7076 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
7078 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
7080 PR rtl-optimization/57960
7081 * lra-constraints.c (process_alt_operands): Use the right mode
7082 when checking strict_low.
7084 2013-07-25 Jan Hubicka <jh@suse.cz>
7086 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
7087 * cgraph.c (cgraph_remove_node): Do not release function body
7088 when in cgraph streaming.
7089 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
7090 in other partitions are not considered reachable; fix handling of
7093 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7095 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
7097 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7102 * config/arm/arm.md ("*sibcall_insn): Replace use of
7103 Ss with US. Adjust output for v5 and v4t.
7104 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
7106 * config/arm/constraints.md ("Ss"): Rename to US.
7108 2013-07-25 Terry Guo <terry.guo@arm.com>
7110 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
7111 shift_add/shift_sub0/shift_sub1 RTXs.
7113 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
7114 Anton Blanchard <anton@au1.ibm.com>
7116 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
7117 (altivec_vpks<VI_char>ss): Likewise.
7118 (altivec_vpks<VI_char>us): Likewise.
7119 (altivec_vpku<VI_char>us): Likewise.
7120 (altivec_vpku<VI_char>um): Likewise.
7122 2013-07-24 David Malcolm <dmalcolm@redhat.com>
7124 Introduce context class.
7126 * Makefile.in (CONTEXT_H): New.
7127 (OBJS): Add context.o.
7128 (toplev.o): Add CONTEXT_H to dependencies.
7131 * toplev.c (general_init): Create the singleton gcc::context instance.
7137 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
7139 PR rtl-optimization/57968
7140 * mode-switching.c (create_pre_exit): Allow instructions that
7141 don't set a return register to need a non-exit mode.
7143 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7144 Anton Blanchard <anton@au1.ibm.com>
7146 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
7147 operands to vperm for little endian.
7148 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
7149 of lvsl to create the control mask for a vperm for little endian.
7151 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7152 Anton Blanchard <anton@au1.ibm.com>
7154 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
7155 two operands for little-endian.
7157 2013-07-23 Steve Ellcey <sellcey@mips.com>
7159 * config/mips/mips.c (mips_case_values_threshold): New.
7160 (TARGET_CASE_VALUES_THRESHOLD): Define.
7162 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7163 Anton Blanchard <anton@au1.ibm.com>
7165 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
7166 selection of field for vector splat in little endian mode.
7168 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7170 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
7171 expanders to rs6000.md.
7172 (ior<mode>3): Likewise.
7173 (and<mode>3): Likewise.
7174 (one_cmpl<mode>2): Likewise.
7175 (nor<mode>3): Likewise.
7176 (andc<mode>3): Likewise.
7177 (eqv<mode>3): Likewise.
7178 (nand<mode>3): Likewise.
7179 (orc<mode>3): Likewise.
7181 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
7184 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
7185 to split multi-word logical operations.
7186 (rs6000_split_logical_di): Likewise.
7187 (rs6000_split_logical): Likewise.
7189 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
7190 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
7191 and allow TImode operations in 32-bit.
7192 (vsx_and<mode>3_64bit): Likewise.
7193 (vsx_ior<mode>3_32bit): Likewise.
7194 (vsx_ior<mode>3_64bit): Likewise.
7195 (vsx_xor<mode>3_32bit): Likewise.
7196 (vsx_xor<mode>3_64bit): Likewise.
7197 (vsx_one_cmpl<mode>2_32bit): Likewise.
7198 (vsx_one_cmpl<mode>2_64bit): Likewise.
7199 (vsx_nor<mode>3_32bit): Likewise.
7200 (vsx_nor<mode>3_64bit): Likewise.
7201 (vsx_andc<mode>3_32bit): Likewise.
7202 (vsx_andc<mode>3_64bit): Likewise.
7203 (vsx_eqv<mode>3_32bit): Likewise.
7204 (vsx_eqv<mode>3_64bit): Likewise.
7205 (vsx_nand<mode>3_32bit): Likewise.
7206 (vsx_nand<mode>3_64bit): Likewise.
7207 (vsx_orc<mode>3_32bit): Likewise.
7208 (vsx_orc<mode>3_64bit): Likewise.
7210 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
7211 logical types in GPRs.
7213 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
7214 logical insns to rs6000.md, and allow TImode operations in
7216 (altivec_ior<mode>3): Likewise.
7217 (altivec_xor<mode>3): Likewise.
7218 (altivec_one_cmpl<mode>2): Likewise.
7219 (altivec_nor<mode>3): Likewise.
7220 (altivec_andc<mode>3): Likewise.
7222 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
7223 attributes for moving the 128-bit logical operations into
7225 (BOOL_REGS_OUTPUT): Likewise.
7226 (BOOL_REGS_OP1): Likewise.
7227 (BOOL_REGS_OP2): Likewise.
7228 (BOOL_REGS_UNARY): Likewise.
7229 (BOOL_REGS_AND_CR0): Likewise.
7230 (one_cmpl<mode>2): Add support for DI logical operations on
7231 32-bit, splitting the operations to 32-bit.
7235 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
7236 changes to combine the 32/64-bit code, allow logical operations on
7237 TI mode in 32-bit, and to use similar match_operator patterns like
7238 scalar mode uses. Combine the Altivec and VSX code for logical
7239 operations, and move it here.
7240 (ior<mode>3, 128-bit types): Likewise.
7241 (xor<mode>3, 128-bit types): Likewise.
7242 (one_cmpl<mode>3, 128-bit types): Likewise.
7243 (nor<mode>3, 128-bit types): Likewise.
7244 (andc<mode>3, 128-bit types): Likewise.
7245 (eqv<mode>3, 128-bit types): Likewise.
7246 (nand<mode>3, 128-bit types): Likewise.
7247 (orc<mode>3, 128-bit types): Likewise.
7248 (and<mode>3_internal): Likewise.
7249 (bool<mode>3_internal): Likewise.
7250 (boolc<mode>3_internal1): Likewise.
7251 (boolc<mode>3_internal2): Likewise.
7252 (boolcc<mode>3_internal1): Likewise.
7253 (boolcc<mode>3_internal2): Likewise.
7254 (eqv<mode>3_internal1): Likewise.
7255 (eqv<mode>3_internal2): Likewise.
7256 (one_cmpl1<mode>3_internal): Likewise.
7258 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
7260 * config/microblaze/microblaze.c (microblaze_expand_prologue):
7261 Rename flag_stack_usage to flag_stack_usage_info.
7263 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
7265 * config/microblaze/sync.md: New file.
7266 * config/microblaze/microblaze.md: Include sync.md
7267 * config/microblaze/microblaze.c: Add print_operand 'y'.
7268 * config/microblaze/constraints.md: Add memory_contraint
7269 'Q' which is a single register.
7271 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
7273 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
7275 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
7277 * reload.c (find_reloads): Exit loop once we find this operand
7278 cannot be reloaded somehow for this alternative.
7280 * reload.c (find_reloads): Exit loop once we find a hard register.
7282 * rtlanal.c (computed_jump_p): Exit loop once we find label
7285 * i386.c (ix86_pad_returns): Exit loop after setting replace.
7287 * cfgloopmanip.c (remove_path): Exit loop after setting
7290 * gensupport.c (subst_dup): Avoid loop if code is not
7291 MATCH_DUP nor MATCH_OP_DUP.
7293 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
7295 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
7297 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7299 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
7300 true for SP_REGNUM if mode == ptr_mode.
7301 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
7302 with value R0_REGNUM + 31.
7304 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7306 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
7307 pad pointer-typed argument downward.
7309 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7311 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
7312 and __ILP32__ when the ILP32 model is in use.
7314 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7316 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
7317 (aarch64_load_symref_appropriately): In the case of
7318 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
7319 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
7320 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
7321 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
7322 if 'mode' doesn't equal to 'ptr_mode'.
7323 (aarch64_output_mi_thunk): Add an assertion on the alignment of
7324 'vcall_offset'; change to call aarch64_emit_move differently depending
7325 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
7326 to calculate the upper bound of 'vcall_offset'.
7327 (aarch64_cannot_force_const_mem): Change to also return true if
7329 (aarch64_legitimize_reload_address): In the case of large
7330 displacements, add new local variable 'xmode' and an assertion
7331 based on it; change to use 'xmode' to generate the new rtx and
7333 (aarch64_asm_trampoline_template): Change to generate the template
7334 differently depending on TARGET_ILP32 or not; change to use
7335 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
7336 (aarch64_trampoline_size): Removed.
7337 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
7338 and replace immediate literals with it. Change to use 'ptr_mode'
7339 instead of 'DImode' and call convert_memory_address if the mode
7340 of 'fnaddr' doesn't equal to 'ptr_mode'.
7341 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
7343 (aarch64_elf_asm_destructor): Likewise.
7344 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
7345 on TARGET_ILP32 instead of aarch64_trampoline_size.
7346 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
7347 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
7348 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
7349 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
7350 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
7351 (storewb_pair<GPI:mode>_<P:mode>): ... this.
7352 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
7353 depending on the value of 'mode'.
7354 (add_losym_<mode>): New.
7355 (ldr_got_small_<mode>): New, based on ldr_got_small.
7356 (ldr_got_small): Remove.
7357 (ldr_got_small_sidi): New.
7358 * config/aarch64/iterators.md (P): New.
7359 (PTR): Change to 'ptr_mode' in the condition.
7361 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7363 * config.gcc (aarch64*-*-*): Support --with-abi.
7364 (aarch64*-*-elf): Support --with-multilib-list.
7365 (aarch64*-*-linux*): Likewise.
7366 (supported_defaults): Add abi to aarch64*-*-*.
7367 * configure.ac: Mention AArch64 for --with-multilib-list.
7368 * configure: Re-generated.
7369 * config/aarch64/biarchilp32.h: New file.
7370 * config/aarch64/biarchlp64.h: New file.
7371 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
7373 (MULTILIB_DEFAULTS): Ditto.
7374 (DRIVER_SELF_SPECS): Ditto.
7375 (ASM_SPEC): Update to also substitute -mabi.
7376 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
7377 file whose name depends on -mabi= and -mbig-endian.
7378 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
7380 (POINTER_SIZE): New define.
7381 (POINTERS_EXTEND_UNSIGNED): Ditto.
7382 (enum aarch64_abi_type): New enumeration tag.
7383 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
7384 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
7385 (TARGET_ILP32): New define.
7386 * config/aarch64/aarch64.opt (mabi): New.
7388 (ilp32, lp64): New values for -mabi.
7389 * config/aarch64/t-aarch64 (comma): New define.
7390 (MULTILIB_OPTIONS): Ditto.
7391 (MULTILIB_DIRNAMES): Ditto.
7392 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
7393 * doc/invoke.texi: Document -mabi for AArch64.
7395 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
7397 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
7399 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7400 Anton Blanchard <anton@au1.ibm.com>
7402 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
7403 endianness when selecting field to splat.
7405 2013-07-22 Eric Christopher <echristo@gmail.com>
7407 * dwarf2out.c (die_odr_checksum): New function to use
7408 CHECKSUM_ macros and ULEB128 for DIE tag.
7409 (generate_type_signature): Use.
7411 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
7413 * config.gcc (sparc*-*-*): Accept leon3 processor.
7414 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
7415 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
7416 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
7417 * config/sparc/sparc.opt (enum processor_type): Add leon3.
7418 (mfix-ut699): Adjust comment.
7419 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
7420 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
7421 (CPP_CPU_SPEC): Likewise.
7422 (ASM_CPU_SPEC): Likewise.
7423 * config/sparc/sparc.c (leon3_cost): New constant.
7424 (sparc_option_override): Add leon3 support.
7425 (mem_ref): New function.
7426 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
7427 (sparc_do_work_around_errata): Look into the instruction in the delay
7428 slot and adjust accordingly. Add fix for the data cache nullify issues
7429 of the UT699. Change insertion position for the NOP.
7430 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
7431 (leon3_load): New reservation.
7432 (leon_store): Bump latency to 2.
7433 (grfpu): New automaton.
7434 (grfpu_alu): New unit.
7435 (grfpu_ds): Likewise.
7436 (leon_fp_alu): Adjust.
7437 (leon_fp_mult): Delete.
7438 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
7439 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
7440 * config/sparc/sparc.md (cpu): Add leon3.
7441 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
7443 (atomic_test_and_set): Likewise.
7446 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
7448 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
7449 default for R5900 targets.
7450 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
7451 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
7452 * config/mips/mips.c (mips_option_override): Report an error for
7453 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
7454 for -march=r5900 -mhard-float.
7456 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
7458 * df-problems.c (can_move_insns_across): Exit loop once we
7459 find a non-fixed, non-global register.
7461 * ipa-pure-const.c (propagate_nothrow): Exit loop after
7464 * omega.c (omega_eliminate_red): Break after setting red_found.
7465 (omega_problem_has_red_equations): Similarly after setting found.
7466 (omega_query_variable): Similarly after setting coupled.
7468 2013-07-22 Marek Polacek <polacek@redhat.com>
7470 * gimplify.c: Don't include gimple.h twice.
7472 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7474 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
7475 instead of TARGET_THUMB1.
7476 (Pz): New constraint.
7477 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
7479 (compare_negsi_si): Likewise.
7480 (compare_addsi2_op0): Likewise.
7481 (compare_addsi2_op1): Likewise.
7482 (addsi3_carryin_<optab>): Likewise.
7483 (addsi3_carryin_alt2_<optab>): Likewise.
7484 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
7485 for arm_restrict_it.
7486 (subsi3_carryin): Likewise.
7487 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
7488 (minmax_arithsi): Disable for arm_restrict_it.
7489 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
7490 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
7491 (satsi_<SAT:code>_shift): Likewise.
7492 (arm_shiftsi3): Add alternative for 16-bit encoding.
7493 (arm32_movhf): Disable for arm_restrict_it.
7494 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
7495 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
7497 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
7499 * config/arm/arm.md (attribute "insn"): Delete.
7500 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
7501 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
7502 (not_shiftsi): Update for attribute change.
7503 (not_shiftsi_compare0): Likewise.
7504 (not_shiftsi_compare0_scratch): Likewise.
7505 (arm_one_cmplsi2): Likewise.
7506 (thumb1_one_cmplsi2): Likewise.
7507 (notsi_compare0): Likewise.
7508 (notsi_compare0_scratch): Likewise.
7509 (thumb1_movdi_insn): Likewise.
7510 (arm_movsi_insn): Likewise.
7511 (movhi_insn_arch4): Likewise.
7512 (movhi_bytes): Likewise.
7513 (arm_movqi_insn): Likewise.
7514 (thumb1_movqi_insn): Likewise.
7515 (arm32_movhf): Likewise.
7516 (thumb1_movhf): Likewise.
7517 (arm_movsf_soft_insn): Likewise.
7518 (thumb1_movsf_insn): Likewise.
7519 (thumb_movdf_insn): Likewise.
7520 (movsicc_insn): Likewise.
7521 (movsfcc_soft_insn): Likewise.
7522 (and_scc): Likewise.
7523 (cond_move): Likewise.
7524 (if_move_not): Likewise.
7525 (if_not_move): Likewise.
7526 (if_shift_move): Likewise.
7527 (if_move_shift): Likewise.
7528 (if_shift_shift): Likewise.
7529 (if_not_arith): Likewise.
7530 (if_arith_not): Likewise.
7531 (cond_move_not): Likewise.
7532 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
7533 (neon_mov<mode>): Likewise.
7534 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
7535 (thumb2_movsi_vfp): Likewise.
7536 (movsf_vfp): Likewise.
7537 (thumb2_movsf_vfp): Likewise.
7538 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7540 (cortexa7_older_only): Likewise.
7541 (cortexa7_younger): Likewise.
7542 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
7543 (1020alu_shift_op): Likewise.
7544 (1020alu_shift_reg_op): Likewise.
7545 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
7546 (alu_shift_op): Likewise.
7547 (alu_shift_reg_op): Likewise.
7548 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
7549 (11_alu_shift_op): Likewise.
7550 (11_alu_shift_reg_op): Likewise.
7551 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
7552 (9_alu_shift_reg_op): Likewise.
7553 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7555 (cortex_a15_alu_shift): Likewise.
7556 (cortex_a15_alu_shift_reg): Likewise.
7557 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
7558 (cortex_a5_alu_shift): Likewise.
7559 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
7561 (cortex_a53_alu_shift): Likewise.
7562 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7564 (cortex_a7_alu_reg): Likewise.
7565 (cortex_a7_alu_shift): Likewise.
7566 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7567 (cortex_a8_alu_shift): Likewise.
7568 (cortex_a8_alu_shift_reg): Likewise.
7569 (cortex_a8_mov): Likewise.
7570 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7571 (cortex_a9_dp_shift): Likewise.
7572 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
7573 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
7574 (cortex_r4_mov): Likewise.
7575 (cortex_r4_alu_shift): Likewise.
7576 (cortex_r4_alu_shift_reg): Likewise.
7577 * config/arm/fa526.md (526_alu_op): Update for attribute change.
7578 (526_alu_shift_op): Likewise.
7579 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
7580 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
7581 (626te_alu_shift_op): Likewise.
7582 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
7583 (726te_alu_op): Likewise.
7584 (726te_alu_shift_op): Likewise.
7585 (726te_alu_shift_reg_op): Likewise.
7586 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
7587 (mp626_alu_shift_op): Likewise.
7588 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
7589 (pj4_alu_e1_conds): Likewise.
7590 (pj4_alu): Likewise.
7591 (pj4_alu_conds): Likewise.
7592 (pj4_shift): Likewise.
7593 (pj4_shift_conds): Likewise.
7594 (pj4_alu_shift): Likewise.
7595 (pj4_alu_shift_conds): Likewise.
7597 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7599 * config/arm/predicates.md (shiftable_operator_strict_it):
7601 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
7602 Disable cond_exec version for arm_restrict_it.
7603 (thumb2_smaxsi3): Convert to generate cond_exec.
7604 (thumb2_sminsi3): Likewise.
7605 (thumb32_umaxsi3): Likewise.
7606 (thumb2_uminsi3): Likewise.
7607 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
7608 (thumb2_neg_abssi2): Likewise.
7609 (thumb2_mov_scc): Add alternative for 16-bit encoding.
7610 (thumb2_movsicc_insn): Adjust alternatives.
7611 (thumb2_mov_negscc): Disable for arm_restrict_it.
7612 (thumb2_mov_negscc_strict_it): New pattern.
7613 (thumb2_mov_notscc_strict_it): New pattern.
7614 (thumb2_mov_notscc): Disable for arm_restrict_it.
7615 (thumb2_ior_scc): Likewise.
7616 (thumb2_ior_scc_strict_it): New pattern.
7617 (thumb2_cond_move): Adjust for arm_restrict_it.
7618 (thumb2_cond_arith): Disable for arm_restrict_it.
7619 (thumb2_cond_arith_strict_it): New pattern.
7620 (thumb2_cond_sub): Adjust for arm_restrict_it.
7621 (thumb2_movcond): Likewise.
7622 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
7623 (thumb2_zero_extendhisi2_v6): Likewise.
7624 (thumb2_zero_extendqisi2_v6): Likewise.
7625 (orsi_notsi_si): Likewise.
7626 (orsi_not_shiftsi_si): Likewise.
7628 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
7630 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
7631 instruction sequence is 1 byte shorter.
7633 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
7635 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
7636 it is not needed after split.
7638 2013-07-20 Iain Sandoe <iain@codesourcery.com>
7641 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
7642 second label for nonlocal goto receivers. Don't output pic base labels
7643 unless we're producing PIC; mark that action unreachable().
7644 (ix86_save_reg): If the function contains a nonlocal label, save the
7646 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
7647 * config/darwin.c (emitted_pic_label_num): New GTY.
7648 (update_pic_label_number_if_needed): New.
7649 (machopic_output_function_base_name): Adjust for nonlocal receiver
7651 (machopic_should_output_picbase_label): New.
7652 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
7653 (nonlocal_goto_receiver): New insn and split.
7655 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
7657 * config/aarch64/aarch64-builtins.c
7658 (aarch64_fold_builtin): Fold abs in all modes.
7659 * config/aarch64/aarch64-simd-builtins.def
7660 (abs): Enable for all modes.
7661 * config/aarch64/arm_neon.h
7662 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
7663 (vabs_f64): Add missing intrinsic.
7665 2013-07-19 Ian Bolton <ian.bolton@arm.com>
7667 * config/aarch64/arm_neon.h (vabs_s64): New function
7669 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
7672 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
7673 * config/avr/avr.md (adjust_len): Add `round'.
7674 * config/avr/avr-protos.h (avr_out_round): New prototype.
7675 (avr_out_plus): Add `out_label' argument.
7676 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
7677 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
7678 Handle the case where `insn' is just a pattern.
7679 (avr_out_bitop): Handle the case where `insn' is just a pattern.
7680 (avr_out_round): New function.
7681 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
7683 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
7685 * config/microblaze/microblaze.c (microblaze_expand_prologue):
7686 Add check for flag_stack_usage to handle -fstack-usage support
7688 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
7690 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
7691 interaction for new Power8 flags and VSX.
7693 2013-07-18 Sriraman Tallam <tmsriram@google.com>
7696 * tree-inline.c (expand_call_inline): Emit errors during
7697 early_inlining only if optimization is not turned on.
7699 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7703 * passes.c (init_optimization_passes): Move the construction of
7704 the pass hierarchy into a new passes.def file.
7706 * Makefile.in (passes.o): Add dependency on passes.def.
7708 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7710 * passes.c (init_optimization_passes): Introduce macros for
7711 constructing the tree of passes (INSERT_PASSES_AFTER,
7712 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
7713 TERMINATE_PASS_LIST).
7715 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
7716 Wei Mi <wmi@google.com>
7718 PR rtl-optimization/57878
7719 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
7721 (reload_pseudo_compare_func): Check nregs first for reload
7724 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7726 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
7728 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
7730 * read-rtl.c (validate_const_int): Once an invalid character is
7731 seen, quit the loop.
7733 * gengtype.c (write_roots): Similarly once we find the "deletable"
7734 or "if_marked" option.
7736 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
7738 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
7739 "xtab" and "sat". Move value "clz" from here to ...
7740 (attriubte "type"): ... here.
7741 (satsi_<SAT:code>): Delete "insn" attribute.
7742 (satsi_<SAT:code>_shift): Likewise.
7743 (arm_zero_extendqisi2addsi): Likewise.
7744 (arm_extendqisi2addsi): Likewise.
7745 (clzsi2): Update for attribute changes.
7746 (rbitsi2): Likewise.
7747 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
7749 (arm_usatsihi): Likewise.
7750 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7752 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
7754 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
7755 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
7756 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
7757 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
7758 in alphabetical order.
7759 (attribute "core_cycles"): Update for attribute changes.
7760 (arm_addsi3): Likewise.
7761 (addsi3_compare0): Likewise.
7762 (addsi3_compare0_scratch): Likewise.
7763 (addsi3_compare_op1): Likewise.
7764 (addsi3_compare_op2): Likewise.
7765 (compare_addsi2_op0): Likewise.
7766 (compare_addsi2_op1): Likewise.
7767 (addsi3_carryin_shift_<optab>): Likewise.
7768 (subsi3_carryin_shift): Likewise.
7769 (rsbsi3_carryin_shift): Likewise.
7770 (arm_subsi3_insn): Likewise.
7771 (subsi3_compare0): Likewise.
7772 (subsi3_compare): Likewise.
7773 (arm_andsi3_insn): Likewise.
7774 (thumb1_andsi3_insn): Likewise.
7775 (andsi3_compare0): Likewise.
7776 (andsi3_compare0_scratch): Likewise.
7777 (zeroextractsi_compare0_scratch
7778 (andsi_not_shiftsi_si): Likewise.
7779 (iorsi3_insn): Likewise.
7780 (iorsi3_compare0): Likewise.
7781 (iorsi3_compare0_scratch): Likewise.
7782 (arm_xorsi3): Likewise.
7783 (thumb1_xorsi3_insn): Likewise.
7784 (xorsi3_compare0): Likewise.
7785 (xorsi3_compare0_scratch): Likewise.
7786 (satsi_<SAT:code>_shift): Likewise.
7788 (arm_shiftsi3): Likewise.
7789 (shiftsi3_compare0): Likewise.
7790 (not_shiftsi): Likewise.
7791 (not_shiftsi_compare0): Likewise.
7792 (not_shiftsi_compare0_scratch): Likewise.
7793 (arm_one_cmplsi2): Likewise.
7794 (thumb_one_complsi2): Likewise.
7795 (notsi_compare0): Likewise.
7796 (notsi_compare0_scratch): Likewise.
7797 (thumb1_zero_extendhisi2): Likewise.
7798 (arm_zero_extendhisi2): Likewise.
7799 (arm_zero_extendhisi2_v6): Likewise.
7800 (arm_zero_extendhisi2addsi): Likewise.
7801 (thumb1_zero_extendqisi2): Likewise.
7802 (thumb1_zero_extendqisi2_v6): Likewise.
7803 (arm_zero_extendqisi2): Likewise.
7804 (arm_zero_extendqisi2_v6): Likewise.
7805 (arm_zero_extendqisi2addsi): Likewise.
7806 (thumb1_extendhisi2): Likewise.
7807 (arm_extendhisi2): Likewise.
7808 (arm_extendhisi2_v6): Likewise.
7809 (arm_extendqisi): Likewise.
7810 (arm_extendqisi_v6): Likewise.
7811 (arm_extendqisi2addsi): Likewise.
7812 (thumb1_extendqisi2): Likewise.
7813 (thumb1_movdi_insn): Likewise.
7814 (arm_movsi_insn): Likewise.
7815 (movsi_compare0): Likewise.
7816 (movhi_insn_arch4): Likewise.
7817 (movhi_bytes): Likewise.
7818 (arm_movqi_insn): Likewise.
7819 (thumb1_movqi_insn): Likewise.
7820 (arm32_movhf): Likewise.
7821 (thumb1_movhf): Likewise.
7822 (arm_movsf_soft_insn): Likewise.
7823 (thumb1_movsf_insn): Likewise.
7824 (movdf_soft_insn): Likewise.
7825 (thumb_movdf_insn): Likewise.
7826 (arm_cmpsi_insn): Likewise.
7827 (cmpsi_shiftsi): Likewise.
7828 (cmpsi_shiftsi_swp): Likewise.
7829 (arm_cmpsi_negshiftsi_si): Likewise.
7830 (movsicc_insn): Likewise.
7831 (movsfcc_soft_insn): Likewise.
7832 (arith_shiftsi): Likewise.
7833 (arith_shiftsi_compare0
7834 (arith_shiftsi_compare0_scratch
7835 (sub_shiftsi): Likewise.
7836 (sub_shiftsi_compare0
7837 (sub_shiftsi_compare0_scratch
7838 (and_scc): Likewise.
7839 (cond_move): Likewise.
7840 (if_plus_move): Likewise.
7841 (if_move_plus): Likewise.
7842 (if_move_not): Likewise.
7843 (if_not_move): Likewise.
7844 (if_shift_move): Likewise.
7845 (if_move_shift): Likewise.
7846 (if_shift_shift): Likewise.
7847 (if_not_arith): Likewise.
7848 (if_arith_not): Likewise.
7849 (cond_move_not): Likewise.
7850 (thumb1_ashlsi3): Set type attribute.
7851 (thumb1_ashrsi3): Likewise.
7852 (thumb1_lshrsi3): Likewise.
7853 (thumb1_rotrsi3): Likewise.
7854 (shiftsi3_compare0_scratch): Likewise.
7855 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
7856 (neon_mov<mode>): Likewise.
7857 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
7859 (thumb2_movsi_insn): Likewise.
7860 (thumb2_cmpsi_neg_shiftsi): Likewise.
7861 (thumb2_extendqisi_v6): Likewise.
7862 (thumb2_zero_extendhisi2_v6): Likewise.
7863 (thumb2_zero_extendqisi2_v6): Likewise.
7864 (thumb2_shiftsi3_short): Likewise.
7865 (thumb2_addsi3_compare0_scratch): Likewise.
7866 (orsi_not_shiftsi_si): Likewise.
7867 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7868 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
7870 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7871 (1020alu_shift_op): Likewise.
7872 (1020alu_shift_reg_op): Likewise.
7873 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7874 (alu_shift_op): Likewise.
7875 (alu_shift_reg_op): Likewise.
7876 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7877 (11_alu_shift_op): Likewise.
7878 (11_alu_shift_reg_op): Likewise.
7879 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7880 (9_alu_shift_reg_op): Likewise.
7881 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7883 (cortex_a15_alu_shift): Likewise.
7884 (cortex_a15_alu_shift_reg): Likewise.
7885 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
7887 (cortex_a5_alu_shift): Likewise.
7888 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
7890 (cortex_a53_alu_shift): Likewise.
7891 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7893 (cortex_a7_alu_reg): Likewise.
7894 (cortex_a7_alu_shift): Likewise.
7895 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
7897 (cortex_a8_alu_shift): Likewise.
7898 (cortex_a8_alu_shift_reg): Likewise.
7899 (cortex_a8_mov): Likewise.
7900 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7901 (cortex_a9_dp_shift): Likewise.
7902 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
7904 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
7906 (cortex_r4_mov): Likewise.
7907 (cortex_r4_alu_shift): Likewise.
7908 (cortex_r4_alu_shift_reg): Likewise.
7909 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7910 (526_alu_shift_op): Likewise.
7911 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7912 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7913 (626te_alu_shift_op): Likewise.
7914 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
7915 (726te_alu_op): Likewise.
7916 (726te_alu_shift_op): Likewise.
7917 (726te_alu_shift_reg_op): Likewise.
7918 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7919 (mp626_alu_shift_op): Likewise.
7920 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
7921 (pj4_alu_e1_conds): Likewise.
7922 (pj4_alu): Likewise.
7923 (pj4_alu_conds): Likewise.
7924 (pj4_shift): Likewise.
7925 (pj4_shift_conds): Likewise.
7926 (pj4_alu_shift): Likewise.
7927 (pj4_alu_shift_conds): Likewise.
7928 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7930 (cortexa7_older_only): Likewise.
7931 (cortexa7_younger): Likewise.
7933 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7935 * ipa-pure-const.c (generate_summary): Rename to...
7936 (pure_const_generate_summary): ... this.
7938 2013-07-17 Iain Sandoe <iain@codesourcery.com>
7940 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
7942 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
7945 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
7948 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7950 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
7951 enabled without -march=zEC12.
7952 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
7955 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7957 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
7958 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
7959 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
7960 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
7961 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
7962 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
7963 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
7965 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
7966 and ISA_HAS_NMADD3_NMSUB3.
7967 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
7968 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
7969 (nmsub4<mode>, nmsub3<mode>): Likewise.
7970 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
7972 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7974 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
7975 TARGET_MIPS5400 checking.
7977 2013-07-16 Jakub Jelinek <jakub@redhat.com>
7978 Peter Bergner <bergner@vnet.ibm.com>
7980 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
7981 registers in the comment.
7982 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
7983 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
7984 rather than FIRST_PSEUDO_REGISTERS.
7986 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
7988 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
7989 enable extra ISA flags with TARGET_HTM.
7991 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7993 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
7996 2013-07-15 Cong Hou <congh@google.com>
7998 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
7999 in compare function for sorting.
8001 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8003 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8004 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8005 * config/rs6000/rs6000.opt: Add -mhtm option.
8006 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8007 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8008 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8009 __HTM__ if the HTM instructions are available.
8010 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8011 htm_spr_reg_operand): New define_predicates.
8012 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8013 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8015 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8016 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8017 HTM builtin functions.
8018 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8019 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8020 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8021 (rs6000_builtin_mask_calculate): Likewise.
8022 (rs6000_option_override_internal): Likewise.
8023 (bdesc_htm): Add new HTM builtin support.
8024 (htm_spr_num): New function.
8025 (htm_spr_regno): Likewise.
8026 (rs6000_htm_spr_icode): Likewise.
8027 (htm_expand_builtin): Likewise.
8028 (htm_init_builtins): Likewise.
8029 (rs6000_expand_builtin): Add support for HTM builtin functions.
8030 (rs6000_init_builtins): Likewise.
8031 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8033 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8034 (TARGET_HTM, MASK_HTM): Define macros.
8035 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8036 (FIXED_REGISTERS): Likewise.
8037 (CALL_USED_REGISTERS): Likewise.
8038 (CALL_REALLY_USED_REGISTERS): Likewise.
8039 (REG_ALLOC_ORDER): Likewise.
8040 (enum reg_class): Likewise.
8041 (REG_CLASS_NAMES): Likewise.
8042 (REG_CLASS_CONTENTS): Likewise.
8043 (REGISTER_NAMES): Likewise.
8044 (ADDITIONAL_REGISTER_NAMES): Likewise.
8045 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8046 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8047 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8048 * config/rs6000/htm.md: New file.
8049 * config/rs6000/htmintrin.h: New file.
8050 * config/rs6000/htmxlintrin.h: New file.
8052 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
8054 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8055 Define SYMBOL_TINY_GOT, update comment.
8056 * config/aarch64/aarch64.c
8057 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
8058 (aarch64_expand_mov_immediate): Likewise.
8059 (aarch64_print_operand): Likewise.
8060 (aarch64_classify_symbol): Likewise.
8061 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
8062 (ldr_got_tiny): Define.
8064 2013-07-13 Tobias Grosser <tobias@grosser.es>
8066 PR tree-optimization/54094
8067 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
8068 scheduling dimension for the parallelism check from the polyhedral
8069 information in the AST.
8070 * graphite-dependences.c (carries_deps): Do not assume the schedule is
8073 2013-07-13 Jason Merrill <jason@redhat.com>
8075 * print-tree.c (debug_vec_tree): Use debug_raw.
8076 (debug_raw (vec<tree, va_gc> &)): New.
8077 (debug_raw (vec<tree, va_gc> *)): New.
8078 * tree.h: Declare them.
8080 2013-07-13 Bin Cheng <bin.cheng@arm.com>
8082 * ifcvt.c (ifcvt_after_combine): New static variable.
8083 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
8085 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
8086 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
8087 rest_of_handle_if_after_reload): Pass new argument for if_convert.
8089 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
8091 * config/mips/mips.c (mips_expand_call): Remove empty statement.
8093 2013-07-12 Michael Matz <matz@suse.de>
8096 * convert.c (convert_to_real): Reject non-float inner types.
8098 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
8100 * config/aarch64/aarch64-protos.h
8101 (aarch64_simd_immediate_valid_for_move): Remove.
8102 * config/aarch64/aarch64.c (simd_immediate_info): New member.
8103 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
8105 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
8107 2013-07-11 Steve Ellcey <sellcey@mips.com>
8109 * config/mips/mips.c (mips_conditional_register_usage): Do not
8110 use t[0-7] registers in MIPS16 mode when optimizing for size.
8112 2013-07-11 Sriraman Tallam <tmsriram@google.com>
8114 * config/i386/i386.c (dispatch_function_versions): Fix array
8115 indexing of function_version_info to match actual_versions.
8117 2013-07-11 Teresa Johnson <tejohnson@google.com>
8119 * vec.h (struct va_gc): Move release out-of-line.
8120 (va_gc::release): Call ggc_free on released vec.
8122 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8124 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
8125 Require GOT register as additional operand in UNSPEC.
8126 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
8127 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8128 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8129 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
8130 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
8131 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
8132 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
8134 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
8137 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
8138 name seen by assembler/linker rather if available.
8140 2013-07-11 Andreas Schwab <schwab@suse.de>
8142 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
8144 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
8146 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
8148 2013-07-10 Joseph Myers <joseph@codesourcery.com>
8150 * doc/tm.texi.in: Move hook documentation to ....
8151 * target.def: ... here.
8153 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
8155 * doc/tm.texi: Regenerate.
8157 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
8160 * doc/invoke.texi: Document -Wconditionally-supported.
8162 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8165 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
8168 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8171 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
8172 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
8173 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
8174 Remove duplicate devices.
8175 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
8176 * config/avr/t-multilib: Regenerate.
8177 * config/avr/avr-tables.opt: Regenerate.
8178 * doc/avr-mmcu.texi: Regenerate.
8180 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8183 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
8185 2013-07-10 Graham Stott <graham.stott@btinternet.com>
8187 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
8188 the cost of MULT when optimizing for size.
8190 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8192 * config/cr16/cr16-protos.h: Don't include target.h.
8194 2013-07-09 Joseph Myers <joseph@codesourcery.com>
8196 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
8197 adjust register size for TDmode and TFmode for VSX registers.
8199 2013-07-08 Kai Tietz <ktietz@redhat.com>
8202 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
8203 hook_bool_const_tree_true.
8205 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8207 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
8208 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
8209 * config/s390/s390.md: Define FPR*_REGNUM constants.
8210 Fix FPR2_REGNUM constant (18 -> 17).
8211 ("*trunc<BFP:mode><DFP_ALL:mode>2")
8212 ("*trunc<DFP_ALL:mode><BFP:mode>2")
8213 ("trunc<BFP:mode><DFP_ALL:mode>2")
8214 ("trunc<DFP_ALL:mode><BFP:mode>2")
8215 ("*extend<BFP:mode><DFP_ALL:mode>2")
8216 ("*extend<DFP_ALL:mode><BFP:mode>2")
8217 ("extend<BFP:mode><DFP_ALL:mode>2")
8218 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
8221 2013-07-08 Graham Stott <graham.stott@btinternet.com>
8223 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
8225 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8227 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
8228 and cfun_fpr_bit_p to cfun_fpr_save_p.
8229 (s390_frame_area, s390_register_info, s390_frame_info)
8230 (s390_emit_prologue, s390_emit_epilogue)
8231 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
8233 * config/s390/s390.h: Define *_REGNUM macros for floating point
8236 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
8238 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
8240 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
8242 PR rtl-optimization/57786
8243 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
8244 and break out of the loop when it is set to false.
8246 2013-07-08 Jakub Jelinek <jakub@redhat.com>
8249 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
8250 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
8251 (const_int 63)) 0)).
8252 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
8253 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
8254 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
8256 PR rtl-optimization/57829
8257 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
8258 mask bits outside of mode are just sign-extension from mode to HWI.
8260 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8262 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
8263 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
8264 adjust_address instead of change_address to keep info about alignment.
8265 (emit_strmov): Remove.
8266 (emit_memmov): New function.
8267 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
8268 (expand_movmem_epilogue): Likewise and return updated rtx for
8270 (expand_constant_movmem_prologue): Likewise and return updated rtx for
8271 destination and source.
8272 (decide_alignment): Refactor, handle vector_loop.
8273 (ix86_expand_movmem): Likewise.
8274 (ix86_expand_setmem): Likewise.
8275 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
8277 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
8279 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
8280 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
8282 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
8284 * config/i386/sse.md (sse_movlhps): Change alternative 3
8285 of operand 2 to "m".
8287 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
8290 * config/i386/sse.md (iptr): New mode attribute.
8291 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
8292 (<sse>_vm<plusminus_insn><mode>3): Ditto.
8293 (<sse>_vmmul<mode>3): Ditto.
8294 (<sse>_vmdiv<mode>3): Ditto.
8295 (sse_vmrcpv4sf2): Ditto.
8296 (<sse>_vmsqrt<mode>2): Ditto.
8297 (sse_vmrsqrtv4sf2): Ditto.
8298 (<sse>_vm<code><mode>3): Ditto.
8299 (avx_vmcmp<mode>3): Ditto.
8300 (<sse>_vmmaskcmp<mode>3): Ditto.
8301 (<sse>_comi): Ditto.
8302 (<sse>_ucomi): Ditto.
8303 (*xop_vmfrcz_<mode>): Ditto.
8304 (*fmai_fmadd_<mode>): Ditto.
8305 (*fmai_fmsub_<mode>): Ditto.
8306 (*fmai_fnmadd_<mode>): Ditto.
8307 (*fmai_fnmsub_<mode>): Ditto.
8308 (*fma4i_vmfmadd_<mode>): Ditto.
8309 (*fma4i_vmfmsub_<mode>): Ditto.
8310 (*fma4i_vmfnmadd_<mode>): Ditto.
8311 (*fma4i_vmfnmsub_<mode>): Ditto.
8312 (*xop_vmfrcz_<mode>): Ditto.
8313 (sse_cvtps2pi): Ditto.
8314 (sse_cvttps2pi): Ditto.
8315 (sse_cvtss2si): Ditto.
8316 (sse_cvtss2si_2): Ditto.
8317 (sse_cvtss2siq_2): Ditto.
8318 (sse_cvttss2si): Ditto.
8319 (sse_cvttss2siq): Ditto.
8320 (sse_cvtsd2si): Ditto.
8321 (sse_cvtsd2si_2): Ditto.
8322 (sse_cvtsd2siq_2): Ditto.
8323 (sse_cvttsd2si): Ditto.
8324 (sse_cvttsd2siq): Ditto.
8325 (sse_cvtsd2ss): Ditto.
8326 (sse_cvtss2sd): Ditto.
8327 (avx2_pbroadcast<mode>): Ditto.
8328 (avx2_pbroadcast<mode>_1): Ditto.
8329 (*avx_vperm_broadcast_v4sf): Ditto.
8331 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
8332 (sse_movlhps): Ditto.
8333 (sse_storehps): Ditto.
8334 (sse_loadhps): Ditto.
8335 (sse_storelps): Ditto.
8336 (sse_loadlps): Ditto.
8337 (*vec_concatv4sf): Ditto.
8338 (*vec_interleave_highv2df): Ditto.
8339 (*vec_interleave_lowv2df): Ditto.
8340 (*vec_extractv2df_1_sse): Ditto.
8341 (*vec_extractv2df_0_sse): Ditto.
8342 (sse2_storelpd): Ditto.
8343 (sse2_loadlpd): Ditto.
8344 (sse2_movsd): Ditto.
8345 (*vec_concatv4si): Ditto.
8346 (vec_concatv2di): Ditto.
8348 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
8349 for Intel asm dialect.
8350 (mmx_punpcklwd): Ditto.
8351 (mmx_punpckldq): Ditto.
8353 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
8354 for intel assembler dialect.
8356 2013-07-06 Jakub Jelinek <jakub@redhat.com>
8359 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
8360 for BUILT_IN_C{LZ,LRSB}*.
8361 * tree.h (CASE_INT_FN): Add FN##IMAX case.
8362 * tree-vrp.c (extract_range_basic): Handle
8363 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
8364 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
8365 fall thru to code calling set_value*.
8366 * builtins.c (expand_builtin): Remove *IMAX cases.
8367 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
8368 if width is bigger than 2*HWI.
8370 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
8372 PR rtl-optimization/55342
8373 * lra-int.h (lra_subreg_reload_pseudos): New.
8374 * lra.c: Add undoing optional reloads to the block diagram.
8375 (lra_subreg_reload_pseudos): New.
8376 (lra_optional_reload_pseudos): Change comments.
8377 (lra): Init and clear lra_subreg_reload_pseudos. Clear
8378 lra_optional_reload_pseudos after undo transformations.
8379 * lra-assigns.c (pseudo_prefix_title): New.
8380 (lra_setup_reg_renumber): Use it.
8381 (spill_for): Ditto. Check subreg reload pseudos too.
8382 (assign_by_spills): Consider subreg reload pseudos too.
8383 * lra-constraints.c (simplify_operand_subreg): Use
8384 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
8385 (curr_insn_transform): Recognize and do optional reloads.
8386 (undo_optional_reloads): New.
8387 (lra_undo_inheritance): Call undo_optional_reloads.
8389 2013-07-05 Thomas Quinot <quinot@adacore.com>
8391 * tree-complex.c (expand_complex_operations_1): Fix typo.
8393 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
8395 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
8396 (tune_params): New member 'const vec_costs'.
8397 * config/aarch64/aarch64.c (generic_vector_cost): New.
8398 (generic_tunings): New member 'generic_vector_cost'.
8399 (aarch64_builtin_vectorization_cost): New.
8400 (aarch64_add_stmt_cost): New.
8401 (TARGET_VECTORIZE_ADD_STMT_COST): New.
8402 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
8404 2013-07-03 Jakub Jelinek <jakub@redhat.com>
8407 * config/i386/predicates.md (vsib_address_operand): Disallow
8408 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
8410 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
8413 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
8414 expand_builtin_setjmp_receiver.
8415 (expand_label): Adjust, call expand_builtin_setjmp_receiver
8416 with NULL for the label parameter.
8417 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
8418 the frame-pointer. Adjust comments.
8419 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
8420 only if LABEL is non-NULL.
8422 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
8424 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
8425 (ARM_ABI_AAPCS64): Ditto.
8427 (ARM_DEFAULT_ABI): Ditto.
8429 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
8431 * config/aarch64/aarch64-builtins.c
8432 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
8433 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
8435 * config/aarch64/aarch64-simd.md
8436 (aarch64_ld1<VALL:mode>): New.
8437 (aarch64_st1<VALL:mode>): Likewise.
8438 * config/aarch64/arm_neon.h
8439 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
8441 2013-07-02 Sriraman Tallam <tmsriram@google.com>
8443 * config/i386/i386.c (gate_insert_vzeroupper): Check if
8445 (ix86_option_override_internal):Turn on all -mavx target flags by
8446 default as they are dependent on AVX anyway.
8448 2013-07-02 Cary Coutant <ccoutant@google.com>
8450 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
8452 (loc_checksum_ordered): Likewise.
8453 (hash_loc_operands): Remove inline keyword.
8455 2013-07-02 Jakub Jelinek <jakub@redhat.com>
8457 PR tree-optimization/57741
8458 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
8459 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
8460 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
8461 Allow REAL_CST step_exprs if flag_associative_math.
8462 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
8464 2013-07-02 Ian Bolton <ian.bolton@arm.com>
8466 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
8468 2013-07-02 Ian Bolton <ian.bolton@arm.com>
8470 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
8472 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8474 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
8476 (iorsi3_insn): Likewise.
8477 (arm_xorsi3): Likewise.
8479 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
8481 * arm.md (attribute "wtype"): Delete. Move attribute values from here
8483 (attribute "type"): ... here, and prefix with "wmmx_".
8484 (attribute "core_cycles"): Update for attribute changes.
8485 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
8486 (tbcstv4hi): Likewise.
8487 (tbcstv2si): Likewise.
8488 (iwmmxt_iordi3): Likewise.
8489 (iwmmxt_xordi3): Likewise.
8490 (iwmmxt_anddi3): Likewise.
8491 (iwmmxt_nanddi3): Likewise.
8492 (iwmmxt_arm_movdi): Likewise.
8493 (iwmmxt_movsi_insn): Likewise.
8494 (mov<mode>_internal): Likewise.
8495 (and<mode>3_iwmmxt): Likewise.
8496 (ior<mode>3_iwmmxt): Likewise.
8497 (xor<mode>3_iwmmxt): Likewise.
8498 (add<mode>3_iwmmxt): Likewise.
8499 (ssaddv8qi3): Likewise.
8500 (ssaddv4hi3): Likewise.
8501 (ssaddv2si3): Likewise.
8502 (usaddv8qi3): Likewise.
8503 (usaddv4hi3): Likewise.
8504 (usaddv2si3): Likewise.
8505 (sub<mode>3_iwmmxt): Likewise.
8506 (sssubv8qi3): Likewise.
8507 (sssubv4hi3): Likewise.
8508 (sssubv2si3): Likewise.
8509 (ussubv8qi3): Likewise.
8510 (ussubv4hi3): Likewise.
8511 (ussubv2si3): Likewise.
8512 (mulv4hi3_iwmmxt): Likewise.
8513 (smulv4hi3_highpart): Likewise.
8514 (umulv4hi3_highpart): Likewise.
8515 (iwmmxt_wmacs): Likewise.
8516 (iwmmxt_wmacsz): Likewise.
8517 (iwmmxt_wmacu): Likewise.
8518 (iwmmxt_wmacuz): Likewise.
8519 (iwmmxt_clrdi): Likewise.
8520 (iwmmxt_clrv8qi): Likewise.
8521 (iwmmxt_clr4hi): Likewise.
8522 (iwmmxt_clr2si): Likewise.
8523 (iwmmxt_uavgrndv8qi3): Likewise.
8524 (iwmmxt_uavgrndv4hi3): Likewise.
8525 (iwmmxt_uavgv8qi3): Likewise.
8526 (iwmmxt_uavgv4hi3): Likewise.
8527 (iwmmxt_tinsrb): Likewise.
8528 (iwmmxt_tinsrh): Likewise.
8529 (iwmmxt_tinsrw): Likewise.
8530 (iwmmxt_textrmub): Likewise.
8531 (iwmmxt_textrmsb): Likewise.
8532 (iwmmxt_textrmuh): Likewise.
8533 (iwmmxt_textrmsh): Likewise.
8534 (iwmmxt_textrmw): Likewise.
8535 (iwmxxt_wshufh): Likewise.
8536 (eqv8qi3): Likewise.
8537 (eqv4hi3): Likewise.
8538 (eqv2si3): Likewise.
8539 (gtuv8qi3): Likewise.
8540 (gtuv4hi3): Likewise.
8541 (gtuv2si3): Likewise.
8542 (gtv8qi3): Likewise.
8543 (gtv4hi3): Likewise.
8544 (gtv2si3): Likewise.
8545 (smax<mode>3_iwmmxt): Likewise.
8546 (umax<mode>3_iwmmxt): Likewise.
8547 (smin<mode>3_iwmmxt): Likewise.
8548 (umin<mode>3_iwmmxt): Likewise.
8549 (iwmmxt_wpackhss): Likewise.
8550 (iwmmxt_wpackwss): Likewise.
8551 (iwmmxt_wpackdss): Likewise.
8552 (iwmmxt_wpackhus): Likewise.
8553 (iwmmxt_wpackwus): Likewise.
8554 (iwmmxt_wpackdus): Likewise.
8555 (iwmmxt_wunpckihb): Likewise.
8556 (iwmmxt_wunpckihh): Likewise.
8557 (iwmmxt_wunpckihw): Likewise.
8558 (iwmmxt_wunpckilb): Likewise.
8559 (iwmmxt_wunpckilh): Likewise.
8560 (iwmmxt_wunpckilw): Likewise.
8561 (iwmmxt_wunpckehub): Likewise.
8562 (iwmmxt_wunpckehuh): Likewise.
8563 (iwmmxt_wunpckehuw): Likewise.
8564 (iwmmxt_wunpckehsb): Likewise.
8565 (iwmmxt_wunpckehsh): Likewise.
8566 (iwmmxt_wunpckehsw): Likewise.
8567 (iwmmxt_wunpckelub): Likewise.
8568 (iwmmxt_wunpckeluh): Likewise.
8569 (iwmmxt_wunpckeluw): Likewise.
8570 (iwmmxt_wunpckelsb): Likewise.
8571 (iwmmxt_wunpckelsh): Likewise.
8572 (iwmmxt_wunpckelsw): Likewise.
8573 (ror<mode>3): Likewise.
8574 (ashr<mode>3_iwmmxt): Likewise.
8575 (lshr<mode>3_iwmmxt): Likewise.
8576 (ashl<mode>3_iwmmxt): Likewise.
8577 (ror<mode>3_di): Likewise.
8578 (ashr<mode>3_di): Likewise.
8579 (lshr<mode>3_di): Likewise.
8580 (ashl<mode>3_di): Likewise.
8581 (iwmmxt_wmadds): Likewise.
8582 (iwmmxt_wmaddu): Likewise.
8583 (iwmmxt_tmia): Likewise.
8584 (iwmmxt_tmiaph): Likewise.
8585 (iwmmxt_tmiabb): Likewise.
8586 (iwmmxt_tmiatb): Likewise.
8587 (iwmmxt_tmiabt): Likewise.
8588 (iwmmxt_tmiatt): Likewise.
8589 (iwmmxt_tmovmskb): Likewise.
8590 (iwmmxt_tmovmskh): Likewise.
8591 (iwmmxt_tmovmskw): Likewise.
8592 (iwmmxt_waccb): Likewise.
8593 (iwmmxt_wacch): Likewise.
8594 (iwmmxt_waccw): Likewise.
8595 (iwmmxt_waligni): Likewise.
8596 (iwmmxt_walignr): Likewise.
8597 (iwmmxt_walignr0): Likewise.
8598 (iwmmxt_walignr1): Likewise.
8599 (iwmmxt_walignr2): Likewise.
8600 (iwmmxt_walignr3): Likewise.
8601 (iwmmxt_wsadb): Likewise.
8602 (iwmmxt_wsadh): Likewise.
8603 (iwmmxt_wsadbz): Likewise.
8604 (iwmmxt_wsadhz): Likewise.
8605 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
8606 (iwmmxt_wabsdiffb): Likewise.
8607 (iwmmxt_wabsdiffh): Likewise.
8608 (iwmmxt_wabsdiffw): Likewise.
8609 (iwmmxt_waddsubhx): Likewise
8610 (iwmmxt_wsubaddhx): Likewise.
8611 (addc<mode>3): Likewise.
8612 (iwmmxt_avg4): Likewise.
8613 (iwmmxt_avg4r): Likewise.
8614 (iwmmxt_wmaddsx): Likewise.
8615 (iwmmxt_wmaddux): Likewise.
8616 (iwmmxt_wmaddsn): Likewise.
8617 (iwmmxt_wmaddun): Likewise.
8618 (iwmmxt_wmulwsm): Likewise.
8619 (iwmmxt_wmulwum): Likewise.
8620 (iwmmxt_wmulsmr): Likewise.
8621 (iwmmxt_wmulumr): Likewise.
8622 (iwmmxt_wmulwsmr): Likewise.
8623 (iwmmxt_wmulwumr): Likewise.
8624 (iwmmxt_wmulwl): Likewise.
8625 (iwmmxt_wqmulm): Likewise.
8626 (iwmmxt_wqmulwm): Likewise.
8627 (iwmmxt_wqmulmr): Likewise.
8628 (iwmmxt_wqmulwmr): Likewise.
8629 (iwmmxt_waddbhusm): Likewise.
8630 (iwmmxt_waddbhusl): Likewise.
8631 (iwmmxt_wqmiabb): Likewise.
8632 (iwmmxt_wqmiabt): Likewise.
8633 (iwmmxt_wqmiatb): Likewise.
8634 (iwmmxt_wqmiatt): Likewise.
8635 (iwmmxt_wqmiabbn): Likewise.
8636 (iwmmxt_wqmiabtn): Likewise.
8637 (iwmmxt_wqmiatbn): Likewise.
8638 (iwmmxt_wqmiattn): Likewise.
8639 (iwmmxt_wmiabb): Likewise.
8640 (iwmmxt_wmiabt): Likewise.
8641 (iwmmxt_wmiatb): Likewise.
8642 (iwmmxt_wmiatt): Likewise.
8643 (iwmmxt_wmiabbn): Likewise.
8644 (iwmmxt_wmiabtn): Likewise.
8645 (iwmmxt_wmiatbn): Likewise.
8646 (iwmmxt_wmiattn): Likewise.
8647 (iwmmxt_wmiawbb): Likewise.
8648 (iwmmxt_wmiawbt): Likewise.
8649 (iwmmxt_wmiawtb): Likewise.
8650 (iwmmxt_wmiawtt): Likewise.
8651 (iwmmxt_wmiawbbn): Likewise.
8652 (iwmmxt_wmiawbtn): Likewise.
8653 (iwmmxt_wmiawtbn): Likewise.
8654 (iwmmxt_wmiawttn): Likewise.
8655 (iwmmxt_wmerge): Likewise.
8656 (iwmmxt_tandc<mode>3): Likewise.
8657 (iwmmxt_torc<mode>3): Likewise.
8658 (iwmmxt_torvsc<mode>3): Likewise.
8659 (iwmmxt_textrc<mode>3): Likewise.
8660 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
8661 (wmmxt_pack): Likewise.
8662 (wmmxt_mult_c1): Likewise.
8663 (wmmxt_mult_c2): Likewise.
8664 (wmmxt_alu_c1): Likewise.
8665 (wmmxt_alu_c2): Likewise.
8666 (wmmxt_alu_c3): Likewise.
8667 (wmmxt_transfer_c1): Likewise.
8668 (wmmxt_transfer_c2): Likewise.
8669 (wmmxt_transfer_c3): Likewise.
8670 (marvell_f_iwmmxt_wstr): Likewise.
8671 (marvell_f_iwmmxt_wldr): Likewise.
8673 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
8675 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
8677 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8680 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8681 * lra-constraints.c (need_for_split_p): Check call used hard regs
8682 living through calls.
8684 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
8685 call used regs for call insn.
8687 2013-06-28 Jakub Jelinek <jakub@redhat.com>
8690 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
8691 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
8693 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
8695 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
8697 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8699 * lra-constraints.c (need_for_split_p): Check call used hard regs
8700 living through calls.
8702 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
8705 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
8706 to tie with any other modes. Eliminate Altivec vector mode tests,
8707 since these are a subset of ALTIVEC or VSX vector modes. Simplify
8708 code, to return 0 if testing MODE2 for a condition, if we've
8709 already tested MODE1 for the same condition.
8711 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8713 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
8716 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8718 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8719 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
8721 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8723 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
8725 (aarch64_symbolic_constant_p): Remove.
8726 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
8727 static. Fix line length and white space.
8728 (aarch64_symbolic_constant_p): Remove.
8729 * config/aarch64/predicates.md (aarch64_valid_symref):
8730 Use aarch64_classify_symbol_expression.
8732 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8734 * config/arm/constraints.md (Ts): New constraint.
8735 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
8737 (compare_scc): Use "Ts" constraint for operand 0.
8738 (ior_scc_scc): Likewise.
8739 (and_scc_scc): Likewise.
8740 (and_scc_scc_nodom): Likewise.
8741 (ior_scc_scc_cmp): Likewise for operand 7.
8742 (and_scc_scc_cmp): Likewise.
8743 * config/arm/thumb2.md (thumb2_movsi_insn):
8744 Add alternatives for 16-bit encodings.
8745 (thumb2_movhi_insn): Likewise.
8746 (thumb2_movsicc_insn): Likewise.
8747 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
8748 (thumb2_negscc): Use "Ts" constraint.
8749 Move mvn instruction outside cond_exec block.
8750 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
8751 for 16-bit encodings.
8753 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8755 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
8757 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
8758 (mulsi3subsi): Likewise.
8759 (mulsidi3adddi): Likewise.
8760 (mulsidi3_v6): Likewise.
8761 (umulsidi3_v6): Likewise.
8762 (umulsidi3adddi_v6): Likewise.
8763 (smulsi3_highpart_v6): Likewise.
8764 (umulsi3_highpart_v6): Likewise.
8765 (mulhisi3tb): Likewise.
8766 (mulhisi3bt): Likewise.
8767 (mulhisi3tt): Likewise.
8768 (maddhisi4): Likewise.
8769 (maddhisi4tb): Likewise.
8770 (maddhisi4tt): Likewise.
8771 (maddhidi4): Likewise.
8772 (maddhidi4tb): Likewise.
8773 (maddhidi4tt): Likewise.
8774 (zeroextractsi_compare0_scratch): Likewise.
8775 (insv_zero): Likewise.
8776 (insv_t2): Likewise.
8777 (anddi_notzesidi_di): Likewise.
8778 (anddi_notsesidi_di): Likewise.
8779 (andsi_notsi_si): Likewise.
8780 (iordi_zesidi_di): Likewise.
8781 (xordi_zesidi_di): Likewise.
8782 (andsi_iorsi3_notsi): Likewise.
8784 (smax_m1): Likewise.
8786 (not_shiftsi): Likewise.
8787 (unaligned_loadsi): Likewise.
8788 (unaligned_loadhis): Likewise.
8789 (unaligned_loadhiu): Likewise.
8790 (unaligned_storesi): Likewise.
8791 (unaligned_storehi): Likewise.
8792 (extv_reg): Likewise.
8793 (extzv_t2): Likewise.
8795 (udivsi3): Likewise.
8796 (arm_zero_extendhisi2addsi): Likewise.
8797 (arm_zero_extendqisi2addsi): Likewise.
8798 (compareqi_eq0): Likewise.
8799 (arm_extendhisi2_v6): Likewise.
8800 (arm_extendqisi2addsi): Likewise.
8801 (arm_movt): Likewise.
8802 (thumb2_ldrd): Likewise.
8803 (thumb2_ldrd_base): Likewise.
8804 (thumb2_ldrd_base_neg): Likewise.
8805 (thumb2_strd): Likewise.
8806 (thumb2_strd_base): Likewise.
8807 (thumb2_strd_base_neg): Likewise.
8808 (arm_negsi2): Add alternative for 16-bit encoding.
8809 (arm_one_cmplsi2): Likewise.
8811 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8813 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
8814 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
8815 (movdfcc): Likewise.
8816 * config/arm/vfp.md (*thumb2_movsf_vfp):
8817 Disable predication for arm_restrict_it.
8818 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
8819 (*thumb2_movdfcc_vfp): Likewise.
8820 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
8821 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
8822 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
8823 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
8824 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
8825 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
8826 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
8827 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
8828 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
8829 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
8830 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
8831 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
8832 Disable predication for arm_restrict_it.
8834 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
8836 * config/i386/bmiintrin.h (_bextr_u32): New.
8837 (_bextr_u64): Ditto.
8839 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
8841 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
8842 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
8843 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
8844 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
8845 * config/mips/n32-elf.h: ...this new file.
8847 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
8850 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
8851 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
8853 2013-06-27 Catherine Moore <clm@codesourcery.com>
8855 * config/mips/mips-tables.opt: Regenerate.
8856 * config/mips/mips-cpus.def: Add m14ke and m14kec.
8857 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8858 * doc/invoke.texi: Add -m14kc.
8860 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8863 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
8864 constraints of operand 1 and 2.
8867 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
8868 to match RTL canonicalization. Swap predicates and
8869 constraints of operand 1 and 2.
8871 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
8873 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
8874 Process OP_INOUT regs for splitting too.
8876 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8878 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
8879 decl before the loop, initialize to NULL.
8880 (vectorizable_load): Initialize ptr_incr to NULL.
8882 2013-06-27 Martin Jambor <mjambor@suse.cz>
8885 * ipa-ref.h (ipa_maybe_record_reference): Declare.
8886 * ipa-ref.c (ipa_maybe_record_reference): New function.
8887 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
8888 * ipa-cp.c (create_specialized_node): Record potential references from
8890 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
8892 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8894 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
8895 parameter 'mode' of type 'enum machine_mode mode'; change to pass
8896 'mode' to force_reg.
8897 (aarch64_add_offset): Update calls to aarch64_force_temporary.
8898 (aarch64_expand_mov_immediate): Likewise.
8900 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8902 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
8903 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
8905 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8907 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
8908 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8909 (struct machine_function): Add tbegin_p.
8910 (s390_canonicalize_comparison): Fold CC mode compares to
8911 conditional jump if possible.
8912 (s390_emit_jump): Return the emitted jump.
8913 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
8914 Handle CCRAWmode compares.
8915 (s390_option_override): Default to -mhtm if available.
8916 (s390_reg_clobbered_rtx): Handle floating point regs as well.
8917 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
8918 FPRs instead of df_regs_ever_live_p.
8919 (s390_optimize_nonescaping_tx): New function.
8920 (s390_init_frame_layout): Extend clobbered_regs array to cover
8922 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
8923 (s390_expand_tbegin): New function.
8924 (enum s390_builtin): New enum definition.
8925 (code_for_builtin): New array definition.
8926 (s390_init_builtins): New function.
8927 (s390_expand_builtin): New function.
8928 (TARGET_INIT_BUILTINS): Define.
8929 (TARGET_EXPAND_BUILTIN): Define.
8930 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
8931 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
8932 (s390_alc_comparison): Likewise.
8933 * config/s390/s390-modes.def: Add CCRAWmode.
8934 * config/s390/s390.h (processor_flags): Add PF_TX.
8935 (TARGET_CPU_HTM): Define macro.
8936 (TARGET_HTM): Define macro.
8937 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
8938 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
8939 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8940 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
8941 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
8942 (TBEGIN_MASK, TBEGINC_MASK): New constants.
8943 ("*cc_to_int"): Move up.
8944 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
8945 constants other than 0.
8946 ("*ccraw_to_int"): New insn and splitter definition.
8947 ("tbegin", "tbegin_nofloat", "tbegin_retry")
8948 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
8949 ("tx_assist"): New expander.
8950 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
8951 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
8952 * config/s390/s390.opt: Add -mhtm option.
8953 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
8954 * config/s390/htmxlintrin.h: New file.
8955 * config/s390/htmintrin.h: New file.
8956 * config/s390/s390intrin.h: New file.
8957 * doc/extend.texi: Document htm builtins.
8958 * config.gcc: Add the new header files to extra_headers.
8960 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
8962 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
8963 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
8965 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
8966 Pat Haugen <pthaugen@us.ibm.com>
8967 Peter Bergner <bergner@vnet.ibm.com>
8969 * config/rs6000/power8.md: New.
8970 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
8971 setting for power8 entry.
8972 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
8973 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
8974 test for Power4/Power5 only.
8975 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
8977 (force_new_group): Adjust comment.
8978 * config/rs6000/rs6000.md: Include power8.md.
8980 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
8982 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
8983 * config/arm/arm-protos.h (arm_max_conditional_execute): New
8985 (tune_params): Update comment.
8986 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
8987 (arm_max_conditional_execute): New function.
8988 (thumb2_final_prescan_insn): Use max_insn_skipped and
8989 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
8991 2013-06-25 Jakub Jelinek <jakub@redhat.com>
8993 PR tree-optimization/57705
8994 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
8995 SSA_NAME step, provided that it is not defined inside the loop.
8996 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
8997 (get_initial_def_for_induction): Handle SSA_NAME IV step.
8999 2013-06-25 Martin Jambor <mjambor@suse.cz>
9002 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9003 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9005 (ipa_note_param_call): Initialize member_ptr flag.
9006 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9007 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9008 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9009 (ipa_read_indirect_edge_info): Likewise.
9011 2013-06-25 Richard Biener <rguenther@suse.de>
9014 * passes.c (init_optimization_passes): Move pass_fold_builtins
9015 and pass_dce earlier with -Og.
9017 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9019 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9020 <BIT_FIELD_REF>: Remove trailing TAB.
9021 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9024 2013-06-24 Martin Jambor <mjambor@suse.cz>
9026 PR tree-optimization/57358
9027 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9028 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9029 (ipa_analyze_params_uses): Generate pessimistic info when true.
9031 2013-06-24 Martin Jambor <mjambor@suse.cz>
9033 PR tree-optimization/57539
9034 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9035 global.inlined_to of the new node to it. All callers changed.
9036 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9037 inlining_into, pass it to cgraph_clone_node.
9038 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9039 ipa_free_edge_args_substructures.
9040 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9041 rdesc linked list. Do not assert rdesc edges have inlined caller.
9042 Assert we have found an rdesc in the rdesc list.
9044 2013-06-24 Richard Biener <rguenther@suse.de>
9046 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9047 (pointer_set_lookup): Declare.
9048 (class pointer_map): New template class implementing a
9049 generic pointer to T map.
9050 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9051 pointer_map<T>::contains, pointer_map<T>::insert,
9052 pointer_map<T>::traverse): New functions.
9053 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
9054 (pointer_set_lookup): New function.
9055 (pointer_set_contains): Use pointer_set_lookup.
9056 (pointer_set_insert): Likewise.
9057 (insert_aux): Remove.
9058 (struct pointer_map_t): Embed a pointer_set_t.
9059 (pointer_map_create): Adjust.
9060 (pointer_map_destroy): Likewise.
9061 (pointer_map_contains): Likewise.
9062 (pointer_map_insert): Likewise.
9063 (pointer_map_traverse): Likewise.
9064 * tree-streamer.h (struct streamer_tree_cache_d): Use a
9065 pointer_map<unsigned> instead of a pointer_map_t.
9066 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
9067 (streamer_tree_cache_lookup): Likewise.
9068 (streamer_tree_cache_create): Likewise.
9069 (streamer_tree_cache_delete): Likewise.
9070 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
9071 pointer_map<unsigned> instead of a pointer_map_t.
9072 (lto_init_tree_ref_encoder): Adjust.
9073 (lto_destroy_tree_ref_encoder): Likewise.
9074 * lto-section-out.c (lto_output_decl_index): Likewise.
9075 (lto_record_function_out_decl_state): Likewise.
9076 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
9078 2013-06-24 Richard Biener <rguenther@suse.de>
9080 PR tree-optimization/57488
9081 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
9083 2013-06-24 Alan Modra <amodra@gmail.com>
9085 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
9086 (gen_easy_altivec_constant): Likewise.
9087 * config/rs6000/predicates.md (easy_vector_constant_add_self,
9088 easy_vector_constant_msb): Likewise.
9090 2013-06-23 Jakub Jelinek <jakub@redhat.com>
9093 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
9094 add missing return true.
9096 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
9099 * config/sh/predicates.md (general_extend_operand): Invoke
9100 general_movsrc_operand for memory operands.
9101 (general_movsrc_operand): Allow reg+reg addressing, do not use
9102 general_operand for memory operands.
9104 2013-06-23 Sriraman Tallam <tmsriram@google.com>
9106 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
9107 when current target options does not apply.
9108 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
9109 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
9110 * config/i386/bmiintrin.h: Pass appropriate target
9111 attributes to header.
9112 * config/i386/mmintrin.h: Ditto.
9113 * config/i386/nmmintrin.h: Ditto.
9114 * config/i386/avx2intrin.h: Ditto.
9115 * config/i386/fxsrintrin.h: Ditto.
9116 * config/i386/tbmintrin.h: Ditto.
9117 * config/i386/xsaveintrin.h: Ditto.
9118 * config/i386/f16cintrin.h: Ditto.
9119 * config/i386/xtestintrin.h: Ditto.
9120 * config/i386/xsaveoptintrin.h: Ditto.
9121 * config/i386/bmi2intrin.h: Ditto.
9122 * config/i386/lzcntintrin.h: Ditto.
9123 * config/i386/smmintrin.h: Ditto.
9124 * config/i386/wmmintrin.h: Ditto.
9125 * config/i386/x86intrin.h: Remove all header include guards.
9126 * config/i386/prfchwintrin.h: Ditto.
9127 * config/i386/pmmintrin.h: Ditto.
9128 * config/i386/tmmintrin.h: Ditto.
9129 * config/i386/xmmintrin.h: Ditto.
9130 * config/i386/popcntintrin.h: Ditto.
9131 * config/i386/rdseedintrin.h: Ditto.
9132 * config/i386/ammintrin.h: Ditto.
9133 * config/i386/emmintrin.h: Ditto.
9134 * config/i386/immintrin.h: Remove all header include guards.
9135 * config/i386/fma4intrin.h: Ditto.
9136 * config/i386/lwpintrin.h: Ditto.
9137 * config/i386/xopintrin.h: Ditto.
9138 * config/i386/ia32intrin.h: Ditto.
9139 * config/i386/avxintrin.h: Ditto.
9140 * config/i386/rtmintrin.h: Ditto.
9141 * config/i386/fmaintrin.h: Ditto.
9142 * config/i386/mm3dnow.h: Ditto.
9144 2013-06-22 Sriraman Tallam <tmsriram@google.com>
9146 * common/config/i386/i386-common.c: Handle LZCNT.
9148 2013-06-22 Andi Kleen <ak@linux.intel.com>
9150 * doc/extend.texi: Use __atomic_store_n instead of
9151 __atomic_store in HLE example.
9153 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
9155 * config/sh/sh.c: Remove <cstdlib> workaround.
9157 2013-06-21 Andi Kleen <ak@linux.intel.com>
9159 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
9161 2013-06-21 Andi Kleen <ak@linux.intel.com>
9163 * doc/extend.texi: Document that __atomic_clear and
9164 __atomic_test_and_set should only be used with bool.
9166 2013-06-20 Jan Hubicka <jh@suse.cz>
9168 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
9170 * tree.c (decls_same_for_odr): New function.
9171 (same_for_edr): New function.
9172 (types_same_for_odr): New function.
9173 (get_binfo_at_offset): Use it.
9174 * tree.h (types_same_for_odr): Declare.
9176 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
9177 Jason Merrill <jason@redhat.com>
9179 * system.h: Include <cstdlib> as well as <stdlib.h>.
9181 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
9184 * config/i386/i386.c (construct_container): Report error if
9185 long double is used with disabled x87 float returns.
9187 2013-06-20 Jan Hubicka <jh@suse.cz>
9189 * lto-cgraph.c (input_symtab): Do not set cgraph state.
9191 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
9193 PR rtl-optimization/57425
9194 PR rtl-optimization/57569
9195 * alias.c (write_dependence_p): Remove parameters mem_mode and
9196 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
9197 Changed all callers.
9198 (canon_anti_dependence): Get comments and semantics in sync.
9199 Add parameter mem_canonicalized. Changed all callers.
9200 * rtl.h (canon_anti_dependence): Update prototype.
9202 2013-06-20 Richard Biener <rguenther@suse.de>
9204 * data-streamer-in.c (streamer_read_uhwi): Optimize single
9205 byte case, inline streamer_read_uchar and defer section
9208 2013-06-20 Richard Biener <rguenther@suse.de>
9210 PR tree-optimization/57584
9211 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
9212 SSA names into the expanded expression that take part in
9213 abnormal coalescing.
9215 2013-06-19 Sharad Singhai <singhai@google.com>
9217 * gcov.c (print_usage): Handle new option.
9218 (process_args): Ditto.
9219 (get_gcov_intermediate_filename): New function.
9220 (output_intermediate_file): New function.
9221 (output_gcov_file): New function
9222 (generate_results): Handle new option.
9223 (release_function): Relase demangled name.
9224 (read_graph_file): Handle demangled name.
9225 (output_lines): Ditto.
9226 * doc/gcov.texi: Document gcov intermediate format.
9228 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
9231 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
9232 (lra_emit_add): Use the functions. Add comment about Y as an
9235 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
9238 * collect2.c (collect_atexit): New.
9239 (collect_exit): Delete.
9240 (main): Register collect_atexit with atexit.
9241 (collect_wait): Change collect_exit to exit.
9243 * collect2.h (collect_exit): Delete.
9244 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
9246 2013-06-19 Wei Mi <wmi@google.com>
9248 PR rtl-optimization/57518
9249 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
9250 if regno is used in paradoxical subreg.
9251 (update_equiv_regs): Check pdx_subregs[regno] before
9252 set a reg to be equivalent with a mem.
9254 2013-06-19 Matthias Klose <doko@ubuntu.com>
9257 * file-find.h (find_a_file): Add a mode parameter.
9258 * file-find.c (find_a_file): Likewise.
9259 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
9260 with X_OK for the executables.
9261 * collect2.c (main): Call find_a_file with X_OK.
9263 2013-06-19 Steve Ellcey <sellcey@mips.com>
9266 * config/mips/mips.md (casesi_internal_mips16_<mode>):
9267 Use NEXT_INSN instead of next_real_insn.
9269 2013-06-19 Jan Hubicka <jh@suse.cz>
9271 * cgraph.h (const_value_known_p): Replace by ...
9272 (ctor_for_folding): .. this one.
9273 * cgraphunit.c (process_function_and_variable_attributes): Use it.
9274 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
9275 * expr.c (expand_expr_real_1): Likewise.
9276 (string_constant): Likewise.
9277 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
9278 * ipa.c (process_references): Likewise.
9279 (symtab_remove_unreachable_nodes): Likewise.
9280 * ipa-inline-analysis.c (param_change_prob): Likewise.
9281 * gimple-fold.c (canonicalize_constructor_val): Likewise.
9282 (get_base_constructor): Likwise.
9283 * varpool.c (varpool_remove_node): Likewise.
9284 (varpool_remove_initializer): LIkewise.
9285 (dump_varpool_node): LIkwise.
9286 (const_value_known_p): Rewrite to ...
9287 (ctor_for_folding): ... this one.
9289 2013-06-19 Jakub Jelinek <jakub@redhat.com>
9292 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
9293 PERSONALITY in $PATH derived prefixes.
9295 2013-06-19 Jeff Law <law@redhat.com>
9297 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
9300 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
9301 (simplify_bitwise_binary): Use it to simpify certain binary ops on
9304 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
9306 * config/arm/vfp.md: Move VFP instruction classification documentation
9308 * config/arm/arm.md: ... here. Update instruction classification
9311 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
9313 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
9314 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
9315 pattern. Use more efficient sequences on ARMv5 and Thumb2.
9317 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
9320 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9321 with NEXT_INSN. Use tablejump_p to check for jump table data
9324 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
9327 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
9328 that now in C++ the value is correct per the C++ standards.
9330 2013-06-19 Richard Biener <rguenther@suse.de>
9332 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
9335 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9338 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9341 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9342 with next_active_insn.
9344 2013-06-18 Sriraman Tallam <tmsriram@google.com>
9346 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
9347 functions are inlined during failures to flag an error.
9348 * tree-inline.c (expand_call_inline): Allow the error to be flagged
9349 in early inline pass.
9351 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
9353 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
9356 2013-06-18 Julian Brown <julian@codesourcery.com>
9358 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
9359 Permit virtual register pre-reload if !strict.
9360 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
9362 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
9364 * config/arm/neon.md (movmisalign<mode>): Use
9365 neon_perm_struct_or_reg_operand instead of
9366 neon_struct_or_register_operand.
9367 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
9368 neon_permissive_struct_operand instead of neon_struct_operand.
9369 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
9370 neon_vector_mem_operand.
9371 * config/arm/predicates.md (neon_struct_operand): Adjust call to
9372 neon_vector_mem_operand.
9373 (neon_permissive_struct_operand): New.
9374 (neon_struct_or_register_operand): Rename to...
9375 (neon_perm_struct_or_reg_operand): This. Adjust call to
9376 neon_vector_mem_operand.
9378 2013-06-18 Richard Biener <rguenther@suse.de>
9380 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
9381 * lto-streamer.h: Include pointer-set.h.
9382 (struct lto_decl_slot): Remove.
9383 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
9384 Remove next_index entry.
9385 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
9386 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
9387 (lto_init_tree_ref_encoder): Adjust.
9388 (lto_destroy_tree_ref_encoder): Likewise.
9389 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
9390 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
9391 (lto_output_decl_index): Adjust.
9392 (lto_new_out_decl_state): Likewise.
9393 (lto_record_function_out_decl_state): Likewise.
9394 * lto-streamer-out.c (copy_function): Likewise.
9396 2013-06-18 Richard Biener <rguenther@suse.de>
9398 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
9399 * cgraphunit.c: Include cfgloop.h.
9400 (init_lowered_empty_function): Initialize the loop tree.
9401 (assemble_thunk): Insert new BBs into loops.
9403 2013-06-18 Richard Biener <rguenther@suse.de>
9405 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
9406 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
9407 the map from cache entry to cache index optional.
9408 (streamer_tree_cache_replace_tree): Adjust accordingly.
9409 (streamer_tree_cache_append): Likewise.
9410 (streamer_tree_cache_delete): Likewise.
9411 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
9412 streamer cache map from cache entry to cache index.
9413 * lto-streamer-out.c (create_output_block): Adjust.
9415 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
9417 * config/arm/arm.md (attribute "insn"): Move multiplication and
9418 division attributes to...
9419 (attribute "type"): ... here. Remove mult.
9420 (attribute "mul32"): New attribute.
9421 (attribute "mul64"): Add umaal.
9422 (*arm_mulsi3): Update attributes.
9423 (*arm_mulsi3_v6): Likewise.
9424 (*thumb_mulsi3): Likewise.
9425 (*thumb_mulsi3_v6): Likewise.
9426 (*mulsi3_compare0): Likewise.
9427 (*mulsi3_compare0_v6): Likewise.
9428 (*mulsi_compare0_scratch): Likewise.
9429 (*mulsi_compare0_scratch_v6): Likewise.
9430 (*mulsi3addsi): Likewise.
9431 (*mulsi3addsi_v6): Likewise.
9432 (*mulsi3addsi_compare0): Likewise.
9433 (*mulsi3addsi_compare0_v6): Likewise.
9434 (*mulsi3addsi_compare0_scratch): Likewise.
9435 (*mulsi3addsi_compare0_scratch_v6): Likewise.
9436 (*mulsi3subsi): Likewise.
9437 (*mulsidi3adddi): Likewise.
9438 (*mulsi3addsi_v6): Likewise.
9439 (*mulsidi3adddi_v6): Likewise.
9440 (*mulsidi3_nov6): Likewise.
9441 (*mulsidi3_v6): Likewise.
9442 (*umulsidi3_nov6): Likewise.
9443 (*umulsidi3_v6): Likewise.
9444 (*umulsidi3adddi): Likewise.
9445 (*umulsidi3adddi_v6): Likewise.
9446 (*smulsi3_highpart_nov6): Likewise.
9447 (*smulsi3_highpart_v6): Likewise.
9448 (*umulsi3_highpart_nov6): Likewise.
9449 (*umulsi3_highpart_v6): Likewise.
9450 (mulhisi3): Likewise.
9451 (*mulhisi3tb): Likewise.
9452 (*mulhisi3bt): Likewise.
9453 (*mulhisi3tt): Likewise.
9454 (maddhisi4): Likewise.
9455 (*maddhisi4tb): Likewise.
9456 (*maddhisi4tt): Likewise.
9457 (maddhidi4): Likewise.
9458 (*maddhidi4tb): Likewise.
9459 (*maddhidi4tt): Likewise.
9461 (udivsi3): Likewise.
9462 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
9463 (thumb2_mulsi_short_compare0): Likewise.
9464 (thumb2_mulsi_short_compare0_scratch): Likewise.
9465 * config/arm/arm1020e.md (1020mult1): Update attribute change.
9466 (1020mult2): Likewise.
9467 (1020mult3): Likewise.
9468 (1020mult4): Likewise.
9469 (1020mult5): Likewise.
9470 (1020mult6): Likewise.
9471 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
9473 (cortex_a15_mult64): Likewise.
9474 (cortex_a15_sdiv): Likewise.
9475 (cortex_a15_udiv): Likewise.
9476 * config/arm/arm1026ejs.md (mult1): Update attribute change.
9482 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
9483 (pj4_ir_div): Likewise.
9484 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
9485 (11_mult2): Likewise.
9486 (11_mult3): Likewise.
9487 (11_mult4): Likewise.
9488 (11_mult5): Likewise.
9489 (11_mult6): Likewise.
9490 (11_mult7): Likewise.
9491 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
9492 (cortex_a8_mla): Likewise.
9493 (cortex_a8_mull): Likewise.
9494 (cortex_a8_smulwy): Likewise.
9495 (cortex_a8_smlald): Likewise.
9496 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
9497 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
9498 (cortex_r4_mul_3): Likewise.
9499 (cortex_r4_mla_4): Likewise.
9500 (cortex_r4_mla_3): Likewise.
9501 (cortex_r4_smlald): Likewise.
9502 (cortex_r4_mull): Likewise.
9503 (cortex_r4_sdiv): Likewise.
9504 (cortex_r4_udiv): Likewise.
9505 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
9506 (cortex_a7_idiv): Likewise.
9507 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
9508 (9_mult2): Likewise.
9509 (9_mult3): Likewise.
9510 (9_mult4): Likewise.
9511 (9_mult5): Likewise.
9512 (9_mult6): Likewise.
9513 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
9514 (cortex_a53_sdiv): Likewise.
9515 (cortex_a53_udiv): Likewise.
9516 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
9517 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
9518 (mp626_mult2): Likewise.
9519 (mp626_mult3): Likewise.
9520 (mp626_mult4): Likewise.
9521 * config/arm/fa526.md (526_mult1): Update attribute change.
9522 (526_mult2): Likewise.
9523 * config/arm/arm-generic.md (mult): Update attribute change.
9524 (mult_ldsched_strongarm): Likewise.
9525 (mult_ldsched): Likewise.
9526 (multi_cycle): Likewise.
9527 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
9528 * config/arm/fa606te.md (606te_mult1): Update attribute change.
9529 (606te_mult2): Likewise.
9530 (606te_mult3): Likewise.
9531 (606te_mult4): Likewise.
9532 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
9533 (cortex_a9_mac16): Likewise.
9534 (cortex_a9_multiply): Likewise.
9535 (cortex_a9_mac): Likewise.
9536 (cortex_a9_multiply_long): Likewise.
9537 * config/arm/fa626te.md (626te_mult1): Update attribute change.
9538 (626te_mult2): Likewise.
9539 (626te_mult3): Likewise.
9540 (626te_mult4): Likewise.
9542 2013-06-18 Richard Biener <rguenther@suse.de>
9545 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
9547 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9550 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9551 with next_active_insn.
9553 2013-06-18 Alan Modra <amodra@gmail.com>
9555 * config/rs6000/rs6000.h (enum data_align): New.
9556 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
9557 (DATA_ABI_ALIGNMENT): Define.
9558 (CONSTANT_ALIGNMENT): Correct comment.
9559 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
9560 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
9562 2013-06-17 David Malcolm <dmalcolm@redhat.com>
9564 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
9565 ATTRIBUTE_UNUSED marking.
9567 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
9569 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
9570 alternative and update.
9571 (aarch64_dup_lanedi): Delete.
9572 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
9573 * config/aarch64/aarch64-simd-builtins.def: Update.
9575 2013-06-17 Richard Biener <rguenther@suse.de>
9577 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
9578 (lto_input_scc): Declare.
9579 (lto_input_tree_1): Likewise.
9580 (struct lto_stats_d): Add num_tree_bodies_output and
9581 num_pickle_refs_output.
9582 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
9583 (lto_read_tree_1): Split out from ...
9584 (lto_read_tree): ... this.
9585 (lto_input_scc): New function.
9586 (lto_input_tree_1): Split out from ...
9587 (lto_input_tree): ... this. Handle LTO_tree_scc.
9588 (lto_data_in_create): Create the streamer cache without hashes.
9589 * lto-streamer-out.c (create_output_block): Create the streamer
9590 cache with hashes when not doing WPA.
9591 (lto_write_tree_1): Split out from ...
9592 (lto_write_tree): ... this.
9593 (get_symbol_initial_value): New function.
9594 (lto_output_tree_1): Split out from ...
9595 (lto_output_tree): ... this. Write trees as series of SCCs
9596 using a DFS walk via DFS_write_tree.
9597 (struct sccs, struct scc_entry): New types.
9598 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
9599 (DFS_write_tree_body): New function.
9600 (DFS_write_tree): Likewise.
9601 (hash_tree): Likewise.
9602 (scc_entry_compare): Likewise.
9603 (hash_scc): Likewise.
9604 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
9605 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
9606 TREE_CHAIN as regular reference.
9607 (streamer_read_integer_cst): Remove.
9608 (streamer_get_pickled_tree): Adjust.
9609 * tree-streamer-out.c (streamer_write_chain): Disable streaming
9610 of DECL_EXTERNALs in BLOCK_VARS for now.
9611 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
9613 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
9614 Add hash value argument and record that if hashes are recorded
9616 (streamer_tree_cache_insert_1): Adjust.
9617 (streamer_tree_cache_insert): Likewise.
9618 (streamer_tree_cache_insert_at): Rename to ...
9619 (streamer_tree_cache_replace_tree): ... this and adjust.
9620 (streamer_tree_cache_append): Adjust.
9621 (record_common_node): Likewise.
9622 (streamer_tree_cache_create): Add argument whether to
9623 record hash values together with trees.
9624 (streamer_tree_cache_delete): Adjust.
9625 * tree-streamer.h (struct streamer_tree_cache_d): Add
9627 (streamer_read_integer_cst): Remove.
9628 (streamer_tree_cache_insert): Adjust.
9629 (streamer_tree_cache_append): Likewise.
9630 (streamer_tree_cache_insert_at): Rename to ...
9631 (streamer_tree_cache_replace_tree): ... this and adjust.
9632 (streamer_tree_cache_create): Add argument whether to record hashes.
9633 (streamer_tree_cache_get): Rename to ...
9634 (streamer_tree_cache_get_tree): ... this.
9635 (streamer_tree_cache_get_hash): New function.
9636 * tree.c (cache_integer_cst): New function.
9637 * tree.h (cache_integer_cst): Declare.
9638 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
9639 * lto-symtab.c (lto_varpool_replace_node): Only release
9640 DECL_INITIAL of non-prevailing decls.
9641 * varpool.c (varpool_remove_initializer): Do not release
9642 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
9644 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
9646 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
9647 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
9648 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
9649 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
9650 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
9651 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
9652 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
9653 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
9654 instead of TARGET_64BIT.
9655 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
9656 Require ISA_HAS_<D>DIV.
9658 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
9660 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
9661 (mips*-*-linux*): Move default with_llsc setting to where other
9663 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
9665 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
9666 Likewise. Remove default with_tune setting. Move default float
9667 setting to its own block. Handle with_llsc in the same block as above.
9669 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
9671 PR rtl-optimization/57425
9672 PR rtl-optimization/57569
9673 * alias.c (write_dependence_p): Add new parameters mem_mode,
9674 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
9675 Changed all callers.
9676 (canon_anti_dependence): New function.
9677 * cse.c (check_dependence): Use canon_anti_dependence.
9678 * cselib.c (cselib_invalidate_mem): Likewise.
9679 * rtl.h (canon_anti_dependence): Declare.
9681 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
9683 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
9684 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
9685 ".set mips3" for 64-bit targets.
9687 2013-06-15 Dehao Chen <dehao@google.com>
9689 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
9690 * gimple-low.c (gimple_check_call_matching_types): Likewise.
9691 (gimple_check_call_args): Likewise.
9692 * value-prof.c (check_ic_target): Likewise.
9693 * ipa-inline.c (early_inliner): Likewise.
9694 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
9695 * cgraph.c (cgraph_create_edge_1): Likewise.
9696 (cgraph_make_edge_direct): Likewise.
9698 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
9701 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
9702 rs6000_output_move_128bit to handle emitting quad memory
9703 operations. Set attribute length to 8 bytes.
9705 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
9707 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
9709 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
9710 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
9711 (aarch64_<su>mlsl<mode>): Likewise.
9713 2013-06-14 Mike Stump <mikestump@comcast.net>
9715 * Makefile.in (TARGET_H): Add insn-codes.h.
9717 2013-06-14 Alan Modra <amodra@gmail.com>
9721 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
9722 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
9723 bitfield expansion when EXPAND_MEMORY.
9724 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
9726 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
9728 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
9729 test for clearing quad memory on 32-bit later.
9731 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
9733 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
9734 (fold_negate_expr): Likewise.
9735 (fold_real_zero_addition_p): Handle vectors.
9736 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
9738 2013-06-14 Alan Modra <amodra@gmail.com>
9740 * varasm.c (force_const_mem): Revert 2013-06-07 change.
9742 2013-06-13 Jan Hubicka <jh@suse.cz>
9744 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
9745 Local comdats are not externally visible.
9746 * symtab.c (dump_symtab_base): Dump externally visible.
9747 (verify_symtab_base): Verify back links in the symtab hash.
9749 2013-06-13 Bin Cheng <bin.cheng@arm.com>
9751 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
9752 CONVERT_EXPR as equal nodes.
9754 2013-06-13 Bin Cheng <bin.cheng@arm.com>
9756 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
9758 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
9760 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
9761 Generalize to complex and vector.
9762 * tree.c (build_all_ones_cst): New function.
9763 * tree.h (build_all_ones_cst): Declare it.
9765 2013-06-13 Alan Modra <amodra@gmail.com>
9767 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
9768 * config/rs6000/rs6000.md (signbittf2): New insn.
9769 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
9770 (abstf2_internal, cmptf_internal2): Likewise.
9771 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
9773 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
9774 Pat Haugen <pthaugen@us.ibm.com>
9775 Peter Bergner <bergner@vnet.ibm.com>
9777 * config/rs6000/rs6000.c (emit_load_locked): Add support for
9778 power8 byte, half-word, and quad-word atomic instructions.
9779 (emit_store_conditional): Likewise.
9780 (rs6000_expand_atomic_compare_and_swap): Likewise.
9781 (rs6000_expand_atomic_op): Likewise.
9783 * config/rs6000/sync.md (larx): Add new modes for power8.
9785 (AINT): New mode iterator to include TImode as well as normal
9786 integer modes on power8.
9787 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
9788 that VSX registers are not considered. Use AINT mode iterator
9789 instead of INT1 to allow inclusion of quad word atomic operations
9791 (load_locked<mode>): Likewise.
9792 (store_conditional<mode>): Likewise.
9793 (atomic_compare_and_swap<mode>): Likewise.
9794 (atomic_exchange<mode>): Likewise.
9795 (atomic_nand<mode>): Likewise.
9796 (atomic_fetch_<fetchop_name><mode>): Likewise.
9797 (atomic_nand_fetch<mode>): Likewise.
9798 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
9800 (ATOMIC): On power8, add QImode, HImode modes.
9801 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
9802 modes that promote to SImode.
9803 (load_lockedti): Convert TImode arguments to PTImode, so that we
9804 get a guaranteed even/odd register pair.
9805 (load_lockedpti): Likewise.
9806 (store_conditionalti): Likewise.
9807 (store_conditionalpti): Likewise.
9809 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
9810 atomic load/store instructions.
9813 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
9815 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
9817 (insn_count): New attribute, with most cases extracted from...
9818 (length): ...here. Redefine most cases in terms of insn_count.
9819 (single_insn): Delete.
9820 (can_delay): Use insn_count to check for single instructions.
9821 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
9822 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
9823 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
9824 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
9825 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
9826 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
9827 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
9828 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
9829 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
9830 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
9831 rather than "length".
9832 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
9833 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
9834 Use "insn_count" rather than "length".
9835 * config/mips/mips-dsp.md
9836 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
9837 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
9840 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
9842 PR tree-optimization/57361
9843 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
9845 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
9847 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
9849 (aarch64_simd_combine<mode>): New instruction expansion.
9850 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
9852 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
9853 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
9855 2013-06-12 Jan Hubicka <jh@suse.cz>
9857 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
9858 decl has when in streaming stage.
9859 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
9860 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
9862 2013-06-12 Roland Stigge <stigge@antcom.de>
9865 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
9867 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9869 PR tree-optimization/57537
9870 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
9871 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
9873 2013-06-12 Richard Biener <rguenther@suse.de>
9875 * data-streamer.h (streamer_write_char_stream): CSE
9876 obs->current_pointer.
9877 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
9878 streamer_write_char_stream manually and optimize the resulting loop.
9879 (streamer_write_hwi_stream): Likewise.
9881 2013-06-12 Jan Hubicka <jh@suse.cz>
9883 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
9884 * cgraph.h (varpool_create_empty_node): Declare.
9885 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
9887 * symtab.c (symtab_unregister_node): Be lax about missin entries
9889 (symtab_get_node): Update comment.
9890 * varpool.c (varpool_create_empty_node): Break out from ...
9891 (varpool_node_for_decl): ... here.
9892 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
9894 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
9896 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
9897 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
9898 part. Use straight-line flow at the end.
9899 <COMPONENT_REF>: Remove superfluous else.
9900 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
9902 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9905 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
9906 target hook even for !TREE_PUBLIC decls. If no resolution info
9907 is available, return false for common and external decls.
9909 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9911 * config/rl78/constraints.md (U): New constraint.
9912 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
9915 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9918 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
9919 to allow returning address to AT_PLATFORM name.
9921 2013-06-11 Jan Hubicka <jh@suse.cz>
9923 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
9924 * cgraph.h (symtab_node_base): Add weakref flag.
9925 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
9926 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
9927 (output_weakrefs): Use weakref flag.
9928 * fold-const.c (simple_operand_p): Handle WEAK.
9929 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
9930 * ipa.c (varpool_externally_visible_p): Drop weakref.
9931 (function_and_variable_visibility): Update comment; fix weakref
9932 sanity checks; do not clear DECL_WEAK on them.
9933 * lto-cgraph.c (lto_output_node): update.
9934 (lto_output_varpool_node): Update.
9935 (input_overwrite_node): Update.
9936 (input_node): Update.
9937 (input_varpool_node): Update.
9938 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
9939 (lto_symtab_merge_symbols): Add sanity check.
9940 (lto_symtab_prevailing_decl): Do not special case weakrefs.
9941 * passes.c (rest_of_decl_compilation): Set static flag, too.
9942 * symtab.c (dump_symtab_base): Dump weakref.
9943 (verify_symtab_base): Sanity check weakrefs.
9944 (symtab_make_decl_local): Remove duplicated code.
9945 (symtab_alias_ultimate_target): Simplify.
9946 * varpool.c (varpool_create_variable_alias): Set weakref flag.
9948 2013-06-11 Tom de Vries <tom@codesourcery.com>
9950 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
9951 sequence_vect == NULL.
9953 2013-06-11 DJ Delorie <dj@redhat.com>
9955 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
9956 (rl78_unwind_word_mode): New.
9958 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9960 * final.c (debug_prefix_maps): Make static.
9962 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9964 * function.c (initial_trampoline): Remove stray copy.
9966 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
9968 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
9970 2013-06-11 Martin Jambor <mjambor@suse.cz>
9972 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
9973 within bounds at the beginning of the function.
9975 2013-06-11 Alan Modra <amodra@gmail.com>
9977 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
9979 (get_named_section): Don't NULL !DECL_P decl.
9981 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
9983 * doc/invoke.texi (core-avx2): Document.
9985 (atom): Updated with MOVBE.
9987 2013-06-11 Richard Biener <rguenther@suse.de>
9989 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
9991 2013-06-11 Anton Blanchard <anton@samba.org>
9993 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
9994 correct shift value in little-endian mode.
9996 2013-06-11 Jakub Jelinek <jakub@redhat.com>
9999 * varasm.c (get_variable_align): Move #endif to the right place.
10001 2013-06-10 Cary Coutant <ccoutant@google.com>
10003 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10004 for hash so that hash table traversal order is deterministic.
10006 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10007 Pat Haugen <pthaugen@us.ibm.com>
10008 Peter Bergner <bergner@vnet.ibm.com>
10010 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10011 of vectors in GPRS if they are direct moves or quad word load or
10014 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10016 (direct_move_p): Likewise.
10017 (quad_load_store_p): Likewise.
10019 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10020 classes into bins based on the physical register type.
10021 (reg_class_to_reg_type): Likewise.
10022 (IS_STD_REG_TYPE): Likewise.
10023 (IS_FP_VECT_REG_TYPE): Likewise.
10024 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10025 use direct move instructions.
10026 (reload_gpr_vsx): Likewise.
10027 (reload_vsx_gpr): Likewise.
10028 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10029 information that is a simplification of register classes. Also
10030 precalculate direct move reload helpers.
10031 (direct_move_p): New function to return true if the operation can
10032 be done as a direct move instruciton.
10033 (quad_load_store_p): New function to return true if the operation
10034 is a quad memory operation.
10035 (rs6000_legitimize_address): If quad memory, only allow register
10036 indirect for TImode addresses.
10037 (rs6000_legitimate_address_p): Likewise.
10038 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10039 (rs6000_reload_register_type): Likewise.
10040 (register_to_reg_type): Return register type.
10041 (rs6000_secondary_reload_simple_move): New helper function for
10042 secondary reload and secondary memory needed to identify anything
10043 that is a simple move, and does not need reloading.
10044 (rs6000_secondary_reload_direct_move): New helper function for
10045 secondary reload to identify cases that can be done with several
10046 instructions via the direct move instructions.
10047 (rs6000_secondary_reload_move): New helper function for secondary
10048 reload to identify moves between register types that can be done.
10049 (rs6000_secondary_reload): Add support for quad memory operations
10050 and for direct move.
10051 (rs6000_secondary_memory_needed): Likewise.
10052 (rs6000_debug_secondary_memory_needed): Change argument names.
10053 (rs6000_output_move_128bit): New function to return the move to
10054 use for 128-bit moves, including knowing about the various
10055 limitations of quad memory operations.
10057 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
10058 memory operations. call rs6000_output_move_128bit for the actual
10059 instruciton(s) to generate.
10060 (vsx_movti_64bit): Likewise.
10062 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
10063 (UNSPEC_P8V_MTVSRWZ): Likewise.
10064 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
10065 (UNSPEC_P8V_MTVSRD): Likewise.
10066 (UNSPEC_P8V_XXPERMDI): Likewise.
10067 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
10068 (UNSPEC_FUSION_GPR): Likewise.
10069 (FMOVE128_GPR): New iterator for direct move.
10070 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
10071 (f32_sv): Likewise.
10072 (f32_dm): Likewise.
10073 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
10074 loads and direct move instructions.
10075 (zero_extendsidi2_lfiwzx): Likewise.
10076 (extendsidi2_lfiwax): Likewise.
10077 (extendsidi2_nocell): Likewise.
10078 (floatsi<mode>2_lfiwax): Likewise.
10079 (lfiwax): Likewise.
10080 (floatunssi<mode>2_lfiwzx): Likewise.
10081 (lfiwzx): Likewise.
10082 (fix_trunc<mode>_stfiwx): Likewise.
10083 (fixuns_trunc<mode>_stfiwx): Likewise.
10084 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
10085 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
10086 (parity<mode>2_cmpb): Set length/type attr.
10087 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
10088 for 'mr.' to fast_compare.
10089 (bpermd_<mode>): Change type attr to popcnt.
10090 (p8_fmrgow_<mode>): New insns for power8 direct move support.
10091 (p8_mtvsrwz_1): Likewise.
10092 (p8_mtvsrwz_2): Likewise.
10093 (reload_fpr_from_gpr<mode>): Likewise.
10094 (p8_mtvsrd_1): Likewise.
10095 (p8_mtvsrd_2): Likewise.
10096 (p8_xxpermdi_<mode>): Likewise.
10097 (reload_vsx_from_gpr<mode>): Likewise.
10098 (reload_vsx_from_gprsf): Likewise.
10099 (p8_mfvsrd_3_<mode>): LIkewise.
10100 (reload_gpr_from_vsx<mode>): Likewise.
10101 (reload_gpr_from_vsxsf): Likewise.
10102 (p8_mfvsrd_4_disf): Likewise.
10103 (multi-word GPR splits): Do not split direct moves or quad memory
10106 2013-06-10 David Malcolm <dmalcolm@redhat.com>
10108 * tree-into-ssa.c (interesting_blocks): Make static.
10110 2013-06-10 Jakub Jelinek <jakub@redhat.com>
10113 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
10114 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
10115 Use DATA_ABI_ALIGNMENT for that case instead if defined.
10116 (get_variable_align): New function.
10117 (get_variable_section, emit_bss, emit_common,
10118 assemble_variable_contents, place_block_symbol): Use
10119 get_variable_align instead of DECL_ALIGN.
10120 (assemble_noswitch_variable): Add align argument, use it
10121 instead of DECL_ALIGN.
10122 (assemble_variable): Adjust caller. Use get_variable_align
10123 instead of DECL_ALIGN.
10124 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
10126 (DATA_ABI_ALIGNMENT): Define.
10127 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
10128 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
10129 opt is false, only return the psABI mandated alignment increase.
10130 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
10131 (DATA_ABI_ALIGNMENT): ... this.
10132 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
10133 (DATA_ABI_ALIGNMENT): ... this.
10134 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
10135 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
10136 (DATA_ABI_ALIGNMENT): ... this.
10137 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
10138 * doc/tm.texi: Regenerated.
10140 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
10142 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
10143 cmp_code to construct REG_EQUAL note.
10145 2013-06-09 Jakub Jelinek <jakub@redhat.com>
10148 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
10149 that operands[2] doesn't overlap with operands[0].
10151 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
10152 Jan Hubicka <jh@suse.cz>
10154 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
10155 hack to mark symbols as used.
10157 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
10159 PR rtl-optimization/57559
10160 * lra-constraints.c (process_alt_operands): Don't discourage
10161 memory with known offset for offsetable memory constraint.
10162 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
10164 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
10166 * varasm.c (struct oc_local_state): Reorder fields.
10167 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
10168 and adjust accordingly.
10169 (output_constructor): Reorder initialization code and adjust call to
10170 output_constructor_bitfield.
10172 2013-06-07 Jan Hubicka <jh@suse.cz>
10174 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
10176 2013-06-07 David Malcolm <dmalcolm@redhat.com>
10178 * tree-object-size.c (unknown): Make const.
10180 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10182 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
10183 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
10184 for last alternative in the cpu_facility attribute.
10186 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10189 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
10190 (xordi3): Change operand 2 constraint to arm_xordi_operand.
10191 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
10192 * config/arm/constraints.md (Dg): New constraint.
10193 * config/arm/neon.md (xordi3_neon): Remove.
10194 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
10195 * config/arm/predicates.md (arm_xordi_operand): New predicate.
10197 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10199 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
10200 Clean up alternatives.
10202 2013-06-07 Alan Modra <amodra@gmail.com>
10204 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
10207 2013-06-07 Alan Modra <amodra@gmail.com>
10209 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
10211 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10213 * config/arm/constraints.md (Df): New constraint.
10214 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
10215 Correct length attribute for last two alternatives.
10217 2013-06-07 Alan Modra <amodra@gmail.com>
10219 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10220 override user -mfp-in-toc.
10221 (offsettable_ok_by_alignment): Consider just the current access
10222 rather than the whole object, unless BLKmode. Handle
10223 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
10224 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
10225 for -mcmodel=medium.
10226 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
10227 override user -mfp-in-toc or -msum-in-toc. Default to
10228 -mno-fp-in-toc for -mcmodel=medium.
10230 2013-06-06 DJ Delorie <dj@redhat.com>
10232 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
10233 TARGET_VALID_POINTER_MODE.
10235 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
10236 Pat Haugen <pthaugen@us.ibm.com>
10237 Peter Bergner <bergner@vnet.ibm.com>
10239 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
10240 Document new power8 builtins.
10242 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
10243 condition code register, to allow 128-bit logical operations to be
10244 done in the VSX or GPR registers.
10245 (nor<mode>3): Use the canonical form for nor.
10246 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
10247 vclz*, and vpopcnt* vector instructions.
10248 (nand<mode>3): Likewise.
10249 (orc<mode>3): Likewise.
10250 (clz<mode>2): LIkewise.
10251 (popcount<mode>2): Likewise.
10253 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
10254 that only the GPRs are recognized.
10256 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10257 support for new power8 builtins.
10259 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
10261 (xscvdpspn): Likewise.
10267 (vpopcnt): Likewise.
10268 (vpopcntb): Likewise.
10269 (vpopcnth): Likewise.
10270 (vpopcntw): Likewise.
10271 (vpopcntd): Likewise.
10273 (vmrgew): Likewise.
10274 (vmrgow): Likewise.
10276 (eqv_v16qi3): Likewise.
10277 (eqv_v8hi3): Likewise.
10278 (eqv_v4si3): Likewise.
10279 (eqv_v2di3): Likewise.
10280 (eqv_v4sf3): Likewise.
10281 (eqv_v2df3): Likewise.
10283 (nand_v16qi3): Likewise.
10284 (nand_v8hi3): Likewise.
10285 (nand_v4si3): Likewise.
10286 (nand_v2di3): Likewise.
10287 (nand_v4sf3): Likewise.
10288 (nand_v2df3): Likewise.
10290 (orc_v16qi3): Likewise.
10291 (orc_v8hi3): Likewise.
10292 (orc_v4si3): Likewise.
10293 (orc_v2di3): Likewise.
10294 (orc_v4sf3): Likewise.
10295 (orc_v2df3): Likewise.
10297 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
10298 allow power8 quad mode in 64-bit.
10299 (rs6000_builtin_vectorized_function): Add support to vectorize
10300 ISA 2.07 count leading zeros, population count builtins.
10301 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
10302 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
10303 (builtin_function_type): Add vgbbd builtin function which takes an
10305 (altivec_expand_vec_perm_const): Add support for new power8 merge
10308 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
10309 that does not include TImdoe for use with 32-bit.
10310 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
10312 (UNSPEC_VSX_CVDPSPN): Likewise.
10313 (vsx_xscvdpspn): Likewise.
10314 (vsx_xscvspdpn): Likewise.
10315 (vsx_xscvdpspn_scalar): Likewise.
10316 (vsx_xscvspdpn_directmove): Likewise.
10317 (vsx_and<mode>3): Split logical operations into 32-bit and
10318 64-bit. Add support to do logical operations on TImode as well as
10319 VSX vector types. Allow logical operations to be done in either
10320 VSX registers or in general purpose registers in 64-bit mode. Add
10321 splitters if GPRs were used. For AND, add clobber of CCmode to
10322 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
10324 (vsx_and<mode>3_32bit): Likewise.
10325 (vsx_and<mode>3_64bit): Likewise.
10326 (vsx_ior<mode>3): Likewise.
10327 (vsx_ior<mode>3_32bit): Likewise.
10328 (vsx_ior<mode>3_64bit): Likewise.
10329 (vsx_xor<mode>3): Likewise.
10330 (vsx_xor<mode>3_32bit): Likewise.
10331 (vsx_xor<mode>3_64bit): Likewise.
10332 (vsx_one_cmpl<mode>2): Likewise.
10333 (vsx_one_cmpl<mode>2_32bit): Likewise.
10334 (vsx_one_cmpl<mode>2_64bit): Likewise.
10335 (vsx_nor<mode>3): Likewise.
10336 (vsx_nor<mode>3_32bit): Likewise.
10337 (vsx_nor<mode>3_64bit): Likewise.
10338 (vsx_andc<mode>3): Likewise.
10339 (vsx_andc<mode>3_32bit): Likewise.
10340 (vsx_andc<mode>3_64bit): Likewise.
10341 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
10342 and xxlorc instructions.
10343 (vsx_eqv<mode>3_64bit): Likewise.
10344 (vsx_nand<mode>3_32bit): Likewise.
10345 (vsx_nand<mode>3_64bit): Likewise.
10346 (vsx_orc<mode>3_32bit): Likewise.
10347 (vsx_orc<mode>3_64bit): Likewise.
10349 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
10351 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
10353 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
10354 (p8_vmrgow): Likewise.
10355 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
10356 GPRs to be split under VSX.
10357 (p8v_clz<mode>2): Add power8 count leading zero support.
10358 (p8v_popcount<mode>2): Add power8 population count support.
10359 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
10362 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
10365 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
10367 (vec_nand): Likewise.
10368 (vec_vclz): Likewise.
10369 (vec_vclzb): Likewise.
10370 (vec_vclzd): Likewise.
10371 (vec_vclzh): Likewise.
10372 (vec_vclzw): Likewise.
10373 (vec_vgbbd): Likewise.
10374 (vec_vmrgew): Likewise.
10375 (vec_vmrgow): Likewise.
10376 (vec_vpopcnt): Likewise.
10377 (vec_vpopcntb): Likewise.
10378 (vec_vpopcntd): Likewise.
10379 (vec_vpopcnth): Likewise.
10380 (vec_vpopcntw): Likewise.
10382 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10384 PR rtl-optimization/57468
10385 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
10388 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10390 PR rtl-optimization/57459
10391 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
10392 type when setting live regs.
10394 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10396 * config/s390/s390.opt (mlra): New option.
10397 * config/s390/s390.c (s390_decompose_address): Check displacement
10398 for all registers for LRA.
10399 (s390_secondary_reload): Don't used secondary reloads for LRA.
10400 (s390_lra_p): New function.
10401 (TARGET_LRA_P): Define.
10402 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
10403 of attribute cpu_facility to zarch for the last alternative.
10404 (*cmpmem_short): Ditto.
10406 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
10408 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
10409 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
10410 (arm_expand_prologue): Likewise.
10412 2013-06-06 Teresa Johnson <tejohnson@google.com>
10415 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
10416 as this is now done by redirect_edge_and_branch_force.
10417 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
10418 barriers, and fix interaction with splitting.
10419 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
10420 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
10421 reflect changes made in the routine.
10422 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
10423 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
10424 since this is called in cfglayout mode, and replace partition fixup
10425 with assert as that is now done by force_nonfallthru_and_redirect.
10426 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
10427 already be marked with region crossing note.
10428 (insert_section_boundary_note): Make non-static, gate on flag
10429 has_bb_partition, rewrite to also check for multiple partitions.
10430 (rest_of_handle_reorder_blocks): Remove call to
10431 insert_section_boundary_note, now done later during free_cfg.
10432 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
10433 * bb-reorder.h (insert_section_boundary_note): Declare.
10434 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
10435 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
10436 invoke insert_section_boundary_note.
10437 (try_redirect_by_replacing_jump): Remove unnecessary
10438 check for region crossing note.
10439 (fixup_partition_crossing): New function.
10440 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
10441 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
10442 in non-cfglayout mode.
10443 (force_nonfallthru_and_redirect): Fixup partition boundaries,
10444 remove old code that tried to do this. Emit barrier correctly
10445 when we are in cfglayout mode.
10446 (last_bb_in_partition): New function.
10447 (rtl_split_edge): Correctly fixup partition boundaries.
10448 (commit_one_edge_insertion): Remove old code that tried to
10449 fixup region crossing edge since this is now handled in
10450 split_block, and set up insertion point correctly since
10451 block may now end in a jump.
10452 (verify_hot_cold_block_grouping): Guard against checking when not in
10453 linearized RTL mode.
10454 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
10456 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
10457 rtl_verify_flow_info, so not called in cfglayout mode.
10458 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
10459 (fixup_reorder_chain): Remove old code that attempted to fixup region
10460 crossing note as this is now handled in force_nonfallthru_and_redirect.
10461 (duplicate_insn_chain): Don't duplicate switch section notes.
10462 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
10464 * basic-block.h (emit_barrier_after_bb): Declare.
10466 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10468 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
10469 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
10470 arm_usatsihi): Adjust alternatives for arm_restrict_it.
10472 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10474 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
10476 * config/arm/ldmstm.md: Regenerate.
10478 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10480 * config/arm/sync.md (atomic_loaddi_1):
10481 Disable predication for arm_restrict_it.
10482 (arm_load_exclusive<mode>): Likewise.
10483 (arm_load_exclusivesi): Likewise.
10484 (arm_load_exclusivedi): Likewise.
10485 (arm_load_acquire_exclusive<mode>): Likewise.
10486 (arm_load_acquire_exclusivesi): Likewise.
10487 (arm_load_acquire_exclusivedi): Likewise.
10488 (arm_store_exclusive<mode>): Likewise.
10489 (arm_store_exclusive<mode>): Likewise.
10490 (arm_store_release_exclusivedi): Likewise.
10491 (arm_store_release_exclusive<mode>): Likewise.
10493 2013-06-06 Richard Biener <rguenther@suse.de>
10495 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
10497 (lto_tag_is_tree_code_p): Adjust.
10498 (lto_tag_is_gimple_code_p): Likewise.
10499 (lto_gimple_code_to_tag): Likewise.
10500 (lto_tag_to_gimple_code): Likewise.
10501 (lto_tree_code_to_tag): Likewise.
10502 (lto_tag_to_tree_code): Likewise.
10503 * data-streamer.h (streamer_write_hwi_in_range): Use
10504 uhwi streaming to stream the normalized range.
10505 (streamer_read_hwi_in_range): Likewise.
10507 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10509 * config/arm/arm.md (enabled_for_depr_it): New attribute.
10510 (predicable_short_it): Likewise.
10511 (predicated): Likewise.
10512 (enabled): Handle above.
10513 (define_cond_exec): Set predicated attribute to yes.
10515 2013-06-05 Mike Stump <mikestump@comcast.net>
10517 * gdbinit.in (__FUNCTION__): Add.
10519 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
10521 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
10522 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
10524 2013-06-05 Jan Hubicka <jh@suse.cz>
10526 * varasm.c (mark_decl_referenced): Revert the removal until targets
10529 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
10531 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
10532 instead of mark_decl_referenced.
10534 2013-06-05 Jan Hubicka <jh@suse.cz>
10536 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
10537 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
10538 and symtab_used_from_object_file_p.
10539 (cgraph_make_node_local_1): Clear forced_by_abi.
10540 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
10541 * cgraph.h (symtab_node_base): Add forced_by_abi.
10542 (decide_is_variable_needed): Remove.
10543 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
10544 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
10545 (decide_is_symbol_needed): ... this one; handle symbols in general;
10546 always analyze virtuals; honnor forced_by_abi.
10547 (cgraph_finalize_function): Update.
10548 (varpool_finalize_decl): Update.
10549 (symbol_defined_and_needed): Remove.
10550 (analyze_functions): Update.
10551 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
10552 output_refs, input_overwrite_node): Handle forced_by_abi.
10553 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
10554 (address_taken_from_non_vtable_p): ... this one.
10555 (comdat_can_be_unshared_p_1): New function.
10556 (cgraph_comdat_can_be_unshared_p): Rename to ...
10557 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
10558 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
10559 (function_and_variable_visibility): Clear forced_by_abi as needed.
10560 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
10561 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
10562 * symtab.c (dump_symtab_base): Dump forced_by_abi.
10563 * varpool.c (decide_is_variable_needed): Remove.
10565 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10567 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
10568 (arm_option_override): Override arm_restrict_it where appropriate.
10569 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
10570 * config/arm/arm.opt (mrestrict-it): New command-line option.
10571 * doc/invoke.texi: Document -mrestrict-it.
10573 2013-06-05 David Malcolm <dmalcolm@redhat.com>
10575 * tsan.c (tsan_atomic_table): Make const.
10577 2013-06-05 Richard Biener <rguenther@suse.de>
10579 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
10580 index associated with the tree we are supposed to replace.
10581 * tree-streamer-out.c (pack_ts_base_value_fields): Output
10582 TREE_ASM_WRITTEN as zero for everything but SSA names.
10584 2013-06-05 David Malcolm <dmalcolm@redhat.com>
10586 * tree-ssa-structalias.c (call_stmt_vars): Make static.
10588 2013-06-04 Jan Hubicka <jh@suse.cz>
10590 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
10591 (input_node, input_varpool_node): Handle correctly external same
10593 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
10594 nodes at ltrans stage.
10596 2013-06-04 Jan Hubicka <jh@suse.cz>
10598 * ipa-inline.c (update_caller_keys): Fix availability test.
10599 (update_callee_keys): Likewise.
10600 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
10601 to follow ELF standard.
10603 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
10605 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
10606 (mips64r5900el-*-elf*): New configurations.
10607 * config/mips/mips-cpus.def (r5900): New processor.
10608 * config/mips/mips-tables.opt: Regenerate.
10609 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
10610 (mips_issue_rate): Handle PROCESSOR_R5900.
10611 (mips_reorg_process_insns): Force reorder mode for the R5900.
10612 * config/mips/mips.h (TARGET_MIPS5900): Define.
10613 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
10615 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
10617 * config/mips/mips.md (processor): Add r5900.
10618 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
10620 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10622 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
10623 into function to generate MOVI instruction.
10624 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
10625 (aarch64_preferred_simd_mode): Turn into wrapper.
10626 (aarch64_output_scalar_simd_mov_immediate): New function.
10627 * config/aarch64/aarch64-protos.h: Add prototype for above.
10629 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10631 * config/aarch64/aarch64.c (simd_immediate_info): Remove
10632 element_char member.
10633 (sizetochar): Return signed char.
10634 (aarch64_simd_valid_immediate): Remove elchar and other
10635 unnecessary variables.
10636 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
10637 Calculate element_char as required.
10638 * config/aarch64/aarch64-protos.h: Update and move prototype
10639 for aarch64_output_simd_mov_immediate.
10640 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
10643 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10645 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
10646 information completed by aarch64_simd_valid_immediate.
10647 (aarch64_legitimate_constant_p): Update arguments.
10648 (aarch64_simd_valid_immediate): Work with struct rather than many
10650 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
10651 (aarch64_simd_make_constant): Update arguments.
10652 (aarch64_output_simd_mov_immediate): Work with struct rather than
10653 many pointers. Output immediate directly rather than as operand.
10654 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
10656 * config/aarch64/constraints.md (Dn): Update arguments.
10658 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10660 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
10662 (aarch64_simd_immediate_valid_for_move): Remove.
10663 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
10664 (aarch64_simd_make_constant): Update call.
10665 (aarch64_output_simd_mov_immediate): Update call.
10666 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
10668 * config/aarch64/constraints.md (Dn): Update call.
10670 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10672 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
10673 return type to bool for prototype.
10674 (aarch64_legitimate_constant_p): Check for true instead of not -1.
10675 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
10676 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
10678 2013-06-04 Catherine Moore <clm@codesourcery.com>
10680 * config/mips/mips.opt (meva): New.
10681 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
10682 (ASM_SPEC): Handle -meva.
10683 * doc/invoke.texi (meva): Document.
10685 2013-06-04 Alan Modra <amodra@gmail.com>
10687 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
10690 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10692 * rtl.def: Add extra fourth optional field to define_cond_exec.
10693 * gensupport.c (process_one_cond_exec): Process attributes from
10695 * doc/md.texi: Document fourth field in define_cond_exec.
10697 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
10699 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
10700 out the processing order as in store_bit_field_1.
10702 2013-06-04 Jan Hubicka <jh@suse.cz>
10704 PR middle-end/57500
10705 * cgraphunit.c (cgraph_process_same_body_aliases): Create
10706 non-VAR_DECL node if it does not exist yet.
10708 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
10710 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
10711 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
10712 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
10713 target_cpu_default setting.
10715 2013-06-03 Teresa Johnson <tejohnson@google.com>
10717 * dumpfile.c (opt_info_switch_p): Change -fopt-info
10718 default to -fopt-info=optimized instead of all.
10719 * doc/invoke.texi: Ditto.
10720 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
10721 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
10722 (execute_vect_slp): Emit BB vectorization success under
10723 MSG_OPTIMIZED_LOCATIONS.
10724 * tree-vect-slp.c (vect_slp_transform_bb): Change
10725 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
10726 * tree-vect-loop.c (vect_transform_loop): Ditto.
10728 2013-06-03 Jason Merrill <jason@redhat.com>
10731 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10732 Use TARGET_EXPR for C++.
10734 2013-06-03 Jakub Jelinek <jakub@redhat.com>
10736 PR rtl-optimization/57268
10737 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
10738 if DEBUG_INSN_P (insn).
10741 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10743 PR rtl-optimization/57268
10744 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10745 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10747 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10749 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
10750 (ix86_avoid_lea_for_addr): Likewise.
10751 (exact_dependency_1): Likewise.
10752 (ix86_adjust_cost): Likewise.
10753 (swap_top_of_ready_list): Fix formatting and !reload_completed check
10755 (do_reorder_for_imul): Fix typo, formatting and
10756 !reload_completed check removed.
10757 (ix86_sched_reorder): Fix typo and formatting.
10758 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
10761 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
10763 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
10765 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
10767 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
10768 <STRING_CST>: Likewise.
10769 <VECTOR_CST>: Likewise.
10771 2013-06-01 Janus Weil <janus@gcc.gnu.org>
10772 Mikael Morin <mikael@gcc.gnu.org>
10774 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
10775 * config.in: Regenerated.
10776 * configure: Regenerated.
10778 2013-06-01 Jan Hubicka <jh@suse.cz>
10780 PR middle-end/57366
10781 * cgraphunit.c (compile): When weakref is not supported,
10782 set up transparent aliases before final output pass.
10783 * varasm.c (assemble_alias): Do not try to do it here.
10785 2013-06-01 Jan Hubicka <jh@suse.cz>
10787 PR middle-end/57467
10788 * passes.c (for_per_function): Skip unanalyzed functions.
10790 2013-06-01 Jan Hubicka <jh@suse.cz>
10792 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
10793 (lto_symtab_merge_symbols_1): ... this one.
10794 (lto_symtab_merge_cgraph_nodes): Rename to ...
10795 (lto_symtab_merge_symbols): ... this one; simplify.
10796 * cgraph.c (same_body_aliases_done): Rename to ...
10797 (cpp_implicit_aliases_done): ... this one.
10798 (cgraph_create_function_alias): Update.
10799 (cgraph_same_body_alias): Update.
10800 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
10801 (verify_edge_corresponds_to_fndecl): Simplify.
10802 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
10803 (cgraph_node): Remove same_body_alias.
10804 (varpool_node): Remove alias_of and extra_name_alias.
10805 (same_body_aliases_done): Rename to ..
10806 (cpp_implicit_aliases_done): ... this one.
10807 (symtab_alias_ultimate_target): Add default parameter.
10808 (symtab_resolve_alias): New function.
10809 (fixup_same_cpp_alias_visibility): Declare.
10810 (cgraph_function_node): Add default parameter.
10811 (cgraph_node_asm_name): Likewise.
10812 (cgraph_function_or_thunk_node): Add default parameter; do
10813 not ICE when it is NULL.
10814 (varpool_variable_node): Likewise.
10815 * tree-emutls.c (create_emultls_var): Update.
10816 (ipa_lower_emutls): Update.
10817 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
10818 (cgraph_reset_node): Reset alias info.
10819 (cgraph_finalize_function): Update.
10820 (fixup_same_cpp_alias_visibility): Move to symtab.c.
10821 (analyze_function): Simplify.
10822 (cgraph_process_same_body_aliases): Simplify.
10823 (analyze_functions): Fixup same body aliases.
10824 (handle_alias_pairs): Simplify.
10825 (assemble_thunk): Update.
10826 (assemble_thunks_and_aliases): Update.
10827 (output_weakrefs): Rewrite.
10828 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
10829 (lto_output_varpool_node): Likewise.
10830 (compute_ltrans_boundary): Remve assert.
10831 (get_alias_symbol): New functoin.
10832 (input_node): Rewrite alias handling.
10833 (input_varpool_node): Likewise.
10834 * ipa-pure-const.c (propagate_pure_const): Fix formating.
10835 * ipa.c (process_references): Handle weakrefs correctly.
10836 (symtab_remove_unreachable_nodes): Likewise.
10837 * trans-mem.c (get_cg_data): Update.
10838 (ipa_tm_create_version_alias): Update.
10839 (ipa_tm_execute): Update.
10840 * symtab.c (dump_symtab_base): Dump aliases.
10841 (verify_symtab_base): Verify aliases.
10842 (symtab_node_availability): New function.
10843 (symtab_alias_ultimate_target): Simplify.
10844 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
10845 handle all the fixup cases.
10846 (symtab_resolve_alias): New function.
10847 * passes.c (ipa_write_summaries): Handle weakrefs.
10848 * varpool.c (varpool_analyze_node): Simplify.
10849 (assemble_aliases): Update.
10850 (varpool_create_variable_alias): Simplify.
10851 (varpool_extra_name_alias): Simplify.
10852 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
10853 (lto_symtab_merge_symbols): ... this one.
10855 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
10858 PR rtl-optimization/57268
10859 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10860 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10862 2013-06-01 Tobias Burnus <burnus@net-b.de>
10864 Partially reverted:
10865 2013-05-31 Tobias Burnus <burnus@net-b.de>
10867 PR middle-end/57073
10868 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
10871 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10873 PR rtl-optimization/57268
10874 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10875 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10877 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10879 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
10880 unordered comparison operators when -fno-trapping-math is in effect
10882 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
10883 and implement unordered comparison operators properly on the e500.
10885 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10887 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
10888 for constant scalar integers.
10889 (simplify_relational_operation_1): Likewise.
10891 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
10893 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
10894 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
10897 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10898 Igor Zamyatin <igor.zamyatin@intel.com>
10900 Silvermont (SLM) architecture performance tuning.
10901 * config/i386/i386.h (enum ix86_tune_indices): Add
10902 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
10903 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
10905 * config/i386/i386.c (initial_ix86_tune_features)
10906 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
10907 (ix86_lea_outperforms): Handle Silvermont tuning.
10908 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
10910 (ix86_use_lea_for_mov): Likewise.
10911 (ix86_avoid_lea_for_addr): Likewise.
10912 (ix86_lea_for_add_ok): Likewise.
10913 (exact_dependency_1): New function.
10914 (exact_store_load_dependency): Likewise.
10915 (ix86_adjust_cost): Handle Silvermont tuning.
10916 (do_reoder_for_imul): Likewise.
10917 (swap_top_of_ready_list): New function.
10918 (ix86_sched_reorder): Changed to handle Silvermont tuning.
10920 * config/i386/i386.md (peepholes that split memory operand in fp
10923 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10925 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
10926 Remove un-necessary braces.
10928 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10930 * config/aarch64/aarch64.c (aarch64_classify_symbol):
10931 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
10933 2013-05-31 Tobias Burnus <burnus@net-b.de>
10935 PR middle-end/57073
10936 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
10938 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10941 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
10942 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
10943 * config/arm/neon.md (iordi3_neon): Remove.
10944 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
10945 * config/arm/predicates.md (imm_for_neon_logic_operand):
10946 Move to earlier in the file.
10947 (neon_logic_op2): Likewise.
10948 (arm_iordi_operand_neon): New predicate.
10950 2013-05-31 Richard Biener <rguenther@suse.de>
10952 PR tree-optimization/57478
10953 PR tree-optimization/57453
10954 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
10957 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10959 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
10960 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
10962 2013-05-30 Tobias Burnus <burnus@net-b.de>
10963 Thomas Koenig <tkoenig@gcc.gnu.org>
10965 PR middle-end/57073
10966 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
10967 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
10969 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
10971 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
10973 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
10975 * target.def (register_usage_leveling_p): New hook.
10976 * targhooks.c (default_register_usage_leveling_p): New.
10977 * targhooks.h (default_register_usage_leveling_p): New prototype.
10978 * lra-assigns.c (register_usage_leveling_p): Use the hook.
10979 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
10980 * doc/tm.texi: Update.
10981 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
10983 2013-05-30 Ian Bolton <ian.bolton@arm.com>
10985 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
10986 (*insv_reg<mode>): New define_insn.
10988 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
10990 PR rtl-optimization/57439
10991 * postreload.c (move2add_valid_value_p): Check that we have
10992 a zero subreg_regno_offset when accessing the register in
10993 the requested mode.
10995 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10996 Igor Zamyatin <igor.zamyatin@intel.com>
10998 Silvermont (SLM) architecture pipeline model, tuning and
11000 * config.gcc: Add slm config options and target.
11002 * config/i386/slm.md: New.
11004 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11006 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11008 (ix86_target_macros_internal): Likewise.
11010 * config/i386/i386.c (slm_cost): New cost.
11011 (m_SLM): New macro flag.
11012 (initial_ix86_tune_features): Set m_SLM.
11013 (x86_accumulate_outgoing_args): Likewise.
11014 (x86_arch_always_fancy_math_387): Likewise.
11015 (processor_target_table): Add slm cost.
11016 (cpu_names): Add slm cpu name.
11017 (x86_option_override_internal): Set SLM ISA.
11018 (ix86_issue_rate): New case PROCESSOR_SLM.
11019 (ia32_multipass_dfa_lookahead): Likewise.
11020 (fold_builtin_cpu): Add slm.
11022 * config/i386/i386.h (TARGET_SLM): New target macro.
11023 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11024 (processor_type): Add PROCESSOR_SLM.
11026 * config/i386/i386.md (cpu): Add new value "slm".
11027 (slm.md): Include slm.md.
11029 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11030 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11032 * config/arm/arm-protos.h: Add and update function protos.
11033 * config/arm/arm.c (use_simple_return_p): New added.
11034 (thumb2_expand_return): Check simple_return flag.
11035 * config/arm/arm.md: Add simple_return and conditional simple_return.
11036 * config/arm/iterators.md: Add iterator for return and simple_return.
11038 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11040 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11041 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11042 (arm_emit_vfp_multi_reg_pop): Likewise.
11043 (thumb2_emit_ldrd_pop): Likewise.
11044 (arm_expand_epilogue): Add misc REG_CFA notes.
11045 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11047 2013-05-29 Lawrence Crowl <crowl@google.com>
11049 * config/arm/t-arm: Update for below.
11051 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11052 Change type to hash_table. Update dependent calls and types.
11054 * config/i386/t-cygming: Update for below.
11056 * config/i386/t-interix: Update for below.
11058 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
11059 Change type to hash_table. Update dependent calls and types.
11060 (i386_find_on_wrapper_list::wrappers): Likewise.
11062 * config/ia64/t-ia64: Update for below.
11064 * config/ia64/ia64.c (bundle_state_table):
11065 Change type to hash_table. Update dependent calls and types.
11067 * config/mips/mips.c (mips_reorg_process_insns::htab):
11068 Change type to hash_table. Update dependent calls and types.
11070 * config/sol2.c (solaris_comdat_htab):
11071 Change type to hash_table. Update dependent calls and types.
11073 * config/t-sol2: Update for above.
11075 2013-05-29 Teresa Johnson <tejohnson@google.com>
11077 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
11078 functions are not yet marked as defined.
11080 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
11081 Pat Haugen <pthaugen@us.ibm.com>
11082 Peter Bergner <bergner@vnet.ibm.com>
11084 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
11088 (vrotl<mode>3): Likewise.
11089 (vashl<mode>3): Likewise.
11090 (vlshr<mode>3): Likewise.
11091 (vashr<mode>3): Likewise.
11093 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11094 support for power8 V2DI builtins.
11096 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
11097 power8 V2DI builtins.
11098 (vupkhsw): Likewise.
11099 (vupklsw): Likewise.
11100 (vaddudm): Likewise.
11101 (vminsd): Likewise.
11102 (vmaxsd): Likewise.
11103 (vminud): Likewise.
11104 (vmaxud): Likewise.
11105 (vpkudum): Likewise.
11106 (vpksdss): Likewise.
11107 (vpkudus): Likewise.
11108 (vpksdus): Likewise.
11113 (vsubudm): Likewise.
11114 (vcmpequd): Likewise.
11115 (vcmpgtsd): Likewise.
11116 (vcmpgtud): Likewise.
11117 (vcmpequd_p): Likewise.
11118 (vcmpgtsd_p): Likewise.
11119 (vcmpgtud_p): Likewise.
11120 (vupkhsw): Likewise.
11121 (vupklsw): Likewise.
11122 (vaddudm): Likewise.
11123 (vmaxsd): Likewise.
11124 (vmaxud): Likewise.
11125 (vminsd): Likewise.
11126 (vminud): Likewise.
11127 (vpksdss): Likewise.
11128 (vpksdus): Likewise.
11129 (vpkudum): Likewise.
11130 (vpkudus): Likewise.
11135 (vsubudm): Likewise.
11137 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
11138 support for power8 V2DI instructions.
11140 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
11141 power8 V2DI instructions. Combine pack and unpack insns to use an
11142 iterator for each mode. Check whether a particular mode supports
11143 Altivec instructions instead of just checking TARGET_ALTIVEC.
11144 (UNSPEC_VPKUWUM): Likewise.
11145 (UNSPEC_VPKSHSS): Likewise.
11146 (UNSPEC_VPKSWSS): Likewise.
11147 (UNSPEC_VPKUHUS): Likewise.
11148 (UNSPEC_VPKSHUS): Likewise.
11149 (UNSPEC_VPKUWUS): Likewise.
11150 (UNSPEC_VPKSWUS): Likewise.
11151 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
11152 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
11153 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
11154 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
11155 (UNSPEC_VUPKHSB): Likewise.
11156 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
11157 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
11158 (UNSPEC_VUPKHSH): Likewise.
11159 (UNSPEC_VUPKLSB): Likewise.
11160 (UNSPEC_VUPKLSH): Likewise.
11162 (VI_char): Likewise.
11163 (VI_scalar): Likewise.
11164 (VI_unit): Likewise.
11166 (VP_small): Likewise.
11167 (VP_small_lc): Likewise.
11168 (VU_char): Likewise.
11169 (add<mode>3): Likewise.
11170 (altivec_vaddcuw): Likewise.
11171 (altivec_vaddu<VI_char>s): Likewise.
11172 (altivec_vadds<VI_char>s): Likewise.
11173 (sub<mode>3): Likewise.
11174 (altivec_vsubcuw): Likewise.
11175 (altivec_vsubu<VI_char>s): Likewise.
11176 (altivec_vsubs<VI_char>s): Likewise.
11177 (altivec_vavgs<VI_char>): Likewise.
11178 (altivec_vcmpbfp): Likewise.
11179 (altivec_eq<mode>): Likewise.
11180 (altivec_gt<mode>): Likewise.
11181 (altivec_gtu<mode>): Likewise.
11182 (umax<mode>3): Likewise.
11183 (smax<mode>3): Likewise.
11184 (umin<mode>3): Likewise.
11185 (smin<mode>3): Likewise.
11186 (altivec_vpkuhum): Likewise.
11187 (altivec_vpkuwum): Likewise.
11188 (altivec_vpkshss): Likewise.
11189 (altivec_vpkswss): Likewise.
11190 (altivec_vpkuhus): Likewise.
11191 (altivec_vpkshus): Likewise.
11192 (altivec_vpkuwus): Likewise.
11193 (altivec_vpkswus): Likewise.
11194 (altivec_vpks<VI_char>ss): Likewise.
11195 (altivec_vpks<VI_char>us): Likewise.
11196 (altivec_vpku<VI_char>us): Likewise.
11197 (altivec_vpku<VI_char>um): Likewise.
11198 (altivec_vrl<VI_char>): Likewise.
11199 (altivec_vsl<VI_char>): Likewise.
11200 (altivec_vsr<VI_char>): Likewise.
11201 (altivec_vsra<VI_char>): Likewise.
11202 (altivec_vsldoi_<mode>): Likewise.
11203 (altivec_vupkhsb): Likewise.
11204 (altivec_vupkhs<VU_char>): Likewise.
11205 (altivec_vupkls<VU_char>): Likewise.
11206 (altivec_vupkhsh): Likewise.
11207 (altivec_vupklsb): Likewise.
11208 (altivec_vupklsh): Likewise.
11209 (altivec_vcmpequ<VI_char>_p): Likewise.
11210 (altivec_vcmpgts<VI_char>_p): Likewise.
11211 (altivec_vcmpgtu<VI_char>_p): Likewise.
11212 (abs<mode>2): Likewise.
11213 (vec_unpacks_hi_v16qi): Likewise.
11214 (vec_unpacks_hi_v8hi): Likewise.
11215 (vec_unpacks_lo_v16qi): Likewise.
11216 (vec_unpacks_hi_<VP_small_lc>): Likewise.
11217 (vec_unpacks_lo_v8hi): Likewise.
11218 (vec_unpacks_lo_<VP_small_lc>): Likewise.
11219 (vec_pack_trunc_v8h): Likewise.
11220 (vec_pack_trunc_v4si): Likewise.
11221 (vec_pack_trunc_<mode>): Likewise.
11223 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
11225 (vec_vmaxsd): Likewise.
11226 (vec_vmaxud): Likewise.
11227 (vec_vminsd): Likewise.
11228 (vec_vminud): Likewise.
11229 (vec_vpksdss): Likewise.
11230 (vec_vpksdus): Likewise.
11231 (vec_vpkudum): Likewise.
11232 (vec_vpkudus): Likewise.
11233 (vec_vrld): Likewise.
11234 (vec_vsld): Likewise.
11235 (vec_vsrad): Likewise.
11236 (vec_vsrd): Likewise.
11237 (vec_vsubudm): Likewise.
11238 (vec_vupkhsw): Likewise.
11239 (vec_vupklsw): Likewise.
11241 2013-05-29 Jan Hubicka <jh@suse.cz>
11243 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
11244 flags; reorder rest of fields in more consistent way.
11245 (varpool_node): Remove analyzed, finalized and alias.
11246 (cgraph_ndoe): Likewise.
11247 (symtab_alias_ultimate_target): New function.
11248 (cgraph_function_node): Move offline.
11249 (cgraph_reset_node): Declare.
11250 (cgraph_comdat_can_be_unshared_p): Remove.
11251 (varpool_remove_initializer): Declare.
11252 (varpool_first_defined_variable, varpool_next_defined_variable
11253 cgraph_first_defined_function, cgraph_next_defined_function): Update.
11254 (cgraph_function_with_gimple_body_p): Update.
11255 (varpool_all_refs_explicit_p): Update.
11256 (symtab_alias_target): New function.
11257 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
11258 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
11259 (cgraph_function_or_thunk_node): Simplify using
11260 symtab_alias_ultimate_target.
11261 (varpool_variable_node): Likewise.
11262 * cgraph.c (cgraph_create_function_alias): Update.
11263 (cgraph_add_thunk): Update.
11264 (cgraph_remove_node): Update.
11265 (dump_cgraph_node): Do not dump removed flags.
11266 (cgraph_function_body_availability): Update.
11267 (cgraph_propagate_frequency): Update.
11268 (verify_cgraph_node): Check sanity of local flag.
11269 (cgraph_function_node): Move here from cgraph.h; revamp for
11270 cgraph_function_or_thunk_node.
11271 * lto-symtab.c (lto_varpool_replace_node): Update.
11272 (lto_symtab_resolve_can_prevail_p): Update.
11273 (lto_symtab_merge_cgraph_nodes): Update.
11274 * ipa-cp.c (determine_versionability, initialize_node_lattices,
11275 propagate_constants_accross_call, devirtualization_time_bonus,
11276 ipcp_propagate_stage): Update.
11277 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
11278 * ipa-inline-transform.c (clone_inlined_nodes,
11279 preserve_function_body_p): Update.
11280 * ipa-reference.c (propagate): Update.
11281 (write_node_summary_p): Update.
11282 * toplev.c (wrapup_global_declaration_2): Update.
11283 * cgraphunit.c (cgraph_analyze_function): Rename to ...
11284 (analyze_function) ... this one.
11285 (cgraph_process_new_functions): Update.
11286 (cgraph_reset_node): Export.
11287 (cgraph_finalize_function): Update.
11288 (cgraph_add_new_function): Update.
11289 (process_function_and_variable_attributes): Update.
11290 (varpool_finalize_decl): Update.
11291 (symbol_finalized): Remove.
11292 (symbol_finalized_and_needed): Rename to ...
11293 (symbol_defined_and_needed): ... update.
11294 (cgraph_analyze_functions): Update.
11295 (handle_alias_pairs): Update.
11296 (mark_functions_to_output): Update.
11297 (assemble_thunk): Update.
11298 (output_in_order): Update.
11299 (output_weakrefs): Update.
11300 (finalize_compilation_unit): Update.
11301 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
11302 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
11303 input_node, input_varpool_node): Update.
11304 * dbxout.c (dbxout_expand_expr): Update.
11305 * cgraphclones.c (cgraph_clone_node): Update.
11306 (cgraph_copy_node_for_versioning): Update.
11307 (cgraph_materialize_clone): Update.
11308 (cgraph_materialize_all_clones): Update.
11309 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
11310 propagate_pure_const, propagate_nothrow): Update.
11311 * lto-streamer-out.c (lto_output, write_symbol): Update.
11312 * ipa-utils.c (ipa_reverse_postorder): Update.
11313 * ipa-inline.c (can_inline_edge_p): Update.
11314 (update_caller_keys, ipa_inline): Update.
11315 * dwarf2out.c (reference_to_unused,
11316 premark_types_used_by_global_vars_helper): Update.
11317 * tree-eh.c (tree_could_trap_p): Update.
11318 * ipa-split.c (consider_split, execute_split_functions): Update.
11319 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
11320 has_addr_references_p): Update; move ahead in file for better
11322 (process_references): Simplify.
11323 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
11324 bodies are removed.
11325 (cgraph_comdat_can_be_unshared_p): Make static.
11326 (cgraph_externally_visible_p): Update.
11327 (varpool_externally_visible_p): Update.
11328 (function_and_variable_visibility): Update.
11329 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
11330 ipa_tm_mark_force_output_node): Update.
11331 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
11332 estimate_edge_devirt_benefit, inline_generate_summary,
11333 inline_write_summary): Update.
11334 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
11335 * ipa-prop.c (ipa_compute_jump_functions): Update.
11336 (ipa_print_node_params, ipa_prop_read_section,
11337 ipa_update_after_lto_read, read_replacements_section): Update.
11338 * varasm.c (mark_decl_referenced): Update.
11339 (assemble_alias, dump_tm_clone_pairs): Update.
11340 * tree-inline.c (copy_bb): Update.
11341 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
11343 * symtab.c (dump_symtab_base): Print new flags.
11344 (verify_symtab_base): Verify new flags.
11345 (symtab_alias_ultimate_target): New function.
11346 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
11347 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
11349 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
11351 * i386.c (ix86_get_function_versions_dispatcher,
11352 ix86_generate_version_dispatcher_body): Update.
11353 (fold_builtin_cpu): Use varpool_add_new_variable.
11354 * varpool.c (varpool_remove_initializer): Break out from ...
11355 (varpool_remove_node): ... this one.
11356 (dump_varpool_node, varpool_node_for_asm,
11357 cgraph_variable_initializer_availability, varpool_analyze_node,
11358 varpool_assemble_decl, varpool_remove_unreferenced_decls,
11359 varpool_finalize_named_section_flags, varpool_create_variable_alias):
11362 2013-05-29 Jan Hubicka <jh@suse.cz>
11364 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
11366 2013-05-29 Easwaran Raman <eraman@google.com>
11368 PR tree-optimization/57442
11369 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
11370 when control exits the main loop.
11372 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
11374 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
11376 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
11377 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
11378 * rx/t-rx: Add rx100 under multi library matches option for nofpu
11381 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11383 PR tree-optimization/57441
11384 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
11385 Don't limit size of incr_vec to number of candidates.
11387 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
11389 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
11390 and mips16 directories.
11391 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
11392 (MULTILIB_DIRNAMES): Ditto.
11393 (MULTILIB_EXCEPTIONS): Add new exceptions.
11394 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
11395 (MULTILIB_DIRNAMES): Ditto.
11396 (MULTILIB_EXCEPTIONS): Add new exceptions.
11398 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11399 Marcus Shawcroft <marcus.shawcroft@arm.com>
11401 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
11402 SYMBOL_TINY_ABSOLUTE.
11403 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
11404 SYMBOL_TINY_ABSOLUTE.
11405 (aarch64_expand_mov_immediate): Likewise.
11406 (aarch64_classify_symbol): Likewise.
11407 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
11408 Permit SYMBOL_TINY_ABSOLUTE.
11409 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
11411 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11412 Marcus Shawcroft <marcus.shawcroft@arm.com>
11414 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
11415 Refactor if/switch. Replace gcc_assert with if.
11417 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11419 * config/i386/i386.c (initial_ix86_tune_features): Enable
11420 FP Reassociation for AMD bdver1 and bdver2.
11422 2013-05-29 Martin Jambor <mjambor@suse.cz>
11424 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
11425 and IMAGPART_EXPR do not occur within other handled_components.
11427 2013-05-29 Richard Biener <rguenther@suse.de>
11429 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
11430 access on whether the use is in the BB we currently try to
11432 (vect_bb_vectorization_profitable_p): Pass the BB we currently
11433 vectorize to vect_bb_slp_scalar_cost.
11435 2013-05-29 Richard Biener <rguenther@suse.de>
11437 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
11438 computing scalar cost offsetted by stmts that are kept live
11440 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
11441 for computation of scalar cost.
11443 2013-05-28 Steve Ellcey <sellcey@mips.com>
11445 * config/mips/mips-cpus.def (mips32r2): Change processor type.
11447 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
11449 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
11450 array notation built-in reduction functions.
11451 * doc/passes.texi (Passes): Added documentation about changes done
11453 * doc/invoke.texi (C Dialect Options): Added documentation about
11454 the -fcilkplus flag.
11455 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
11456 (BUILTINS_DEF): Depend on cilkplus.def.
11457 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
11458 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
11459 * cilkplus.def: New file.
11461 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
11463 PR rtl-optimization/57439
11464 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
11466 2013-05-28 Easwaran Raman <eraman@google.com>
11468 PR tree-optimization/57337
11469 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
11470 (find_insert_point): Correctly identify the insertion point
11471 when two statements with the same UID is compared.
11473 2013-05-28 Richard Biener <rguenther@suse.de>
11475 PR tree-optimization/56787
11476 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
11477 from the list of data references.
11478 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
11480 (vect_analyze_loop_operations): Likewise.
11481 (vect_transform_loop): Remove clobbers.
11483 2013-05-28 Martin Jambor <mjambor@suse.cz>
11485 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
11486 and REALPART_EXPRs have scalar type.
11488 2013-05-28 Richard Biener <rguenther@suse.de>
11490 PR tree-optimization/57411
11491 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
11493 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
11494 virtual operand propagation.
11496 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
11498 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
11499 destination register for bmasksi_vis.
11500 (vector_init_bshuffle): Likewise.
11501 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
11503 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
11505 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
11506 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
11507 mode if the instruction isn't available in the original mode.
11508 * config/sparc/sparc.opt (mfix-ut699): New option.
11509 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
11510 (divdf3): Turn into expander.
11511 (divdf3_nofix): New insn.
11512 (divdf3_fix): Likewise.
11513 (divsf3): Disable if -mfix-ut699.
11514 (sqrtdf2): Turn into expander.
11515 (sqrtdf2_nofix): New insn.
11516 (sqrtdf2_fix): Likewise.
11517 (sqrtsf2): Disable if -mfix-ut699.
11519 2013-05-27 Richard Biener <rguenther@suse.de>
11521 PR middle-end/57412
11522 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
11523 block for the new loop.
11525 2013-05-27 Richard Biener <rguenther@suse.de>
11527 PR tree-optimization/57343
11528 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
11529 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
11530 (number_of_iterations_cond): Do not build the folded tree.
11532 2013-05-27 Richard Biener <rguenther@suse.de>
11535 PR middle-end/57381
11536 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
11537 OEP_CONSTANT_ADDRESS_OF retained.
11539 PR tree-optimization/57417
11540 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
11541 for unchanged base.
11542 (set_ssa_val_to): Compare addresses using
11543 get_addr_base_and_unit_offset.
11545 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
11547 PR rtl-optimization/56833
11548 * postreload.c (move2add_record_mode): New function.
11549 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
11550 (move2add_use_add2_insn): Use move2add_record_sym_value.
11551 (move2add_use_add3_insn): Likewise.
11552 (reload_cse_move2add): Use move2add_valid_value_p and
11553 move2add_record_mode. Invalidate call-clobbered and REG_INC
11554 affected regs by setting reg_mode to VOIDmode.
11555 (move2add_note_store): Don't pretend the inside of a SUBREG is
11556 the actual destination. Invalidate single/leading registers by
11557 setting reg_mode to VOIDmode.
11558 Use move2add_record_sym_value, move2add_valid_value_p and
11559 move2add_record_mode.
11561 2013-05-27 Richard Biener <rguenther@suse.de>
11563 PR tree-optimization/57396
11564 * tree-affine.c (double_int_constant_multiple_p): Properly
11565 return false for val == 0 and div != 0.
11567 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11569 * config/mips/mips.h: Use #elif in preprocessor conditions.
11571 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11574 * config/mips/constraints.md (kl): New constraint.
11575 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
11576 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
11577 constraint for operand 0. Split after CSE for MIPS16. Emit a move
11578 from LO for MIPS16.
11579 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
11581 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11584 * config/mips/mips.c (mips_can_inline_p): New function.
11585 (TARGET_CAN_INLINE_P): Define.
11587 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
11589 * sched-int.h (ds_t, dw_t): Make unsigned int.
11590 Fix documentation that describes how all the ds_t bits are used.
11591 Reserve the last bit for delayed-branch scheduling.
11592 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
11593 (BITS_PER_DEP_WEAK): Fix definition and documentation.
11594 (gen_dep_weak_1): Remove prototype.
11595 * sched-deps.c (get_dep_weak_1): Make static.
11596 * target.def (speculate_insn, needs_block_p, gen_spec_check,
11597 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
11598 * doc/tm.texi: Regenerate.
11599 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
11601 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
11604 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
11606 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
11607 Sandra Loosemore <sandra@codesourcery.com>
11609 * config.gcc (powerpc-*): Allow native for with-cpu.
11611 2013-05-24 Jeff Law <law@redhat.com>
11613 PR tree-optimization/57124
11614 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
11615 conversion feeding a condition if the range has an overflow
11616 if -fstrict-overflow. Add warnings for when we do make the
11619 2013-05-24 Dehao Chen <dehao@google.com>
11621 * tree-cfg.c (locus_discrim_map): Fix the typo.
11622 (locus_discrim_hasher): Likewise.
11623 (locus_discrim_hasher::hash): Likewise.
11624 (locus_discrim_hasher::equal): Likewise.
11626 2013-05-24 Martin Jambor <mjambor@suse.cz>
11628 PR tree-optimization/57294
11629 * cgraph.h (ipa_record_stmt_references): Declare.
11630 * cgraphbuild.c (ipa_record_stmt_references): New function.
11631 (build_cgraph_edges): Use ipa_record_stmt_references.
11632 (rebuild_cgraph_edges): Likewise.
11633 (cgraph_rebuild_references): Likewise.
11634 * ipa-prop.c (ipa_modify_call_arguments): Discard references
11635 associated with the old statement and build references from the
11636 newly built statements.
11637 * ipa-ref.c (ipa_remove_stmt_references): New function.
11638 * ipa-ref.h (ipa_remove_stmt_references): Declare.
11640 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
11642 * lra-constraints.c (emit_spill_move): Use smaller mode for
11644 (check_and_process_move): Consider mem-reg moves for secondary
11646 (curr_insn_transform): Don't lose insns emitted before for
11647 secondary memory moves.
11648 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
11649 reg set up in the current insn.
11651 2013-05-24 Dehao Chen <dehao@google.com>
11653 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
11655 (locus_descrim_hasher::equal): Likewise.
11656 (build_gimple_cfg): New discriminator assignment algorithm.
11657 (make_edges): Likewise.
11658 (next_discriminator_for_locus): Likewise.
11659 (same_line_p): Likewise.
11660 (assign_discriminators): Likewise.
11661 (make_cond_expr_edges): Likewise.
11662 (make_gimple_switch_edges): Likewise.
11663 (make_goto_expr_edges): Likewise.
11664 (make_gimple_asm_edges): Likewise.
11666 2013-05-24 Ian Bolton <ian.bolton@arm.com>
11668 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
11669 X format specifier to only display bottom 16 bits.
11670 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
11671 immediate to match for operand 2, since it will be masked.
11673 2013-05-24 Richard Biener <rguenther@suse.de>
11675 PR tree-optimization/57287
11676 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
11677 all SSA names that occur in abnormal PHIs.
11679 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
11682 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
11683 that index is not negative.
11685 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
11687 PR rtl-optimization/55177
11688 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
11689 (simplify_byte_swapping_operation): New.
11690 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
11691 (simplify_relational_operation_1): Deal with BSWAP.
11693 2013-05-23 Richard Henderson <rth@redhat.com>
11696 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
11697 (ix86_reorg): Call it.
11699 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
11702 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
11703 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
11704 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
11706 2013-05-23 Christian Bruel <christian.bruel@st.com>
11709 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
11711 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11712 Marcus Shawcroft <marcus.shawcroft@arm.com>
11714 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
11715 * config/aarch64/constraints.md (Usa): Remove.
11716 * doc/md.texi (AArch64 Usa): Remove.
11718 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11719 Marcus Shawcroft <marcus.shawcroft@arm.com>
11721 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
11722 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
11723 * config/aarch64/predicates.md (aarch64_const_address): Remove.
11724 (aarch64_mov_operand): Use aarch64_mov_operand_p.
11726 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
11728 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
11729 instruction (AdvSIMD).
11730 * config/aarch64/aarch64-builtins.c
11731 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
11732 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
11734 2013-05-23 Martin Jambor <mjambor@suse.cz>
11736 PR middle-end/57347
11737 * tree.h (contains_bitfld_component_ref_p): Declare.
11738 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
11739 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
11741 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
11742 not access a bit-field. Assert all final offsets are byte-aligned.
11744 2013-05-23 Richard Biener <rguenther@suse.de>
11746 PR tree-optimization/57380
11747 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
11748 least one invariant or re-used load.
11749 * passes.c (init_optimization_passes): Move pass_phiprop before
11752 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
11754 * config/aarch64/aarch64-simd.md
11755 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
11757 2013-05-23 Richard Biener <rguenther@suse.de>
11759 PR middle-end/57381
11760 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
11761 OEP_CONSTANT_ADDRESS_OF retained.
11763 2013-05-23 Jakub Jelinek <jakub@redhat.com>
11765 PR middle-end/57344
11766 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
11767 don't lower unit. Handle unit not being always BITS_PER_WORD.
11769 2013-05-23 Richard Biener <rguenther@suse.de>
11771 PR rtl-optimization/57341
11772 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
11773 instead of true_dependence.
11775 2013-05-22 David Malcolm <dmalcolm@redhat.com>
11777 * bb-reorder.c (branch_threshold): Make const.
11778 (exec_threshold): Ditto.
11780 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11781 Pat Haugen <pthaugen@us.ibm.com>
11782 Peter Bergner <bergner@vnet.ibm.com>
11784 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
11785 documentation for the power8 crypto builtins.
11787 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
11789 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
11790 macros for defining power8 builtin functions.
11791 (BU_P8V_AV_2): Likewise.
11792 (BU_P8V_AV_P): Likewise.
11793 (BU_P8V_VSX_1): Likewise.
11794 (BU_P8V_OVERLOAD_1): Likewise.
11795 (BU_P8V_OVERLOAD_2): Likewise.
11796 (BU_CRYPTO_1): Likewise.
11797 (BU_CRYPTO_2): Likewise.
11798 (BU_CRYPTO_3): Likewise.
11799 (BU_CRYPTO_OVERLOAD_1): Likewise.
11800 (BU_CRYPTO_OVERLOAD_2): Likewise.
11801 (XSCVSPDP): Fix typo, point to the correct instruction.
11802 (VCIPHER): Add power8 crypto builtins.
11803 (VCIPHERLAST): Likewise.
11804 (VNCIPHER): Likewise.
11805 (VNCIPHERLAST): Likewise.
11806 (VPMSUMB): Likewise.
11807 (VPMSUMH): Likewise.
11808 (VPMSUMW): Likewise.
11809 (VPERMXOR_V2DI): Likewise.
11810 (VPERMXOR_V4SI: Likewise.
11811 (VPERMXOR_V8HI: Likewise.
11812 (VPERMXOR_V16QI: Likewise.
11813 (VSHASIGMAW): Likewise.
11814 (VSHASIGMAD): Likewise.
11815 (VPMSUM): Likewise.
11816 (VPERMXOR): Likewise.
11817 (VSHASIGMA): Likewise.
11819 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11820 __CRYPTO__ if the crypto instructions are available.
11821 (altivec_overloaded_builtins): Add support for overloaded power8
11824 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
11825 support for power8 crypto builtins.
11826 (builtin_function_type): Likewise.
11827 (altivec_init_builtins): Add support for builtins that take vector
11828 long long (V2DI) arguments.
11830 * config/rs6000/crypto.md: New file, define power8 crypto
11833 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11834 Pat Haugen <pthaugen@us.ibm.com>
11835 Peter Bergner <bergner@vnet.ibm.com>
11837 * doc/invoke.texi (Option Summary): Add power8 options.
11838 (RS/6000 and PowerPC Options): Likewise.
11840 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
11841 constraints.md instead of rs6000.h. Reorder w* constraints. Add
11842 wm, wn, wr documentation.
11844 * config/rs6000/constraints.md (wm): New constraint for VSX
11845 registers if direct move instructions are enabled.
11846 (wn): New constraint for no registers.
11847 (wq): New constraint for quad word even GPR registers.
11848 (wr): New constraint if 64-bit instructions are enabled.
11849 (wv): New constraint if power8 vector instructions are enabled.
11850 (wQ): New constraint for quad word memory locations.
11852 * config/rs6000/predicates.md (const_0_to_15_operand): New
11853 constraint for 0..15 for crypto instructions.
11854 (gpc_reg_operand): If VSX allow registers in VSX registers as well
11855 as GPR and floating point registers.
11856 (int_reg_operand): New predicate to match only GPR registers.
11857 (base_reg_operand): New predicate to match base registers.
11858 (quad_int_reg_operand): New predicate to match even GPR registers
11859 for quad memory operations.
11860 (vsx_reg_or_cint_operand): New predicate to allow vector logical
11861 operations in both GPR and VSX registers.
11862 (quad_memory_operand): New predicate for quad memory operations.
11863 (reg_or_indexed_operand): New predicate for direct move support.
11865 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
11866 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
11867 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
11868 (POWERPC_MASKS): Add power8 options.
11869 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
11872 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11873 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
11875 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
11876 (-mpower8-fusion): New power8 options.
11877 (-mpower8-fusion-sign): Likewise.
11878 (-mpower8-vector): Likewise.
11879 (-mcrypto): Likewise.
11880 (-mdirect-move): Likewise.
11881 (-mquad-memory): Likewise.
11883 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
11884 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
11886 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
11887 (rs6000_debug_vector_unit): Add p8_vector.
11888 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
11889 definitions. Also print fusion state.
11890 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
11891 (rs6000_builtin_mask_calculate): Add power8 builtin support.
11892 (rs6000_option_override_internal): Add support for power8.
11893 (rs6000_common_init_builtins): Add debugging for skipped builtins
11894 if -mdebug=builtin.
11895 (rs6000_adjust_cost): Add power8 support.
11896 (rs6000_issue_rate): Likewise.
11897 (insn_must_be_first_in_group): Likewise.
11898 (insn_must_be_last_in_group): Likewise.
11899 (force_new_group): Likewise.
11900 (rs6000_register_move_cost): Likewise.
11901 (rs6000_opt_masks): Likewise.
11903 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
11904 power8 capable assembler, default to power7 options.
11905 (TARGET_DIRECT_MOVE): Likewise.
11906 (TARGET_CRYPTO): Likewise.
11907 (TARGET_P8_VECTOR): Likewise.
11908 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
11909 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
11910 (VECTOR_MEM_P8_VECTOR_P): Likewise.
11911 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
11912 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
11913 (TARGET_XSCVDPSPN): Likewise.
11914 (TARGET_XSCVSPDPN): Likewsie.
11915 (TARGET_SYNC_HI_QI): Likewise.
11916 (TARGET_SYNC_TI): Likewise.
11917 (MASK_CRYPTO): Likewise.
11918 (MASK_DIRECT_MOVE): Likewise.
11919 (MASK_P8_FUSION): Likewise.
11920 (MASK_P8_VECTOR): Likewise.
11921 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
11922 temporary used by some of the direct move instructions to get two FP
11923 temporary registers does not force creation of a stack frame.
11924 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
11925 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
11926 that any VSX registers are tieable, even if they are also an
11927 Altivec vector mode.
11928 (r6000_reg_class_enum): Add wm, wr, wv constraints.
11929 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
11930 (RS6000_BTM_CRYPTO): Likewise.
11931 (RS6000_BTM_COMMON): Likewise.
11933 * config/rs6000/rs6000.md (cpu attribute): Add power8.
11934 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
11935 (enum rs6000_vector): Add power8 vector support.
11937 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11941 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
11942 (any_sibcall_could_use_r3): this and handle indirect calls.
11943 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
11945 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11947 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
11949 2013-05-22 Richard Biener <rguenther@suse.de>
11951 PR middle-end/57349
11952 * profile.c (branch_prob): Do not split blocks that are
11953 abnormally receiving from ECF_RETURNS_TWICE functions.
11955 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11957 * recog.c (offsettable_address_addr_space_p): Fix calculation of
11958 address mode. Move pointer mode initialization to the same place.
11960 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11962 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
11963 while it has any effect.
11965 2013-05-21 Easwaran Raman <eraman@google.com>
11967 PR tree-optimization/57322
11968 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
11969 UID of the statement added to the BB to be 1.
11971 2013-05-21 Jakub Jelinek <jakub@redhat.com>
11973 PR tree-optimization/57331
11974 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
11975 of conversion from pointer type to integral type with integer.
11977 2013-05-21 Martin Jambor <mjambor@suse.cz>
11980 * ipa-prop.c (ipa_read_node_info): Process param_used and
11981 controlled_uses in the same order as when writing.
11983 2013-05-21 Magnus Granberg <baldrick@free.fr>
11986 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
11988 2013-05-21 Richard Biener <rguenther@suse.de>
11990 PR tree-optimization/57318
11991 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
11992 estimate stmts with side-effects as likely eliminated.
11994 2013-05-21 Richard Biener <rguenther@suse.de>
11996 PR tree-optimization/57330
11997 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
11998 preserve the call stmts fntype.
12000 2013-05-21 Richard Biener <rguenther@suse.de>
12002 PR tree-optimization/57303
12003 * tree-ssa-sink.c (statement_sink_location): Improve killing
12004 stmt detection and properly handle self-assignments.
12006 2013-05-21 Christian Bruel <christian.bruel@st.com>
12008 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12009 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12010 registers. Set register size out of the PARALLEL loop.
12012 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12015 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12016 (*fmasf4, *fmasf4_media): New insns.
12018 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12020 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12021 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12022 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12023 (mips_idiv_insns): Update the comments to say that the returned
12024 instruction counts are in units of BASE_INSN_LENGTH.
12025 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12026 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12027 using 2 rather than 4 as the length of indirect MIPS16 and
12028 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12029 length of a NOP. Don't divide MIPS16 lengths by 2.
12030 (mips16_split_long_branches): Assume a branch is long if the
12031 length is greater than 4 rather than 8.
12032 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12033 rather than multiplying them by 2. Multiply instruction counts
12034 by BASE_INSN_LENGTH rather than 4.
12035 (*jump_mips16, tls_get_tp_mips16_<mode>)
12036 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12038 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12040 * config/mips/mips.md (extended_mips16): Remove branch case.
12041 (length): Remove duplicated extended_mips16 test.
12043 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12045 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12047 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12049 * recog.h (Recog_data): Rename to...
12050 (recog_data_d): ...this.
12051 (recog_data): Update accordingly.
12052 * recog.c (recog_data): Likewise.
12053 * reload.c (save_recog_data): Likewise.
12054 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
12055 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
12057 2013-05-17 Julian Brown <julian@codesourcery.com>
12059 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
12060 found in a REG_EQUAL note, invalidate it.
12062 2013-05-17 Easwaran Raman <eraman@google.com>
12064 * tree-ssa-reassoc.c (find_insert_point): New function.
12065 (insert_stmt_after): Likewise.
12066 (get_def_stmt): Likewise.
12067 (ensure_ops_are_available): Likewise.
12068 (not_dominated_by): Likewise.
12069 (rewrite_expr_tree): Do not move statements beyond what is
12070 necessary. Remove call to swap_ops_for_binary_stmt...
12071 (reassociate_bb): ... and move it here.
12072 (build_and_add_sum): Assign UIDs for new statements.
12073 (linearize_expr): Likewise.
12074 (do_reassoc): Renumber gimple statement UIDs.
12076 2013-05-17 Jan Hubicka <jh@suse.cz>
12078 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
12080 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
12081 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
12083 (output_weakrefs): Update.
12085 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
12086 Martin Jambor <mjambor@suse.cz>
12088 PR middle-end/57276
12089 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
12090 value that corresponds to the given aggval is found in values vector.
12092 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
12094 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
12095 sse, sse2, sse3, ssse3 and sse4a flags to options.
12097 2013-05-17 David Malcolm <dmalcolm@redhat.com>
12099 * gengtype-state.c: (s_expr_writer): New class, to handle
12100 prettifying of output layout of s-expressions.
12101 (state_writer): New class, to write out gtype.state.
12102 (state_written_type_count): Move this variable into member data of
12104 (s_expr_writer::s_expr_writer): New code: constructor for new class
12105 (state_writer::state_writer(): ditto
12106 (s_expr_writer::write_new_line): New function
12107 (s_expr_writer::write_any_indent): ditto
12108 (s_expr_writer::begin_s_expr): ditto
12109 (s_expr_writer::end_s_expr): ditto
12110 (write_state_fileloc): convert to method of state_writer...
12111 (state_writer:: write_state_fileloc): ...and use methods of
12112 s_expr_writer to write indentation into the gtype.state output file
12113 to visually represent the hierarchical structure of the list
12115 (write_state_fields): ditto, renaming to...
12116 (state_writer::write_state_fields)
12117 (write_state_a_string): ditto, renaming to...
12118 (state_writer::write_state_a_string)
12119 (write_state_string_option): ditto, renaming to...
12120 (state_writer::write_state_string_option)
12121 (write_state_type_option): ditto, renaming to...
12122 (state_writer::write_state_type_option)
12123 (write_state_nested_option): ditto, renaming to...
12124 (state_writer::write_state_nested_option)
12125 (write_state_option): ditto, renaming to...
12126 (state_writer::write_state_option)
12127 (write_state_options): ditto, renaming to...
12128 (state_writer::write_state_options)
12129 (write_state_lang_bitmap): ditto, renaming to...
12130 (state_writer::write_state_lang_bitmap)
12131 (write_state_version): ditto, renaming to...
12132 (state_writer::write_state_version)
12133 (write_state_scalar_type): ditto, renaming to...
12134 (state_writer::write_state_scalar_type)
12135 (write_state_string_type): ditto, renaming to...
12136 (state_writer::write_state_string_type)
12137 (write_state_undefined_type): ditto, renaming to...
12138 (state_writer::write_state_undefined_type)
12139 (write_state_struct_union_type): ditto, renaming to...
12140 (state_writer::write_state_struct_union_type)
12141 (write_state_struct_type): ditto, renaming to...
12142 (state_writer::write_state_struct_type)
12143 (write_state_user_struct_type): ditto, renaming to...
12144 (state_writer::write_state_user_struct_type)
12145 (write_state_lang_struct_type): ditto, renaming to...
12146 (state_writer::write_state_lang_struct_type)
12147 (write_state_param_struct_type): ditto, renaming to...
12148 (state_writer::write_state_param_struct_type)
12149 (write_state_pointer_type): ditto, renaming to...
12150 (state_writer::write_state_pointer_type)
12151 (write_state_array_type): ditto, renaming to...
12152 (state_writer::write_state_array_type)
12153 (write_state_gc_used): ditto, renaming to...
12154 (state_writer::write_state_gc_used)
12155 (write_state_common_type_content): ditto, renaming to...
12156 (state_writer::write_state_common_type_content)
12157 (write_state_type): ditto, renaming to...
12158 (state_writer::write_state_type)
12159 (write_state_pair_list): ditto, renaming to...
12160 (state_writer::write_state_pair_list)
12161 (write_state_pair): ditto, renaming to...
12162 (state_writer::write_state_pair)
12163 (write_state_typedefs): ditto, renaming to...
12164 (state_writer::write_state_typedefs)
12165 (write_state_structures): ditto, renaming to...
12166 (state_writer::write_state_structures)
12167 (write_state_param_structs): ditto, renaming to...
12168 (state_writer::write_state_param_structs)
12169 (write_state_variables): ditto, renaming to...
12170 (state_writer::write_state_variables)
12171 (write_state_srcdir): ditto, renaming to...
12172 (state_writer::write_state_srcdir)
12173 (write_state_files_list): ditto, renaming to...
12174 (state_writer::write_state_files_list)
12175 (write_state_languages): ditto, renaming to...
12176 (state_writer::write_state_languages)
12177 (write_state): create a state_writer instance and use it when
12178 writing out the state file
12180 2013-05-17 Mike Stump <mikestump@comcast.net>
12182 PR rtl-optimization/57304
12183 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
12184 accessing DF_REF_REAL_LOC.
12186 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12188 PR rtl-optimization/57281
12189 PR rtl-optimization/57300
12190 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
12191 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
12192 what the other splitter did if the registers are dead.
12194 2013-05-17 Richard Biener <rguenther@suse.de>
12196 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
12199 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12201 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
12204 2013-05-17 Marek Polacek <polacek@redhat.com>
12206 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
12207 length when doing non-zero store of storing '\0' to '\0'.
12209 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12211 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
12212 vect_external_def oprnd1 with loop_vinfo, try to emit
12213 optional cast, negation and and stmts on the loop preheader
12214 edge instead of into the pattern def seq.
12216 PR tree-optimization/57051
12217 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
12218 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
12220 2013-05-16 Nick Clifton <nickc@redhat.com>
12222 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
12223 (rl78_is_naked_func): New function.
12224 (rl78_expand_prologue): Skip prologue generation for naked functions.
12225 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
12226 * doc/extend.texi (naked): Add RL78 to the list of processors
12227 that supports this attribute.
12229 2013-05-16 Jeff Law <law@redhat.com>
12231 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
12233 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12235 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
12236 cache parameters using detect_caches_amd also for CYRIX,
12237 NSC and TM2 signatures.
12239 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12240 Dzianis Kahanovich <mahatma@eu.by>
12244 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12245 VIA/Centaur processors and determine their cache parameters
12246 using detect_caches_amd.
12248 2013-05-16 Teresa Johnson <tejohnson@google.com>
12250 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
12251 (rtl_verify_edges): New function.
12252 (rtl_verify_bb_insns): Ditto.
12253 (rtl_verify_bb_pointers): Ditto.
12254 (rtl_verify_bb_insn_chain): Ditto.
12255 (rtl_verify_fallthru): Ditto.
12256 (rtl_verify_bb_layout): Ditto.
12257 (rtl_verify_flow_info_1): Outline checks into new functions.
12258 (rtl_verify_flow_info): Ditto.
12260 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
12262 * cfghooks.c (copy_bbs): Add update_dominance argument.
12263 * cfghooks.h (copy_bbs): Update prototype.
12264 * tree-cfg.c (gimple_duplicate_sese_region):
12265 Add update_dominance argument.
12266 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
12267 * tree-ssa-loop-ch.c (copy_loop_headers): Update
12268 gimple_duplicate_sese_region call.
12269 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
12270 Update copy_bbs call.
12271 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
12272 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
12274 2013-05-16 Jakub Jelinek <jakub@redhat.com>
12276 * tree-vectorizer.h (NUM_PATTERNS): Increment.
12277 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
12278 vect_recog_rotate_pattern.
12279 (vect_recog_rotate_pattern): New function.
12281 2013-05-16 Jason Merrill <jason@redhat.com>
12283 * Makefile.in (LLINKER): New variable.
12284 (mostlyclean): Remove link mutex.
12285 * configure.ac: Handle --enable-link-mutex.
12286 * lock-and-run.sh: New script.
12288 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12291 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
12294 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12296 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
12298 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
12300 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
12301 * config/arm/arm.c (next_consecutive_mem): New function.
12302 (gen_movmem_ldrd_strd): Likewise.
12303 * config/arm/arm.md (movmemqi): Update condition and code.
12304 (unaligned_loaddi, unaligned_storedi): New patterns.
12306 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12308 * config.gcc: Obsolete *-*-solaris2.9*.
12309 * doc/install.texi (Specific, *-*-solaris2*): Document it.
12311 2013-05-16 Richard Biener <rguenther@suse.de>
12313 * passes.c (init_optimization_passes): Move pass_parallelize_loops
12314 earlier, after GRAPHITE transforms and IV canonicalization.
12316 2013-05-16 Jakub Jelinek <jakub@redhat.com>
12318 * omp-low.c (extract_omp_for_data): For collapsed loops,
12319 if at least one of the loops is known at compile time to
12320 iterate zero times, set count to 0.
12321 (expand_omp_regimplify_p): New function.
12322 (expand_omp_for_generic): For collapsed loops, if at least
12323 one of the loops isn't known to iterate at least once,
12324 add runtime check with setting count to 0.
12325 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
12326 For unsigned types if it isn't known at compile time that
12327 the loop will iterate at least once, add runtime check to bypass
12328 the whole loop if initial condition isn't true.
12330 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
12332 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
12334 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
12336 PR middle-end/57286
12337 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
12338 transformations to avoid an infinite loop.
12340 2013-05-16 Marek Polacek <polacek@redhat.com>
12342 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
12344 2013-05-15 Leif Ekblad <leif@rdos.net>
12346 * config/i386/i386.c (ix86_decompose_address): Use
12347 DEFAULT_TLS_SEG_REG to access TLS segment register.
12348 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
12349 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
12350 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
12352 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
12355 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
12356 sibling calls to functions that would normally be lazily bound,
12357 unless $gp is call-clobbered.
12359 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
12361 * config/i386/i386.c (ix86_option_override_internal): Update
12362 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
12363 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
12364 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
12365 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
12367 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
12369 2013-05-15 Andreas Schwab <schwab@suse.de>
12371 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
12372 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
12375 2013-05-15 Teresa Johnson <tejohnson@google.com>
12377 * loop-unroll.c (report_unroll_peel): Check decision before
12378 emitting unroll/peel message.
12380 2013-05-15 Teresa Johnson <tejohnson@google.com>
12382 * function.h (has_bb_partition): New rtl_data flag.
12383 (bb_reorder_complete): Ditto.
12384 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
12385 instead of flag_reorder_blocks_and_partition.
12386 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
12387 with some enhancements.
12388 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
12389 * bb-reorder.c (connect_traces): Check for has_bb_partition
12390 instead of flag_reorder_blocks_and_partition.
12391 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
12392 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
12393 verify_hot_cold_block_grouping.
12394 (partition_hot_cold_basic_blocks): Set has_bb_partition.
12396 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12399 * config/arm/predicates.md (call_insn_operand): New predicate.
12400 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
12401 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
12402 if insn is not a tail call.
12403 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
12405 * config/arm/arm.h (enum reg_class): New caller save register class.
12406 (REG_CLASS_NAMES): Likewise.
12407 (REG_CLASS_CONTENTS): Likewise.
12408 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
12411 2013-05-15 Richard Biener <rguenther@suse.de>
12413 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
12414 of MSG_OPTIMIZED_LOCATIONS.
12415 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
12416 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
12418 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
12419 of MSG_OPTIMIZED_LOCATIONS.
12420 (execute_vect_slp): Likewise.
12421 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
12422 (vect_create_cond_for_alias_checks): Likewise.
12423 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
12424 (vect_recog_widen_mult_pattern): Likewise.
12425 (vect_recog_widen_sum_pattern): Likewise.
12426 (vect_recog_over_widening_pattern): Likewise.
12427 (vect_recog_widen_shift_pattern): Likewise.
12428 (vect_recog_vector_vector_shift_pattern): Likewise.
12429 (vect_recog_divmod_pattern): Likewise.
12430 (vect_recog_mixed_size_cond_pattern): Likewise.
12431 (vect_recog_bool_pattern): Likewise.
12432 (vect_pattern_recog_1): Likewise.
12434 2013-05-15 Martin Jambor <mjambor@suse.cz>
12436 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
12437 non-functions to builtin_unreachable.
12438 * ipa-inline-transform.c (inline_call): Do not assert estimates were
12439 correct when new direct edges were discovered.
12441 2013-05-15 Martin Jambor <mjambor@suse.cz>
12443 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
12444 header, print symbol order instead of node uid, print more information
12445 about indirect edge targets.
12446 (ipa_make_edge_direct_to_target): Print symbol order instead of node
12448 (ipa_make_edge_direct_to_target): Likewise.
12449 (remove_described_reference): Likewise.
12450 (propagate_controlled_uses): Likewise.
12451 (ipa_print_node_params): Also print symbol order.
12452 (ipcp_transform_function): Print symbol order instead of node uids.
12453 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
12454 (cgraph_get_create_real_symbol_node): Likewise.
12455 * ipa-cp.c (print_lattice): Likewise.
12456 (print_all_lattices): Likewise.
12457 (determine_versionability): Likewise.
12458 (initialize_node_lattices): Likewise.
12459 (estimate_local_effects): Likewise.
12460 (update_profiling_info): Likewise.
12461 (create_specialized_node): Likewise.
12462 (perhaps_add_new_callers): Likewise.
12463 (decide_about_value): Likewise.
12464 (decide_whether_version_node): Likewise.
12465 (identify_dead_nodes): Likewise.
12466 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
12467 (dump_inline_summary): Likewise.
12468 (estimate_node_size_and_time): Likewise.
12469 (inline_analyze_function): Likewise.
12470 * ipa-inline.c (report_inline_failed_reason): Likewise.
12471 (want_early_inline_function_p): Likewise.
12472 (edge_badness): Likewise.
12473 (update_edge_key): Likewise.
12474 (inline_small_functions): Likewise. Add dumping of order to two other
12476 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
12477 instead of node uids.
12478 (propagate_pure_const): Likewise.
12479 (propagate_pure_const): Likewise.
12480 * ipa-utils.c (dump_cgraph_node_set): Likewise.
12481 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
12482 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
12484 * tree-pretty-print.c (dump_function_header): Likewise.
12485 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
12486 Print symbol order instead of node uids.
12488 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12490 * config/s390/s390.c (s390_register_move_cost): Don't impose the
12491 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
12493 2013-05-15 Richard Biener <rguenther@suse.de>
12495 PR tree-optimization/57275
12496 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
12497 return value for fail to do runtime alias checks for gather loads.
12499 2013-05-15 Jan Hubicka <jh@suse.cz>
12503 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
12504 weakrefs are not external.
12505 (lto_symtab_merge_decls): Fix thinko when dealing with
12506 non-lto_symtab decls.
12507 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
12508 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
12509 * varpool.c (dump_varpool_node): Dump more flags.
12511 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12513 * config/i386/i386.c (processor_alias_table): Add instruction
12514 FSGSBASE for AMD bdver3 architecture.
12516 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12518 * tree.c (warn_deprecated_use): Print file:line using locus color.
12519 * diagnostic.c (diagnostic_report_current_module): Print file:line
12520 and file:line:column using locus color.
12522 2013-05-14 Mike Stump <mikestump@comcast.net>
12524 * gdbinit.in: Add __null.
12526 2013-05-14 Mike Stump <mikestump@comcast.net>
12528 * recog.h: Rename struct recog_data to Recog_data.
12529 * recog.c: Likewise.
12530 * reload.c (can_reload_into): Likewise.
12531 * config/picochip/picochip.c: Likewise.
12533 2013-05-14 Mike Stump <mikestump@comcast.net>
12535 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
12537 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
12539 * resource.h (struct resources): Remove unch_memory member.
12540 (CLEAR_RESOURCE): Don't clear unch_memory.
12541 * resource.c (mark_referenced_resources): Don't set it.
12542 (mark_set_resources): Likewise.
12543 (mark_target_live_regs): Don't clear it.
12544 (init_resource_info): Likewise.
12545 * reorg.c (resource_conflicts_p): Don't compare it.
12546 (redundant_insn): Don't set it.
12548 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
12550 * emit-rtl.c (next_label): Remove unused function.
12551 (skip_consecutive_labels, link_cc0_insns): Move to ...
12552 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
12553 only place where these functions are used, and make them static.
12555 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
12557 * fold-const.c (fold_negate_expr): Handle vectors.
12558 (fold_truth_not_expr): Make it static.
12559 (fold_invert_truthvalue): New static function.
12560 (invert_truthvalue_loc): Handle vectors. Do not call
12561 fold_truth_not_expr directly.
12562 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
12563 <TRUTH_NOT_EXPR>: Do not cast to boolean.
12564 (fold_comparison): Handle vector constants.
12565 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
12566 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
12567 * tree.h (fold_truth_not_expr): Remove declaration.
12569 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
12571 * config/aarch64/aarch64-simd.md
12572 (aarch64_vcond_internal<mode>): Rename to...
12573 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
12574 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
12575 float modes. Clarify all iterator modes.
12576 (vcond<mode><mode>): Use new name for vcond expanders.
12577 (vcond<v_cmp_result><mode>): Likewise.
12578 (vcondu<mode><mode>: Likewise.
12579 * config/aarch64/iterators.md (VDQF_COND): New.
12581 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
12584 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
12585 variable for the shift amount. Check that we shift by non-negative
12588 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
12591 * config/arm/arm.h (EPILOGUE_USES): Only return true
12592 for LR_REGNUM after epilogue_completed.
12594 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
12596 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
12597 is error_mark_node.
12599 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12602 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
12603 and Solaris 11+/x86 with gld.
12604 * configure: Regenerate.
12606 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12608 * expmed.c (expand_shift_1): Canonicalize rotates by
12609 constant bitsize / 2 to bitsize - 1.
12610 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
12611 case ROTATERT>: Likewise.
12614 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12616 * config/i386/i386.md (rotateinv): New code attr.
12617 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12618 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12619 roll $31, %eax, etc.
12621 2013-05-14 Richard Biener <rguenther@suse.de>
12623 PR middle-end/57235
12624 * tree-eh.c (sink_clobbers): Give up for successors with
12625 multiple predecessors and no virtual uses.
12627 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
12629 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
12630 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
12632 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12634 PR middle-end/57251
12635 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
12636 the case when both op0 and op1 have VOIDmode.
12638 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12640 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
12641 in multiply-accumulate mode.
12643 2013-05-13 Guozhi Wei <carrot@google.com>
12645 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
12647 2013-05-13 Kai Tietz <ktietz@redhat.com>
12650 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
12651 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
12652 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
12653 * config/i386/i386.c (ix86_option_override_internal): Likewise.
12654 (ix86_expand_prologue): Likewise.
12655 (ix86_expand_split_stack_prologue): Likewise.
12656 (legitimate_pic_address_disp_p): Likewise.
12657 (legitimize_pic_address): Likewise.
12658 (legitimize_tls_address): Likewise.
12659 (legitimize_pe_coff_symbol): Likewise.
12660 (output_pic_addr_const): Likewise.
12661 (construct_plt_address): Likewise.
12662 (ix86_expand_call): Likewise.
12663 (x86_output_mi_thunk): Likewise.
12664 (x86_function_profiler): Likewise.
12666 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
12668 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
12669 similar switch cases.
12670 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
12671 (aarch64_simd_mov_to_<mode>low): Delete.
12672 (aarch64_simd_mov_to_<mode>high): Delete.
12673 (move_lo_quad_<mode>): Add w<-r alternative.
12674 (aarch64_simd_move_hi_quad_<mode>): Likewise.
12675 (aarch64_simd_mov_from_*): Update type attribute.
12676 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
12679 2013-05-13 Jan Hubicka <jh@suse.cz>
12681 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
12682 * config/i386/i386.c (ix86_compute_frame_layout,
12683 ix86_expand_epilogue, emit_i387_cw_initialization,
12684 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
12685 ix86_local_alignment): Fix use of size/speed predicates.
12687 2013-05-13 Jakub Jelinek <jakub@redhat.com>
12689 PR tree-optimization/45216
12690 PR tree-optimization/57157
12691 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
12692 the (-Y) & (B - 1) variant if OP is |.
12693 * expmed.c (expand_shift_1): For rotations by const0_rtx just
12694 return shifted. Use (-op1) & (prec - 1) as other_amount
12695 instead of prec - op1.
12697 2013-05-13 Martin Jambor <mjambor@suse.cz>
12699 PR middle-end/42371
12700 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
12701 (ipa_constant_data): New type.
12702 (ipa_jump_func): Use ipa_constant_data to hold information about
12703 constant jump functions.
12704 (ipa_get_jf_constant): Adjust to jump function type changes.
12705 (ipa_get_jf_constant_rdesc): New function.
12706 (ipa_param_descriptor): New field controlled_uses.
12707 (ipa_get_controlled_uses): New function.
12708 (ipa_set_controlled_uses): Likewise.
12709 * ipa-ref.h (ipa_find_reference): Declare.
12710 * ipa-prop.c (ipa_cst_ref_desc): New type.
12711 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
12713 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
12714 New parameter cs. Adjust all callers.
12715 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
12716 (remove_described_reference): New function.
12717 (jfunc_rdesc_usable): Likewise.
12718 (try_make_edge_direct_simple_call): Decrement controlled use count,
12719 attempt to remove reference if it hits zero.
12720 (combine_controlled_uses_counters): New function.
12721 (propagate_controlled_uses): Likewise.
12722 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
12723 (ipa_edge_duplication_hook): Duplicate reference descriptions.
12724 (ipa_print_node_params): Print described use counter.
12725 (ipa_write_jump_function): Adjust to jump function type changes.
12726 (ipa_read_jump_function): New parameter CS, pass it to
12727 ipa_set_jf_constant. Adjust caller.
12728 (ipa_write_node_info): Stream controlled use count
12729 (ipa_read_node_info): Likewise.
12730 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
12732 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
12733 count. Remove cloning-added reference if it reaches zero.
12734 * ipa-ref.c (ipa_find_reference): New function.
12736 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12738 * config/i386/i386.c (processor_target_table): Modified default
12739 alignment values for AMD BD and BT architectures.
12741 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
12743 * tree-vect-generic.c (uniform_vector_p): Move ...
12744 * tree.c (uniform_vector_p): ... here.
12745 * tree.h (uniform_vector_p): Declare it.
12746 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
12749 2013-05-13 Jakub Jelinek <jakub@redhat.com>
12751 PR tree-optimization/57230
12752 * tree-ssa-strlen.c (handle_char_store): Record length for
12753 array store from STRING_CST.
12755 PR tree-optimization/57230
12756 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
12759 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
12761 * config/epiphany/epiphany.c (epiphany_init): Check size of
12762 NUM_MODES_FOR_MODE_SWITCHING.
12763 (epiphany_expand_prologue):
12764 Remove CONFIG_REGNUM initial value handling code.
12765 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
12766 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
12767 (emit_set_fp_mode, epiphany_mode_after): Likewise.
12768 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
12769 Don't return 1 for FP_MODE_NONE.
12770 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
12771 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
12772 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
12773 * config/epiphany/epiphany.md (save_config): New pattern.
12775 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
12777 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
12779 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
12781 * config/i386/i386.md (memory): Handle sseishft1.
12782 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
12783 (*vec_extractv2di_1): Ditto.
12785 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
12787 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
12790 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
12792 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
12793 Add mthumb/march=armv7-a multilib.
12794 Add mthumb/march=armv7-r multilib.
12795 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
12797 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
12799 * config/v850/t-rtems: Add more multilibs.
12801 2013-05-10 Richard Biener <rguenther@suse.de>
12803 PR tree-optimization/57214
12804 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
12805 not propagate from SSA names that occur in abnormal PHI nodes.
12807 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
12809 * stor-layout.c (element_precision): New function.
12810 * machmode.h (element_precision): Declare it.
12811 * tree.c (build_minus_one_cst): New function.
12812 (element_precision): Likewise.
12813 * tree.h (build_minus_one_cst): Declare new function.
12814 (element_precision): Likewise.
12815 * fold-const.c (operand_equal_p): Use element_precision.
12816 (fold_binary_loc): Handle vector types.
12817 * convert.c (convert_to_integer): Use element_precision.
12818 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
12821 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
12823 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
12824 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
12825 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
12826 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12827 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
12828 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
12829 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
12830 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12831 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
12832 (Uuw8): New constraints.
12833 (Usb4): Move into alphabetical order.
12834 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
12835 (sd8_operand, ub8_operand, uw8_operand): New predicates.
12836 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
12837 previously unnamed patterns.
12838 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
12839 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
12840 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
12841 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
12842 of set_attr_alternative/if_then_else. Use extended_mips16 instead
12843 of specific lengths.
12845 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12847 * config/i386/i386.md (rotateinv): New code attr.
12848 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12849 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12850 roll $31, %eax, etc.
12852 PR tree-optimization/45216
12853 PR tree-optimization/57157
12854 * tree-ssa-forwprop.c (simplify_rotate): New function.
12855 (ssa_forward_propagate_and_combine): Call it.
12857 2013-05-10 Richard Biener <rguenther@suse.de>
12859 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
12860 disable peeling when we version for aliasing.
12861 (vector_alignment_reachable_p): Honor explicit user alignment.
12862 (vect_supportable_dr_alignment): Likewise.
12863 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
12864 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
12865 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
12866 then peeling to arrange for the cost-model check to come first.
12868 2013-05-10 Alan Modra <amodra@gmail.com>
12870 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
12871 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
12872 * configure: Regenerate.
12874 2013-05-10 Alan Modra <amodra@gmail.com>
12877 * varasm.c (default_elf_select_section): Move !DECL_P check..
12878 (get_named_section): ..to here before calling get_section_name.
12880 (default_section_type_flags): Add DECL_P check.
12881 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
12882 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
12884 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
12886 * config/epiphany/epiphany.c (epiphany_expand_prologue):
12887 When using gen_stack_adjust_str with a register offset, add a
12888 REG_FRAME_RELATED_EXPR note.
12890 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
12892 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
12893 (*vec_extractv4si_zext_mem): Ditto.
12894 (*vec_extractv2di): Add 0->x and x->x alternatives.
12895 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
12896 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
12898 2013-05-09 Jason Merrill <jason@redhat.com>
12900 N3639 C++1y VLA support
12901 * gimplify.c (gimplify_vla_decl): Don't touch an existing
12904 * tree.c (build_constructor_va): New.
12905 * tree.h: Declare it.
12907 2013-05-09 Martin Jambor <mjambor@suse.cz>
12910 * gimple-fold.c (canonicalize_constructor_val): Call
12911 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
12913 2013-05-09 Jan Hubicka <jh@suse.cz>
12914 Richard Biener <rguenther@suse.de>
12917 * symtab.c (symtab_make_decl_local): Do not add private names.
12919 2013-05-09 Jan Hubicka <jh@suse.cz>
12922 * symtab.c (insert_to_assembler_name_hash): Handle clones.
12923 (unlink_from_assembler_name_hash): Likewise.
12924 (symtab_prevail_in_asm_name_hash, symtab_register_node,
12925 symtab_unregister_node, symtab_initialize_asm_name_hash,
12926 change_decl_assembler_name): Update.
12928 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
12930 * config/aarch64/aarch64.md: New movtf split.
12931 (*movtf_aarch64): Update.
12932 (aarch64_movdi_tilow): Handle TF modes and rename to
12933 aarch64_movdi_<mode>low.
12934 (aarch64_movdi_tihigh): Handle TF modes and rename to
12935 aarch64_movdi_<mode>high
12936 (aarch64_movtihigh_di): Handle TF modes and rename to
12937 aarch64_mov<mode>high_di
12938 (aarch64_movtilow_di): Handle TF modes and rename to
12939 aarch64_mov<mode>low_di
12940 (aarch64_movtilow_tilow): Remove spurious whitespace.
12941 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
12943 (aarch64_print_operand): Update.
12945 2013-05-09 Alan Modra <amodra@gmail.com>
12947 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
12949 * configure: Regenerate.
12951 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12953 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
12954 splitter preparation statements.
12955 * config/i386/sse.md (*vec_extract* splitters): Ditto.
12956 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
12959 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12961 * gimple-ssa-strength-reduction.c (count_candidates): Change
12962 return value to int.
12963 (analyze_candidates_and_replace): Change type of length to int.
12965 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12967 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
12968 (*vec_extract<mode>): Use VI12_128 mode iterator.
12969 (*vec_extract<mode>_mem): Ditto.
12970 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
12973 2013-05-08 Diego Novillo <dnovillo@google.com>
12978 2012-08-17 Diego Novillo <dnovillo@google.com>
12981 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
12982 * config.in: Regenerate.
12983 * configure: Regenerate.
12984 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
12986 2013-05-08 Jan Hubicka <jh@suse.cz>
12989 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
12990 * cgraph.h (symtab_node_base): Add unique_name.
12991 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12992 input_overwrite_node, input_varpool_node): Stream unique_name.
12993 * cgraphclones.c (cgraph_create_virtual_clone,
12994 cgraph_function_versioning): Set unique_name.
12995 * ipa.c (function_and_variable_visibility): Set unique_name.
12997 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12999 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13000 (alloc_cand_and_find_basis): Restrict conditional candidate
13001 processing to CAND_MULTs.
13003 2013-05-08 Jan Hubicka <jh@suse.cz>
13006 lto-symtab.c (lto_symtab_symbol_p): New function.
13007 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13008 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13009 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13010 Skip static symbols.
13012 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13014 PR tree-optimization/57200
13015 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13016 Only call inform if the preceding warning_at returns true.
13018 2013-05-07 Han Shen <shenhan@google.com>
13020 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13021 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13022 * common.opt (fstack-protector-strong): New option.
13023 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13024 * doc/invoke.texi (Optimization Options): Document
13025 "-fstack-protector-strong".
13026 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13028 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13030 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13032 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13034 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13035 it is not lhs of the insn.
13036 (setup_reg_equiv): Remove insn having equiv memory which it is not
13038 * lra-constraints.c (process_address): Try to improve generation
13039 code for address base + disp.
13040 (lra_constraints): Make correct the code for checking insn setting
13041 up backward equivalence. Remove insn only if it is in the init
13043 * lra-eliminations.c (update_reg_eliminate): Change return value.
13044 (lra_eliminate): Use the result.
13046 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13048 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13049 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13050 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13051 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13052 register target operands.
13053 (*vec_extractv8hi_sse2): New pattern.
13054 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
13055 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
13056 (*vec_extract<mode>_mem): New insn and split pattern.
13058 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
13060 * config/arm/arm.c (arm_asan_shadow_offset): New function.
13061 (TARGET_ASAN_SHADOW_OFFSET): Define.
13062 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
13063 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
13065 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13067 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
13068 (incr_vec_index): Return -1 if increment not found.
13069 (create_add_on_incoming_edge): Assert if increment not found.
13070 (record_increment): Limit number of increments recorded.
13071 (all_phi_incrs_profitable): Return false if an increment not found.
13072 (replace_profitable_candidates): Don't process increments that were
13074 (analyze_candidates_and_replace): Limit size of incr_vec.
13076 2013-05-07 Richard Biener <rguenther@suse.de>
13078 * calls.c (special_function_p): setjmp-like functions are leaf.
13079 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
13080 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
13082 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
13084 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
13085 (aarch64_simd_mov<mode>): New expander.
13086 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
13087 (aarch64_simd_mov_to_<mode>high): Likewise.
13088 (aarch64_simd_mov_from_<mode>low): Likewise.
13089 (aarch64_simd_mov_from_<mode>high): Likewise.
13090 (aarch64_dup_lane<mode>): Update.
13091 (aarch64_dup_lanedi): New instruction pattern.
13092 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
13093 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
13095 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13097 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
13098 (replace_mult_candidate): Remove unnecessary argument; remove
13099 unnecessary parameter from call to introduce_cast_before_cand.
13100 (replace_unconditional_candidate): Remove unnecessary parameter
13101 from call to replace_mult_candidate.
13102 (replace_conditional_candidate): Likewise.
13103 (insert_initializers): Use make_temp_ssa_name.
13104 (introduce_cast_before_cand): Remove unnecessary argument; use
13105 make_temp_ssa_name.
13106 (replace_one_candidate): Remove unnecessary argument; remove
13107 unnecessary parameter from calls to introduce_cast_before_cand.
13108 (replace_profitable_candidates): Remove unnecessary parameters
13109 from calls to replace_one_candidate.
13111 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13113 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
13114 phi def as possibly hiding a basis for a CAND_ADD whose operands
13115 have been commuted in the analysis.
13116 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
13118 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13120 * config/aarch64/aarch64.md
13121 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
13122 shift value between 0-4.
13124 2013-05-07 Richard Biener <rguenther@suse.de>
13126 * double-int.h (rshift): New overload.
13127 * double-int.c (rshift): New function.
13128 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
13129 (create_reference_ops_from_ref): Remove.
13130 (vn_reference_insert): Use shared ops for constructing the
13131 reference and copy it.
13133 2013-05-07 Richard Biener <rguenther@suse.de>
13135 PR middle-end/57190
13136 * tree-eh.c (sink_clobbers): Properly propagate
13137 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
13139 2013-05-07 Jakub Jelinek <jakub@redhat.com>
13141 PR tree-optimization/57149
13142 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
13143 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
13144 collect_phi_def_edges, execute_late_warn_uninitialized): Use
13145 uninit_undefined_value_p instead of ssa_undefined_value_p.
13148 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
13149 for modifier == EXPAND_INITIALIZER.
13151 2013-05-07 Anton Blanchard <anton@samba.org>
13153 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
13154 for powerpc64 little endian.
13155 * configure: Regenerate.
13157 2013-05-06 Graham Stott <grahams@btinternet.com>
13159 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
13160 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
13161 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
13162 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
13164 2013-05-06 Graham Stott <grahams@btinternet.com>
13166 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
13167 codes which allow non-lvalues.
13169 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
13171 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
13172 components are all 1s.
13173 (integer_minus_onep): New function.
13174 * tree.h (integer_minus_onep): Declare it.
13175 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
13176 integer_minus_onep instead of integer_all_onesp.
13178 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13181 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
13182 variations of these patterns.
13184 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13186 * config/i386/i386.md (isa): Add x64_sse4 member.
13187 (enabled): Handle x64_sse4.
13188 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
13189 instruction for 64bit SSE4_1 targets. Update insn attributes.
13190 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
13191 instruction for SSE4_1 targets. Update insn attributes.
13192 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13193 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
13194 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
13196 (*vec_extractv4si): Rename from *sse4_1_pextrd.
13197 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
13198 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
13200 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13203 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
13205 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
13207 * final.c (do_assembler_dialects): Don't handle curly braces and
13208 vertical bar escaped by % as dialect delimiters.
13209 (output_asm_insn): Print curly braces and vertical bar if escaped
13210 by % and ASSEMBLER_DIALECT defined.
13211 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
13212 * doc/tm.texi: Regenerated.
13214 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13216 * config/mips/mips.c: Include tree-pass.h.
13217 (mips_reorg): Split in pre- and post-dbr_schedule parts.
13218 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
13219 (pass_mips_machine_reorg2): New machine specific pass.
13220 (insert_pass_mips_machine_reorg2): New pass plugin definition.
13221 (mips_option_override): Register the new pass.
13222 * rtl.h (cleanup_barriers): Remove prototype.
13223 (dbr_schedule): Likewise.
13224 * jump.c (cleanup_barriers): Make static.
13225 * reorg.c (dbr_schedule): Likewise.
13227 2013-05-06 Richard Biener <rguenther@suse.de>
13229 PR tree-optimization/57185
13230 * tree-parloops.c (add_field_for_reduction): Handle anonymous
13231 SSA names properly.
13233 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13236 * config/i386/i386.c (add_parameter_dependencies): Add dependence
13237 between "first_arg" and "insn", not "last" and "insn".
13239 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13241 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
13242 (find_candidates_in_block): Re-enable slsr_process_phi.
13243 (create_phi_basis): Fix double counting of candidate adjustment.
13245 2013-05-06 Richard Biener <rguenther@suse.de>
13247 PR middle-end/57147
13248 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
13249 the edge is also fallthru, preserve it and just clear the
13251 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
13252 also complex, preserve that and just clear the fallthru flag.
13253 * tree-inline.c (update_ssa_across_abnormal_edges): Also
13254 update virtual operands.
13256 2013-05-06 Alan Modra <amodra@gmail.com>
13258 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
13259 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
13260 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
13261 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
13262 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
13263 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
13265 2013-05-06 Alan Modra <amodra@gmail.com>
13267 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
13268 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
13269 (DEFAULT_ASM_ENDIAN): Define.
13270 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
13271 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
13272 Update -K PIC clause from sysv4.h.
13273 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
13274 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
13276 2013-05-06 Alan Modra <amodra@gmail.com>
13278 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
13279 twice for little-endian.
13280 (ashrdi3_no_power, ashrdi3): Support little-endian.
13282 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13285 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
13286 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
13288 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
13289 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
13290 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
13293 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
13294 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13296 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
13297 * config/pa/pa.opt: Make mbig-switch a no-op.
13298 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
13299 (CASE_VECTOR_MODE): Always return SImode.
13300 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
13301 for the !TARGET_BIG_SWITCH case.
13302 * config/pa/pa-linux.h: Likewise.
13303 * config/pa/pa-openbsd.h: Likewise.
13304 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
13305 * config/pa/pa.md (short_jump): Remove define_insn.
13306 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
13307 (casesi0): Remove define_insn.
13308 (type): Remove btable_branch.
13309 (pa_combine_type): Likewise.
13310 (in_nullified_branch_delay): Likewise.
13311 (in_call_delay): Likewise.
13312 (define_delay): Likewise.
13313 (define_insn_reservation "Z3"): Likewise.
13314 (define_insn_reservation "Z4"): Likewise.
13315 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
13316 (pa_adjust_insn_length): Remove adjustment for btable branches.
13317 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
13320 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
13322 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13323 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
13324 Add m->r,x alternatives.
13325 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
13326 splitters using SWI48x mode iterator.
13327 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
13328 TARGET_64BIT. Add m->x alternative.
13329 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
13330 Add o->x alternative. Enable for TARGET_SSE.
13331 (sse_storeq): Remove expander.
13332 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
13333 with memory input operand.
13334 (*vec_extractv2di_1 splitter): New.
13335 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
13336 * config/i386/i386.md (ssevecmodelower): New mode attribute.
13338 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13340 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
13341 (INT_LOWPART): Delete.
13342 (extract_MB): Adjust.
13343 (extract_ME): Adjust.
13344 (print_operand): Adjust.
13346 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13348 * config/rs6000/predicates.md (reg_or_add_cint_operand,
13349 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13350 (reg_or_logical_cint_operand, easy_fp_constant,
13351 logical_const_operand): Delete "CONST_DOUBLE" case.
13352 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
13353 "HOST_BITS_PER_WIDE_INT == 64" test.
13354 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
13355 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
13356 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
13357 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
13359 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
13360 CONST_DOUBLE DImode/VOIDmode case.
13361 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
13362 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
13363 CONST_DOUBLE VOIDmode case.
13364 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13365 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
13366 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
13367 Delete CONST_DOUBLE case.
13368 (splitters for mov FMOVE64 const_double): Delete
13369 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
13370 "HOST_BITS_PER_WIDE_INT >= 64" test.
13371 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
13373 (mov DI const_double): Delete.
13375 2013-05-04 Jakub Jelinek <jakub@redhat.com>
13377 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
13378 on op shows all bits zero in mode of a lowpart subreg, return zero.
13380 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
13383 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
13384 to save TFmode registers and DImode to save TImode registers for
13385 caller save operations.
13386 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
13387 mark being partially clobbered since they only use the first
13390 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
13391 and TDmode only use the upper 64-bits of each VSX register.
13393 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13395 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
13396 (find_candidates_in_block): Disable slsr_process_phi.
13398 2013-05-03 Guozhi Wei <carrot@google.com>
13400 * coverage.c (coverage_obj_init): Move the construction of gcov
13402 (build_init_ctor): ... here.
13404 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13406 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
13407 (slsr_cand_d): Redefine def_phi.
13408 (stride_status, phi_adjust_status, count_phis_status): New enums.
13409 (find_phi_def): New.
13410 (find_basis_for_base_expr): New.
13411 (find_basis_for_candidate): Handle hidden bases.
13412 (alloc_cand_and_find_basis): Handle phi candidates.
13413 (slsr_process_phi): New.
13414 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
13415 (create_mul_imm_cand): Likewise.
13416 (create_add_ssa_cand): Exclude phi base candidates.
13417 (create_add_imm_cand): Likewise.
13418 (slsr_process_cast): Likewise.
13419 (slsr_process_copy): Likewise.
13420 (find_candidates_in_block): Handle phi candidates.
13421 (dump_candidate): Likewise.
13422 (unconditional_cands): Delete.
13423 (unconditional_cands_with_known_stride_p): Delete.
13424 (phi_dependent_cand_p): New.
13425 (cand_increment): Handle phi-dependent candidates.
13426 (replace_dependent): Delete.
13427 (replace_mult_candidate): New.
13428 (replace_unconditional_candidate): New.
13429 (incr_vec_index): Move to avoid forward reference.
13430 (create_add_on_incoming_edge): New.
13431 (create_phi_basis): New.
13432 (replace_dependents): Delete.
13433 (replace_conditional_candidate): New.
13434 (phi_add_costs): New.
13435 (replace_uncond_cands_and_profitable_phis): New.
13436 (record_increment): Handle phi adjustments.
13437 (record_phi_increments): New.
13438 (record_increments): Handle phi adjustments.
13439 (phi_incr_cost): New.
13440 (lowest_cost_path): Handle phis.
13441 (total_savings): Likewise.
13442 (analyze_increments): Likewise.
13443 (ncd_with_phi): New.
13444 (ncd_of_cand_and_phis): New.
13445 (nearest_common_dominator_for_cands): Handle phi increments.
13446 (all_phi_incrs_profitable): New.
13447 (replace_profitable_candidates): Handle phi-dependent candidates.
13448 (analyze_candidates_and_replace): Likewise.
13450 2013-05-03 Teresa Johnson <tejohnson@google.com>
13453 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
13454 do not exceed REG_BR_PROB_BASE.
13456 2013-05-03 Jeff Law <law@redhat.com>
13458 PR tree-optimization/57144
13459 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
13460 operand of the condition will bit into the new type when eliminating
13461 a cast feeding a condition.
13463 2013-05-03 Jakub Jelinek <jakub@redhat.com>
13465 PR rtl-optimization/57130
13466 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
13467 of COMPARE as in_code to the recursive call if needed.
13469 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
13471 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
13472 (enabled): Handle new members.
13473 * config/i386/sse.md (*vec_concatv2si): Merge from
13474 *vec_concatv2si_sse2 and vec_concatv2si_sse.
13475 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
13477 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
13479 PR tree-optimization/57027
13480 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
13481 for fnms opportunity, check we got the prerequisite kind
13482 of tree / gimple before using accessor functions.
13484 2013-05-03 Richard Biener <rguenther@suse.de>
13486 * double-int.h (lshift): New overload without precision
13487 and arith argument.
13488 (operator *=, operator +=, operator -=): Move ...
13489 * double-int.c (operator *=, operator +=, operator -=): ... here
13490 and implement more efficiently.
13491 (mul_double_with_sign): Remove.
13492 (lshift_double): Adjust to take unsinged shift argument, push
13493 dispatching code to callers.
13494 (mul_double_wide_with_sign): Add early out for callers that
13495 are not interested in high parts or overflow.
13496 (lshift): New function.
13497 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
13498 dispatch code here.
13499 (lrotate, rrotate): Use logical shifts.
13500 * expr.c (get_inner_reference): Use lshift.
13501 * fixed-value.c (do_fixed_divide): Likewise.
13502 * tree-dfa.c (get_ref_base_and_extent): Likewise.
13503 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
13504 (indirect_refs_may_alias_p): Likewise.
13505 (stmt_kills_ref_p_1): Likewise.
13507 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
13509 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
13511 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
13513 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
13514 scalar form of FABD instruction.
13516 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
13518 * lra-constraints.c (process_alt_operands): Add checking alt
13519 number to choose the best alternative.
13521 2013-05-02 Richard Biener <rguenther@suse.de>
13523 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
13524 bitmap and its handling.
13525 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
13527 2013-05-02 Richard Biener <rguenther@suse.de>
13529 PR middle-end/57140
13530 * tree-inline.c (copy_loops): Properly handle removed loops.
13531 (copy_cfg_body): Mark destination loops for fixup if source
13532 loops needed fixup.
13534 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
13537 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
13538 generating simple_return for naked functions.
13540 2013-05-02 Martin Jambor <mjambor@suse.cz>
13542 PR middle-end/56988
13543 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
13544 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
13546 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
13547 ipa_agg_replacement_value structures.
13548 (known_aggs_to_agg_replacement_list): Likewise.
13549 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
13550 (read_agg_replacement_chain): Likewise.
13551 (ipcp_transform_function): Also check that by_ref flags match.
13553 2013-05-02 Richard Biener <rguenther@suse.de>
13555 * graphds.h (struct graph): Add obstack member.
13556 * graphds.c (new_graph): Initialize obstack and allocate
13558 (add_edge): Allocate edge from the obstack.
13559 (free_graph): Free the obstack instead of all edges and vertices.
13561 2013-05-02 Teresa Johnson <tejohnson@google.com>
13563 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
13565 * cfg.c (update_bb_profile_for_threading): Ditto.
13566 * tree-inline.c (copy_bb): Ditto.
13567 (copy_edges_for_bb): Ditto.
13568 (initialize_cfun): Ditto.
13569 (copy_cfg_body): Ditto.
13570 (expand_call_inline): Ditto.
13571 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
13572 (estimate_node_size_and_time): Ditto.
13573 (inline_merge_summary): Ditto.
13574 * cgraphclones.c (cgraph_clone_edge): Ditto.
13575 (cgraph_clone_node): Ditto.
13576 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13577 (compute_trg_info): Ditto.
13579 2013-05-02 Ian Bolton <ian.bolton@arm.com>
13581 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
13582 S reg when fp attribute set.
13583 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
13585 2013-05-02 Ian Bolton <ian.bolton@arm.com>
13587 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
13589 (*and_one_cmplsi3_compare0_uxtw): Likewise.
13590 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
13591 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
13593 2013-05-02 Richard Biener <rguenther@suse.de>
13595 * tree-scalar-evolution.c (scev_info_hasher): Remove.
13596 (struct instantiate_cache_entry): New type.
13597 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
13598 (struct instantiate_cache_type): New type.
13599 (set_instantiated_value, get_instantiated_value): Remove.
13600 (get_instantiated_value_entry): New function.
13601 (instantiate_scev_name): Use the new cache and adjust.
13602 (instantiate_scev_poly): Adjust.
13603 (instantiate_scev_binary): Likewise.
13604 (instantiate_array_ref): Likewise.
13605 (instantiate_scev_convert): Likewise.
13606 (instantiate_scev_not): Likewise.
13607 (instantiate_scev_3): Likewise.
13608 (instantiate_scev_2): Likewise.
13609 (instantiate_scev_r): Likewise.
13610 (instantiate_scev): Likewise.
13611 (resolve_mixers): Likewise.
13613 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
13616 * lra-constraints.c (best_small_class_operands_num): Remove.
13617 (process_alt_operands): Remove small_class_operands_num. Take
13618 small classes operands into losers and only if the operand is not
13619 matched. Modify debugging output.
13620 (curr_insn_transform): Remove best_small_class_operands_num.
13623 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13625 * config/aarch64/aarch64-builtins.c
13626 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
13627 * config/aarch64/aarch64-simd-builtins.def
13628 (reduc_splus_): Add new modes.
13629 (reduc_uplus_): New.
13630 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
13631 (reduc_uplus_v4sf): Likewise.
13632 (reduc_splus_v4sf): Likewise.
13633 (aarch64_addv<mode>): Likewise.
13634 (reduc_uplus_<mode>): Likewise.
13635 (reduc_splus_<mode>): Likewise.
13636 (aarch64_addvv2di): Likewise.
13637 (reduc_uplus_v2di): Likewise.
13638 (reduc_splus_v2di): Likewise.
13639 (aarch64_addvv2si): Likewise.
13640 (reduc_uplus_v2si): Likewise.
13641 (reduc_splus_v2si): Likewise.
13642 (reduc_<sur>plus_<mode>): New.
13643 (reduc_<sur>plus_v2di): Likewise.
13644 (reduc_<sur>plus_v2si): Likewise.
13645 (reduc_<sur>plus_v4sf): Likewise.
13646 (aarch64_addpv4sf): Likewise.
13647 * config/aarch64/arm_neon.h
13648 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
13649 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
13650 add UNSPEC_SADDV, UNSPEC_UADDV.
13652 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
13654 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13656 * config/aarch64/arm_neon.h
13657 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
13659 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13661 * config/aarch64/aarch64-builtins
13662 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
13664 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13666 * config/aarch64/aarch64-simd-builtins.def
13667 (reduc_smax_): New.
13668 (reduc_smin_): Likewise.
13669 (reduc_umax_): Likewise.
13670 (reduc_umin_): Likewise.
13671 (reduc_smax_nan_): Likewise.
13672 (reduc_smin_nan_): Likewise.
13675 (smax): Update for V2SF, V4SF and V2DF modes.
13678 (smin_nan): Likewise.
13679 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
13680 (<su><maxmin><mode>3): ...This, refactor.
13681 (s<maxmin><mode>3): New.
13682 (<maxmin_uns><mode>3): Likewise.
13683 (reduc_<maxmin_uns>_<mode>): Refactor.
13684 (reduc_<maxmin_uns>_v4sf): Likewise.
13685 (reduc_<maxmin_uns>_v2si): Likewise.
13686 (aarch64_<fmaxmin><mode>: Remove.
13687 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
13689 (vmin<q>_f<32,64>): Likewise.
13690 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
13692 (su): Add mappings for smax, smin, umax, umin.
13694 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
13695 (FMAXMIN): Rename as...
13696 (FMAXMIN_UNS): ...This.
13698 (fmaxminv): Likewise.
13699 (fmaxmin): Likewise.
13701 (maxmin_uns_op): Likewise.
13703 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13705 * config/aarch64/arm_neon.h
13706 (vac<ge, gt><sd>_f<32, 64>): Rename to...
13707 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
13708 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
13710 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13712 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
13713 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
13715 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13717 * config/aarch64/aarch64-simd.md
13718 (vcond<mode>_internal): Handle special cases for constant masks.
13719 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
13720 (vcondu<mode><mode>): Likewise.
13721 (vcond<v_cmp_result><mode>): New.
13723 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13725 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
13726 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
13727 * config/aarch64/aarch64-simd-builtins.def
13728 (cmeq): Update to BUILTIN_VALLDI.
13733 * config/aarch64/arm_neon.h
13734 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
13735 to builtins or C as appropriate.
13737 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13739 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
13741 (cmhi): Rename to...
13743 * config/aarch64/aarch64-simd.md
13744 (simd_mode): Add SF.
13745 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
13746 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
13747 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
13748 (cstore<mode>_neg): ...This.
13749 * config/aarch64/iterators.md
13751 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
13752 (COMPARISONS): New.
13753 (UCOMPARISONS): Likewise.
13754 (optab): Add missing comparisons.
13760 (VCMP_S): Likewise.
13761 (VCMP_U): Likewise.
13762 (V_cmp_result): Add DF, SF modes.
13763 (v_cmp_result): Likewise.
13765 (vmtype): Likewise.
13766 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
13768 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
13770 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
13771 define_insn to define_insn_and_split.
13772 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
13773 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
13774 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
13775 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
13776 (thumb2_negscc): Likewise.
13778 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13780 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
13782 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13784 * config/arm/thumb2.md: Remove trailing whitespaces.
13786 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13788 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
13789 Use gen_int_mode rather than GEN_INT.
13791 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
13793 * value-prof.c (stream_in_histogram_value): Remove the strayed
13796 2013-04-30 Richard Biener <rguenther@suse.de>
13798 PR middle-end/57122
13799 * cfghooks.c (split_edge): Properly check for the loop latch edge.
13801 2013-04-30 Richard Biener <rguenther@suse.de>
13803 PR middle-end/57107
13804 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
13806 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
13808 PR rtl-optimization/56957
13809 PR rtl-optimization/57105
13810 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
13811 variable. Use just INSN_UID for determining whether an insn
13812 should be only disconnected from the insn stream.
13813 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
13815 2013-04-30 Jakub Jelinek <jakub@redhat.com>
13817 PR tree-optimization/57104
13818 * tsan.c (instrument_expr): Don't instrument accesses to
13819 DECL_HARD_REGISTER VAR_DECLs.
13821 2013-04-30 Richard Biener <rguenther@suse.de>
13823 * function.h (loops_for_fn): New inline function.
13824 (set_loops_for_fn): Likewise.
13825 * cfgloop.h (place_new_loop): Add struct function parameter.
13826 (get_loop): Likewise.
13827 (get_loops): Likewise.
13828 (number_of_loops): Likewise.
13829 (fel_next): Adjust.
13830 (fel_init): Likewise.
13831 * cfg.c (get_loop_copy): Adjust.
13832 * cfgloop.c (flow_loops_dump): Likewise.
13833 (record_loop_exits): Likewise.
13834 (verify_loop_structure): Likewise.
13835 * cfgloopanal.c (mark_irreducible_loops): Likewise.
13836 (estimate_reg_pressure_cost): Likewise.
13837 (mark_loop_exit_edges): Likewise.
13838 * cfgloopmanip.c (place_new_loop): Likewise.
13839 (add_loop): Likewise.
13840 (duplicate_loop): Likewise.
13841 * graph.c (draw_cfg_nodes): Likewise.
13842 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
13843 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
13844 (extract_affine_chrec): Likewise.
13845 (build_scop_iteration_domain): Likewise.
13846 * graphite.c (graphite_initialize): Likewise.
13847 * ira-build.c (create_loop_tree_nodes): Likewise.
13848 (more_one_region_p): Likewise.
13849 (rebuild_regno_allocno_maps): Likewise.
13850 (mark_loops_for_removal): Likewise.
13851 (mark_all_loops_for_removal): Likewise.
13852 (remove_unnecessary_regions): Likewise.
13853 (ira_build): Likewise.
13854 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
13855 * loop-init.c (fix_loop_structure): Likewise.
13856 (gate_rtl_move_loop_invariants): Likewise.
13857 (gate_rtl_unswitch): Likewise.
13858 (gate_rtl_unroll_and_peel_loops): Likewise.
13859 (rtl_doloop): Likewise.
13860 * lto-streamer-in.c (input_cfg): Likewise.
13861 * lto-streamer-out.c (output_cfg): Likewise.
13862 * modulo-sched.c (sms_schedule): Likewise.
13863 * predict.c (tree_estimate_probability): Likewise.
13864 (tree_estimate_probability_driver): Likewise.
13865 (estimate_loops): Likewise.
13866 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
13867 (move_sese_region_to_fn): Likewise.
13868 (debug_loop_num): Likewise.
13869 * tree-chrec.c (chrec_evaluate): Likewise.
13870 (hide_evolution_in_other_loops_than_loop): Likewise.
13871 (chrec_component_in_loop_num): Likewise.
13872 (reset_evolution_in_loop): Likewise.
13873 (evolution_function_is_invariant_rec_p): Likewise.
13874 * tree-if-conv.c (main_tree_if_conversion): Likewise.
13875 * tree-inline.c (copy_loops): Likewise.
13876 (copy_cfg_body): Likewise.
13877 (tree_function_versioning): Likewise.
13878 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
13879 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
13881 (add_to_evolution_1): Likewise.
13882 (scev_const_prop): Likewise.
13883 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
13884 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
13885 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
13886 (tree_ssa_lim_initialize): Likewise.
13887 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
13888 (verify_loop_closed_ssa): Likewise.
13889 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
13890 (tree_ssa_loop_im): Likewise.
13891 (tree_ssa_loop_unswitch): Likewise.
13892 (tree_vectorize): Likewise.
13893 (check_data_deps): Likewise.
13894 (tree_ssa_loop_ivcanon): Likewise.
13895 (tree_ssa_loop_bounds): Likewise.
13896 (tree_complete_unroll): Likewise.
13897 (tree_complete_unroll_inner): Likewise.
13898 (tree_parallelize_loops): Likewise.
13899 (tree_ssa_loop_prefetch): Likewise.
13900 (tree_ssa_loop_ivopts): Likewise.
13901 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
13902 * tree-vectorizer.c (vectorize_loops): Likewise.
13904 2013-04-29 Mike Frysinger <vapier@gentoo.org>
13906 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
13907 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
13908 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
13909 with EABI_LINK_SPEC.
13911 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13914 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
13917 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
13920 * lra-constraints.c (process_alt_operands): Discourage a bit more
13921 using memory for pseudos. Print cost dump for alternatives.
13922 Modify cost values for conflicts with early clobbers.
13923 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
13925 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13928 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
13930 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13932 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
13933 from/to S register.
13934 (movdi_aarch64): Support LDR/STR from/to D register.
13936 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13938 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
13939 or higher by default.
13941 2013-04-29 Richard Biener <rguenther@suse.de>
13943 PR middle-end/57075
13944 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
13945 even if not adding abnormal edges for calls that can make
13948 2013-04-29 Richard Biener <rguenther@suse.de>
13950 PR middle-end/57103
13951 * tree-cfg.c (move_stmt_op): Fix condition under which to update
13953 (move_stmt_r): Remove redundant checking.
13955 2013-04-29 Teresa Johnson <tejohnson@google.com>
13958 * basic-block.h (apply_scale): New function.
13959 (apply_probability): Use apply_scale.
13960 * gimple-streamer-in.c (input_bb): Ditto.
13961 * lto-streamer-in.c (input_cfg): Ditto.
13962 * lto-cgraph.c (merge_profile_summaries): Ditto.
13963 * tree-optimize.c (execute_fixup_cfg): Ditto.
13964 * tree-inline.c (copy_bb): Update comment to use apply_scale.
13965 (copy_edges_for_bb): Ditto.
13966 (copy_cfg_body): Ditto.
13968 2013-04-29 Tom de Vries <tom@codesourcery.com>
13970 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
13971 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
13972 (tail_merge_optimize): Handle current_loops == NULL.
13974 2013-04-26 Jeff Law <law@redhat.com>
13976 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
13977 (simplify_cond_using_ranges): Generalize code to simplify
13978 COND_EXPRs where one argument is a constant and the other
13979 is an SSA_NAME created by an integral type conversion.
13981 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13983 * config/arm/arm.md (store_minmaxsi): Use only when
13984 optimize_insn_for_size_p.
13986 2013-04-29 Christian Bruel <christian.bruel@st.com>
13989 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
13991 2013-04-29 Richard Biener <rguenther@suse.de>
13993 PR middle-end/57089
13994 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
13995 loop tree make sure to schedule a fixup for the child as well.
13996 (expand_omp_for_generic): Properly add loops.
13997 (expand_omp_for_static_nochunk): Likewise.
13998 (expand_omp_for_static_chunk): Likewise.
13999 (expand_omp_for): For the degenerate case fixup loops.
14000 (expand_omp_sections): Fix default bb placement in loops.
14001 (expand_omp_atomic_pipeline): Properly add loops.
14003 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14005 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14007 2013-04-29 Tom de Vries <tom@codesourcery.com>
14009 * tree-ssa-tail-merge.c: Update header comment.
14011 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14013 * config/aarch64/arm_neon.h
14014 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14015 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14016 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14017 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14018 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14019 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14020 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14021 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14023 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14025 * config/aarch64/aarch64-simd.md
14026 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14027 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14028 fix_trunc, fixuns_trunc.
14029 (ftrunc<VDQF:mode>2): New.
14030 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14031 (fix_trunc_optab): New.
14033 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14035 * config/aarch64/aarch64-builtins.c
14036 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14037 iceilf, lround, iroundf.
14039 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14042 * config/i386/i386.h (enum ix86_tune_indices)
14043 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14044 New, split from X86_TUNE_INTER_UNIT_MOVES.
14045 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14046 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14047 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14048 (TARGET_INTER_UNIT_MOVES): Remove.
14049 * config/i386/i386.c (initial_ix86_tune_features): Update.
14050 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14051 (ix86_expand_convert_uns_didf_sse): Use
14052 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14053 (ix86_expand_vector_init_one_nonzero): Ditto.
14054 (ix86_expand_vector_init_interleave): Ditto.
14055 (inline_secondary_memory_needed): Return true for moves from SSE class
14056 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
14057 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
14058 * config/i386/constraints.md (Yi, Ym): Depend on
14059 TARGET_INTER_UNIT_MOVES_TO_VEC.
14060 (Yj, Yn): New constraints.
14061 * config/i386/i386.md (*movdi_internal): Change constraints of
14062 operand 1 from Yi to Yj and from Ym to Yn.
14063 (*movsi_internal): Ditto.
14064 (*movdf_internal): Ditto.
14065 (*movsf_internal): Ditto.
14066 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
14067 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14068 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
14069 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
14070 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
14071 * config/i386/sse.md (movdi_to_sse): Ditto.
14072 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
14073 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
14074 TARGET_INTER_UNIT_MOVES.
14075 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
14076 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
14077 instead of TARGET_INTER_UNIT_MOVES.
14078 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
14079 operand 1 from Yi to Yj and from Ym to Yn.
14081 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14083 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
14084 (float_truncate_hi_): Likewise.
14085 (float_extend_lo_): Likewise.
14086 (float_truncate_lo_): Likewise.
14087 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
14088 (aarch64_float_extend_lo_v2df): Likewise.
14089 (vec_unpacks_hi_v4sf): Likewise.
14090 (aarch64_float_truncate_lo_v2sf): Likewise.
14091 (aarch64_float_truncate_hi_v4sf): Likewise.
14092 (vec_pack_trunc_v2df): Likewise.
14093 (vec_pack_trunc_df): Likewise.
14095 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14097 * config/aarch64/aarch64-builtins.c
14098 (aarch64_fold_builtin): Fold float conversions.
14099 * config/aarch64/aarch64-simd-builtins.def
14100 (floatv2si, floatv4si, floatv2di): New.
14101 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
14102 * config/aarch64/aarch64-simd.md
14103 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
14104 * config/aarch64/iterators.md (FLOATUORS): New.
14105 (optab): Add float, floatuns.
14106 (su_optab): Likewise.
14108 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14110 * config/aarch64/aarch64-builtins.c
14111 (aarch64_builtin_vectorized_function): Use new names for
14113 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
14114 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
14115 (fcvtzu): Split as...
14116 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
14117 (fcvtas): Split as...
14118 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
14119 (fcvtau): Split as...
14120 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
14121 (fcvtps): Split as...
14122 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
14123 (fcvtpu): Split as...
14124 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
14125 (fcvtms): Split as...
14126 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
14127 (fcvtmu): Split as...
14128 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
14129 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
14130 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
14131 (lfrintnusf, lfrintnudf): Likewise.
14132 * config/aarch64/aarch64-simd.md
14133 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
14135 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
14136 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
14137 (fcvt_pattern): Likewise.
14139 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14141 * config/aarch64/aarch64-simd.md
14142 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
14143 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
14145 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14147 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
14148 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
14149 (vrnd<a,m,n,p>_f32): Implement using builtins.
14150 (vrnd<i,x><q>_f<32, 64>): New.
14152 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14154 * config/aarch64/aarch64-builtins.c
14155 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
14156 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
14157 (frintz): Rename to...
14159 (frintp): Rename to...
14161 (frintm): Rename to...
14163 (frinti): Rename to...
14164 (nearbyint): ...this.
14165 (frintx): Rename to...
14167 (frinta): Rename to...
14169 * config/aarch64/aarch64-simd.md
14170 (aarch64_frint<frint_suffix><mode>): Delete.
14171 (<frint_pattern><mode>2): Convert to insn.
14172 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
14173 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
14174 (frint_pattern): Likewise.
14175 (frint_suffix): Likewise.
14177 2013-04-29 Richard Biener <rguenther@suse.de>
14179 PR tree-optimization/57081
14180 * loop-init.c: Include tree-flow.h.
14181 (loop_optimizer_finalize): Free number of iteration estimates.
14182 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
14184 2013-04-29 Jakub Jelinek <jakub@redhat.com>
14186 PR tree-optimization/57083
14187 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
14188 non-singleton shift count range, zero extend low_bound for uns case.
14190 * config/i386/predicates.md (general_vector_operand): New predicate.
14191 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
14192 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
14193 if they aren't nonimmediate operands. If their original values
14194 satisfy const_vector_equal_evenodd_p, don't shift them.
14195 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
14196 predicates. For the SSE4.1 case force operands[{1,2}] into registers
14197 if not nonimmediate_operand.
14198 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
14199 instead of register_operand.
14200 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
14202 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
14204 * stor-layout.c (finalize_size_functions): Allocate a structure and
14205 reset cfun before dumping the functions.
14207 2013-04-27 Jakub Jelinek <jakub@redhat.com>
14209 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
14212 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
14213 use xop_pmacsdqh if uns_p.
14214 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
14215 the immediate rotate count.
14217 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
14219 * rtl.h (struct rtx_def): Add comment for field jump.
14220 (LRA_SUBREG_P): New macro.
14221 * recog.c (register_operand): Check LRA_SUBREG_P.
14222 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14224 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14225 (lra_final_code_change): Skip subreg change for operators.
14226 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14227 if there are no operand changes.
14228 * lra-constraints.c (curr_insn_set): New.
14229 (match_reload): Set LRA_SUBREG_P.
14230 (emit_spill_move): Ditto.
14231 (check_and_process_move): Use curr_insn_set. Process only single
14232 set insns. Don't initialize sec_mem_p and change_p.
14233 (simplify_operand_subreg): Use LRA_SUBREG_P.
14234 (reg_in_class_p): New function.
14235 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14236 of #ifdef. Add code to remove cycling.
14237 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14238 non-null disp. Reload inner instead of disp when base and index
14239 are null. Try to put lo_sum into register.
14240 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14241 (check_and_process_move): Move code for move cost check to
14242 simple_move_p. Remove equiv_substitution.
14243 (simple_move_p): New function.
14244 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14245 curr_insn_set. Call check_and_process_move only for single set
14246 insns. Use the new function. Move call of check_and_process_move
14247 after operand equiv substitution and address process.
14249 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14252 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
14253 with nonlocal goto receivers or returns twice calls, ignore
14254 unininitialized values from abnormal edges to nl goto receiver
14255 or returns twice call.
14257 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14259 PR tree-optimization/57051
14260 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
14261 and VEC_RSHIFT_EXPR if shift count is a multiple of element
14264 2013-04-26 Richard Biener <rguenther@suse.de>
14266 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
14267 (expand_omp_taskreg): Likewise. Mark loops for fixup.
14268 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
14269 (fixup_loop_arrays_after_move): New function.
14270 (move_sese_region_to_fn): Properly outline the loop tree parts
14271 of the SESE region.
14273 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
14275 * config/i386/i386.md (type, unit): Fix long lines.
14277 2013-04-26 Richard Biener <rguenther@suse.de>
14279 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
14280 (lto-streamer-out.o): Likewise.
14281 * cfgloop.c (init_loops_structure): Export, add struct function
14282 argument and adjust.
14283 (flow_loops_find): Adjust.
14284 * cfgloop.h (enum loop_estimation): Add EST_LAST.
14285 (init_loops_structure): Declare.
14286 * lto-streamer-in.c: Include cfgloop.h.
14287 (input_cfg): Input the loop tree.
14288 * lto-streamer-out.c: Include cfgloop.h.
14289 (output_cfg): Output the loop tree.
14290 (output_struct_function_base): Do not drop PROP_loops.
14292 2013-03-26 Richard Biener <rguenther@suse.de>
14294 * tree-cfg.c (execute_build_cfg): Build the loop tree.
14295 (pass_build_cfg): Provide PROP_loops.
14296 (move_sese_region_to_fn): Remove loops that are outlined into fn
14298 * tree-inline.c: Include cfgloop.h.
14299 (initialize_cfun): Do not drop PROP_loops.
14300 (copy_loops): New function.
14301 (copy_cfg_body): Copy loop structure.
14302 (tree_function_versioning): Initialize destination loop tree.
14303 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
14304 (pass_parallelize_loops): Do IL verification.
14305 * loop-init.c (loop_optimizer_init): Fixup loops if required.
14306 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
14307 the CFG make sure we fixup loops as well.
14308 * tree-ssa-tail-merge.c: Include cfgloop.h.
14309 (replace_block_by): When merging loop latches mark loops for fixup.
14310 * lto-streamer-out.c (output_struct_function_base): Drop
14311 PROP_loops for now.
14312 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
14313 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
14314 * ipa-split.c: Include cfgloop.h.
14315 (split_function): Add the new return block to the loop tree root.
14316 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
14317 whether we have removed the forwarder block.
14318 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
14319 * cfgloop.h (place_new_loop): Declare.
14320 * cfgloopmanip.c (place_new_loop): Export.
14321 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
14322 (tree-switch-conversion.o): Likewise.
14323 (tree-complex.o): Likewise.
14324 (tree-inline.o): Likewise.
14325 (tree-ssa-tailmerge.o): Likewise.
14326 (ipa-split.o): Likewise.
14327 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
14328 (tree-ssa-copy.o): Likewise.
14329 * tree-switch-conversion.c: Include cfgloop.h
14330 (process_switch): If we emit a bit-test cascade, schedule loops
14332 * tree-complex.c: Include cfgloop.h.
14333 (expand_complex_div_wide): Properly add new basic-blocks to loops.
14334 * asan.c: Include cfgloop.h.
14335 (create_cond_insert_point): Properly add new basic-blocks to
14336 loops, schedule loop fixup.
14337 * cfgloop.c (verify_loop_structure): Check that looks are not
14339 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
14341 (expand_omp_for_generic): Likewise.
14342 (expand_omp_sections): Likewise.
14343 (expand_omp_atomic_pipeline): Schedule loops for fixup.
14344 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
14345 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
14346 is initialized, not when loops are present.
14347 * tree-parloops.c (parallelize_loops): Remove checking here.
14348 * passes.c (init_optimization_passes): Schedule a copy-propagation
14349 pass before complete unrolling of inner loops.
14351 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14353 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
14354 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
14355 (colorize_init): Add argument to _WIN32 version.
14356 * toplev.c: Include diagnostic-color.h.
14357 (process_options): Default to -fdiagnostics-color=auto if
14358 GCC_COLORS env var is in the environment.
14359 * common.opt (fdiagnostics-color=): Add Var and Init.
14360 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
14361 env var is in the environment, the default is auto rather than never.
14363 * diagnostic.h (file_name_as_prefix): Add context argument.
14364 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
14365 the string as locus.
14366 * langhooks.c (lhd_print_error_function): Adjust caller.
14368 2013-04-25 Lawrence Crowl <crowl@google.com>
14370 * var-tracking.c (shared_hash_def::htab):
14371 Change type to hash_table. Update dependent calls and types.
14373 2013-04-25 Lawrence Crowl <crowl@google.com>
14375 * Makefile.in: Update as needed below.
14377 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
14378 Move declaration to after the type's method definitons.
14380 * attribs.c (htab_t scoped_attributes::attribute_hash):
14381 Change type to hash_table. Update dependent calls and types.
14383 * bitmap.c (htab_t bitmap_desc_hash):
14384 Change type to hash_table. Update dependent calls and types.
14386 * cselib.c (htab_t cselib_hash_table):
14387 Change type to hash_table. Update dependent calls and types.
14389 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
14390 (hash_string_slot_node): Move implementation into lto-streamer.h
14391 struct string_slot_hasher.
14392 (eq_string_slot_node): Likewise.
14394 * data-streamer-out.c: Update output_block::string_hash_table
14395 dependent calls and types.
14397 * dwarf2cfi.c (htab_t trace_index):
14398 Change type to hash_table. Update dependent calls and types.
14400 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
14401 Change type to hash_table. Update dependent calls and types.
14402 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
14403 (htab_t optimize_external_refs::map): Likewise.
14404 (htab_t output_comp_unit::extern_map): Likewise.
14405 (htab_t output_comdat_type_unit::extern_map): Likewise.
14406 (htab_t output_macinfo::macinfo_htab): Likewise.
14407 (htab_t optimize_location_lists::htab): Likewise.
14408 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
14410 * except.c (htab_t ehspec_hash_type):
14411 Change type to hash_table. Update dependent calls and types.
14412 (assign_filter_values::ttypes): Likewise.
14413 (assign_filter_values::ehspec): Likewise.
14414 (sjlj_assign_call_site_values::ar_hash): Likewise.
14415 (convert_to_eh_region_ranges::ar_hash): Likewise.
14417 * gcse.c (htab_t pre_ldst_table):
14418 Change type to hash_table. Update dependent calls and types.
14420 * ggc-common.c (htab_t saving_htab):
14421 Change type to hash_table. Update dependent calls and types.
14422 (htab_t loc_hash): Likewise.
14423 (htab_t ptr_hash): Likewise.
14424 (call_count): Rename ggc_call_count.
14425 (call_alloc): Rename ggc_call_alloc.
14426 (loc_descriptor): Rename make_loc_descriptor.
14427 (add_statistics): Rename ggc_add_statistics.
14429 * ggc-common.c (saving_htab):
14430 Change type to hash_table. Update dependent calls and types.
14432 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
14433 (push_gimplify_context): Likewise.
14434 (pop_gimplify_context): Likewise.
14435 (struct gimple_temp_hash_elt): Added.
14436 (struct gimplify_hasher): Likewise.
14437 (struct gimplify_ctx.temp_htab):
14438 Change type to hash_table. Update dependent calls and types.
14440 * gimple-fold.c: Include gimplify-ctx.h.
14442 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
14443 Change type to hash_table. Update dependent calls and types.
14444 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
14445 avoid potential global name collision.
14447 * gimplify.c: Include gimplify-ctx.h.
14448 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
14449 (htab_t gimplify_ctx::temp_htab):
14450 Update dependent calls and types for new type hash_table.
14451 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
14452 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
14454 * gimplify-ctx.h: New.
14455 (struct gimple_temp_hash_elt): Move from gimplify.c.
14456 (class gimplify_hasher): New.
14457 (struct gimplify_ctx): Move from gimple.h.
14458 (htab_t gimplify_ctx::temp_htab):
14459 Change type to hash_table. Update dependent calls and types.
14461 * graphite-clast-to-gimple.c: Include graphite-htab.h.
14462 (htab_t ivs_params::newivs_index):
14463 Change type to hash_table. Update dependent calls and types.
14464 (htab_t ivs_params::params_index): Likewise.
14465 (htab_t print_generated_program::params_index): Likewise.
14466 (htab_t gloog::newivs_index): Likewise.
14467 (htab_t gloog::params_index): Likewise.
14469 * graphite.c: Include graphite-htab.h.
14470 4htab_t graphite_transform_loops::bb_pbb_mapping):
14471 Change type to hash_table. Update dependent calls and types.
14473 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
14474 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
14475 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
14477 * graphite-dependences.c: Include graphite-htab.h.
14478 (loop_is_parallel_p): Change hash table type of parameter.
14480 * graphite-htab.h: New.
14481 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
14482 (extern find_pbb_via_hash): Move from graphite-poly.h.
14483 (extern loop_is_parallel_p): Move from graphite-poly.h.
14484 (extern get_loop_body_pbbs): Move from graphite-poly.h.
14486 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
14487 (extern loop_is_parallel_p): Move to graphite-htab.h.
14488 (extern get_loop_body_pbbs): Move to graphite-htab.h.
14490 * haifa-sched.c (htab_t delay_htab):
14491 Change type to hash_table. Update dependent calls and types.
14492 (htab_t delay_htab_i2): Likewise.
14494 * ira-color.c (htab_t allocno_hard_regs_htab):
14495 Change type to hash_table. Update dependent calls and types.
14497 * ira-costs.c (htab_t cost_classes_htab):
14498 Change type to hash_table. Update dependent calls and types.
14500 * loop-invariant.c (htab_t merge_identical_invariants::eq):
14501 Change type to hash_table. Update dependent calls and types.
14503 * loop-iv.c (htab_t bivs):
14504 Change type to hash_table. Update dependent calls and types.
14506 * loop-unroll.c (htab_t opt_info::insns_to_split):
14507 Change type to hash_table. Update dependent calls and types.
14508 (htab_t opt_info::insns_with_var_to_expand): Likewise.
14510 * lto-streamer.h (struct string_slot): Move from data-streamer.h
14511 (struct string_slot_hasher): New.
14512 (htab_t output_block::string_hash_table):
14513 Change type to hash_table. Update dependent calls and types.
14515 * lto-streamer-in.c (freeing_string_slot_hasher): New.
14516 (htab_t file_name_hash_table):
14517 Change type to hash_table. Update dependent calls and types.
14519 * lto-streamer-out.c: Update output_block::string_hash_table dependent
14522 * lto-streamer.c (htab_t tree_htab):
14523 Change type to hash_table. Update dependent calls and types.
14525 * omp-low.c: Include gimplify-ctx.h.
14527 * passes.c (htab_t name_to_pass_map):
14528 Change type to hash_table. Update dependent calls and types.
14529 (pass_traverse): Rename to passes_pass_traverse.
14531 * plugin.c (htab_t event_tab):
14532 Change type to hash_table. Update dependent calls and types.
14534 * postreload-gcse.c (htab_t expr_table):
14535 Change type to hash_table. Update dependent calls and types.
14536 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
14538 * sese.c (debug_rename_map_1): Make extern.
14539 (htab_t copy_bb_and_scalar_dependences::rename_map):
14540 Change type to hash_table. Update dependent calls and types.
14542 * sese.h (extern debug_rename_map): Move to .c file.
14544 * store-motion.c (htab_t store_motion_mems_table):
14545 Change type to hash_table. Update dependent calls and types.
14547 * trans-mem.c (htab_t tm_new_mem_hash):
14548 Change type to hash_table. Update dependent calls and types.
14550 * tree-browser.c (htab_t TB_up_ht):
14551 Change type to hash_table. Update dependent calls and types.
14553 * tree-cfg.c (htab_t discriminator_per_locus):
14554 Change type to hash_table. Update dependent calls and types.
14556 * tree-complex.c: Include tree-hasher.h
14557 (htab_t complex_variable_components):
14558 Change type to hash_table. Update dependent calls and types.
14560 * tree-eh.c (htab_t finally_tree):
14561 Change type to hash_table. Update dependent calls and types.
14563 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
14564 struct int_tree_hasher.
14565 (extern int_tree_map_eq): Likewise.
14566 (uid_decl_map_hash): Removed.
14567 (extern decl_tree_map_eq): Likewise.
14569 * tree-hasher.h: New.
14570 (struct int_tree_hasher): New.
14571 (typedef int_tree_htab_type): New.
14573 * tree-inline.c: Include gimplify-ctx.h.
14575 * tree-mudflap.c: Include gimplify-ctx.h.
14577 * tree-parloops.c: Include tree-hasher.h.
14578 (htab_t eliminate_local_variables_stmt::decl_address):
14579 Change type to hash_table. Update dependent calls and types.
14580 (htab_t separate_decls_in_region::decl_copies): Likewise.
14582 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
14583 Change type to hash_table. Update dependent calls and types.
14585 * tree-sra.c (candidates):
14586 Change type to hash_table. Update dependent calls and types.
14588 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
14590 (int_tree_map_hash): Likewise.
14592 * tree-ssa-dom.c (htab_t avail_exprs):
14593 Change type to hash_table. Update dependent calls and types.
14595 * tree-ssa-live.c (var_map_base_init::tree_to_index):
14596 Change type to hash_table. Update dependent calls and types.
14598 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
14599 Change type to hash_table. Update dependent calls and types.
14601 * tree-ssa-phiopt.c (seen_ssa_names):
14602 Change type to hash_table. Update dependent calls and types.
14604 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
14605 Change type to hash_table. Update dependent calls and types.
14607 * tree-ssa-uncprop.c (equiv):
14608 Change type to hash_table. Update dependent calls and types.
14610 2013-04-25 Jakub Jelinek <jakub@redhat.com>
14612 PR rtl-optimization/57003
14613 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
14614 call note_stores with kill_clobbered_value callback again after
14615 killing regs_invalidated_by_call.
14617 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14619 * config/aarch64/aarch64-simd.md
14620 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
14621 (aarch64_simd_bsl<mode>): Likewise.
14622 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
14624 2013-04-25 Marek Polacek <polacek@redhat.com>
14626 PR tree-optimization/57066
14627 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
14629 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
14631 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
14633 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14635 * config/aarch64/aarch64-builtins.c
14636 (aarch64_fold_builtin): New.
14637 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
14638 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
14639 * config/aarch64/aarch64-simd-builtins.def (abs): New.
14640 * config/aarch64/arm_neon.h
14641 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
14643 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14644 Tejas Belagod <tejas.belagod@arm.com>
14646 * config/aarch64/aarch64-builtins.c
14647 (aarch64_gimple_fold_builtin): New.
14648 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
14649 * config/aarch64/aarch64-simd-builtins.def (addv): New.
14650 * config/aarch64/aarch64-simd.md (addpv4sf): New.
14651 (addvv4sf): Update.
14652 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
14654 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14656 * config/aarch64/aarch64.md
14657 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
14659 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14661 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
14662 (*ngcsi_uxtw): New pattern.
14664 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14665 Julian Brown <julian@codesourcery.com>
14667 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
14668 (TB_DREG): Add T_V4HF.
14669 (v4hf_UP): New macro.
14670 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14671 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14672 Handle initialisation of V4HF. Adjust initialisation of reinterpret
14674 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14675 (arm_vector_mode_supported_p): Handle V4HF.
14676 (arm_mangle_map): Handle V4HFmode.
14677 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
14678 * config/arm/arm_neon_builtins.def: Add entries for
14679 vcvtv4hfv4sf, vcvtv4sfv4hf.
14680 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
14681 (neon_vcvtv4hfv4sf): Likewise.
14682 * config/arm/neon-gen.ml: Handle half-precision floating point
14684 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
14685 * config/arm/arm_neon.h: Regenerate.
14686 * config/arm/neon.ml (type elts): Add F16.
14687 (type vectype): Add T_float16x4, T_floatHF.
14688 (type vecmode): Add V4HF.
14689 (type features): Add Requires_FP_bit feature.
14690 (elt_width): Handle F16.
14691 (elt_class): Likewise.
14692 (elt_of_class_width): Likewise.
14693 (mode_of_elt): Refactor.
14694 (type_for_elt): Handle F16, fix error messages.
14695 (vectype_size): Handle T_float16x4.
14696 (vcvt_sh): New function.
14697 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
14698 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
14699 (string_of_mode): Handle V4HF.
14700 * doc/arm-neon-intrinsics.texi: Regenerate.
14702 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14704 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
14705 format specifier in 'X' case.
14707 2013-04-25 Alan Modra <amodra@gmail.com>
14710 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
14711 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
14712 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
14713 Repeat for many other rotate/shift and mask patterns using subregs.
14714 Name lshiftrt insns.
14715 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
14716 on WORDS_BIG_ENDIAN.
14718 2013-04-25 Alan Modra <amodra@gmail.com>
14720 * config.gcc: Support little-endian powerpc-linux targets.
14721 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
14722 (LINK_OS_LINUX_SPEC): Define.
14723 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
14724 Preserve MASK_LITTLE_ENDIAN.
14725 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
14726 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
14727 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
14728 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
14729 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
14730 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
14731 Correct fp word order for little-endian. Don't shift toc entries
14732 smaller than a word for little-endian.
14733 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
14734 (bswapdi2 splits): Correct low-part subreg for little-endian.
14735 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
14736 low/high where such is correct only for be.
14737 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
14738 little-endian for -mcall-aixdesc.
14740 2013-04-25 Alan Modra <amodra@gmail.com>
14742 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
14743 replace_equiv_address_nv.
14745 2013-04-25 Alan Modra <amodra@gmail.com>
14747 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
14749 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14752 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14753 * rtl.h (struct rtx_def): ...
14755 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14757 PR rtl-optimizations/57046
14758 * lra-constraints (split_reg): Set up lra_risky_transformations_p
14759 for multi-reg splits.
14761 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
14763 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
14765 2013-04-24 Sterling Augustine <saugustine@google.com>
14767 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
14768 (comp_dir_string, debug_str_dwo_section): New.
14769 (DEBUG_STR_DWO_SECTION): Rename to ...
14770 (DEBUG_DWO_STR_SECTION): ... this.
14771 (DEBUG_NORM_STR_SECTION): Delete.
14772 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
14773 (DEBUG_STR_DWO_SECTION_FLAGS): New.
14774 (find_AT_string): Move most logic to ...
14775 (find_AT_string_in_table): ... here. New.
14776 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
14777 add_skeleton_AT_string. Delete logic.
14778 (output_skeleton_debug_sections): Remove call to
14779 add_top_level_skeleton_die_attrs.
14780 (add_comp_dir_attribute): Move logic to comp_dir_string.
14781 (dwarf2out_init): Initialize debug_str_dwo_section.
14782 (output_indirect_string): Call find_string_form.
14783 (output_indirect_strings): Rewrite.
14784 (prune_unused_types): Empty skeleton_debug_str_hash.
14785 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
14786 (dwarf2out_finish): Call output_indirect_strings.
14788 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14790 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
14792 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14794 * rtl.h (struct rtx_def): Add comment for field jump.
14795 (LRA_SUBREG_P): New macro.
14796 * recog.c (register_operand): Check LRA_SUBREG_P.
14797 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14799 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14800 (lra_final_code_change): Skip subreg change for operators.
14801 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14802 if there are no operand changes.
14803 * lra-constraints.c (curr_insn_set): New.
14804 (match_reload): Set LRA_SUBREG_P.
14805 (emit_spill_move): Ditto.
14806 (check_and_process_move): Use curr_insn_set. Process only single
14807 set insns. Don't initialize sec_mem_p and change_p.
14808 (simplify_operand_subreg): Use LRA_SUBREG_P.
14809 (reg_in_class_p): New function.
14810 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14811 of #ifdef. Add code to remove cycling.
14812 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14813 non-null disp. Reload inner instead of disp when base and index
14814 are null. Try to put lo_sum into register.
14815 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14816 (check_and_process_move): Move code for move cost check to
14817 simple_move_p. Remove equiv_substitution.
14818 (simple_move_p): New function.
14819 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14820 curr_insn_set. Call check_and_process_move only for single set
14821 insns. Use the new function. Move call of check_and_process_move
14822 after operand equiv substitution and address process.
14824 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
14826 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
14827 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
14828 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
14830 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14832 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
14834 2013-04-24 Marek Polacek <polacek@redhat.com>
14836 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
14837 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
14838 (select_loops_exit_conditions): Likewise.
14839 (number_of_iterations_for_all_loops): Likewise.
14840 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
14841 (scev_analysis): Likewise.
14843 2013-04-02 Catherine Moore <clm@codesourcery.com>
14844 Chao-ying Fu <fu@mips.com>
14846 * config/mips/micromips.md (jraddiusp): New pattern.
14847 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
14848 instruction if possible.
14850 2013-04-24 Alan Modra <amodra@gmail.com>
14852 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
14854 2013-04-24 Julian Brown <julian@codesourcery.com>
14855 Chung-Lin Tang <cltang@codesourcery.com>
14857 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
14858 dependency behavior in enumeration type DIE generation. Add TODO note
14859 to comments about future DW_FORM_sdata/udata re-work of related code.
14861 2013-04-23 Lawrence Crowl <crowl@google.com>
14863 * Makefile.in: Update as needed below.
14865 * hash-table.h (class hash_table):
14866 Correct many methods with parameter types compare_type to the correct
14867 value_type. (Correct code was unlikely to notice the change.)
14868 (hash_table::elements_with_deleted) New.
14869 (class hashtable::iterator): New.
14870 (hashtable::begin()): New.
14871 (hashtable::end()): New.
14872 (FOR_EACH_HASH_TABLE_ELEMENT): New.
14874 * statistics.c (statistics_hashes):
14875 Change type to hash_table. Update dependent calls and types.
14877 * tree-into-ssa.c (var_infos):
14878 Change type to hash_table. Update dependent calls and types.
14880 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
14881 Change type to hash_table. Update dependent calls and types.
14883 * tree-ssa-loop-im.c (struct mem_ref.refs):
14884 Change type to hash_table. Update dependent calls and types.
14886 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
14887 Change type to hash_table. Update dependent calls and types.
14889 * tree-ssa-sccvn.c (vn_tables_s::nary):
14890 Change type to hash_table. Update dependent calls and types.
14891 (vn_tables_s::phis): Likewise.
14892 (vn_tables_s::references): Likewise.
14894 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
14895 (vn_reference_eq): Update parameter and return types.
14897 * tree-ssa-structalias.c (pointer_equiv_class_table):
14898 Change type to hash_table. Update dependent calls and types.
14899 (location_equiv_class_table): Likewise.
14901 * tree-vect-data-refs.c: Consequential changes for making
14902 peeling a hash_table.
14904 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
14905 (destroy_loop_vec_info): Dependent hash_table update.
14907 * tree-vectorizer.h (peeling_htab):
14908 Change type to hash_table. Update dependent calls and types.
14910 2013-04-23 Shiva Chen <shiva0217@gmail.com>
14912 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
14913 to check the register content is equal or not.
14914 * lra-constraints.c (match_reload): Use lra_assign_reg_val
14915 to assign register content record.
14916 * lra-eliminations.c (update_reg_eliminate): Use
14917 lra_update_reg_val_offset to update register content offset.
14918 * lra-int.h (struct lra_reg): Add offset member.
14919 (lra_reg_val_equal_p): New static inline function.
14920 (lra_update_reg_val_offset): New static inline function.
14921 (lra_assign_reg_val): New static inline function.
14922 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
14923 to assign register content record.
14924 (initialize_lra_reg_info_element): Initial offset to zero.
14926 2013-04-23 Catherine Moore <clm@codesourcery.com>
14928 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
14929 operands. Record compression.
14931 2013-04-23 Xinliang David Li <davidxl@google.com>
14933 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
14935 2013-04-23 Richard Biener <rguenther@suse.de>
14937 PR middle-end/57036
14938 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
14939 parameter, only add abnormal goto edges from the copied body
14940 if the call could perform abnormal gotos.
14941 (copy_cfg_body): Adjust.
14943 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
14945 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
14947 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
14949 * coretypes.h (gimple_stmt_iterator): Add struct to make
14952 2013-04-23 Richard Biener <rguenther@suse.de>
14954 PR tree-optimization/57026
14955 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
14956 from SSA names occuring in abnormal PHI nodes.
14958 2013-04-22 Andi Kleen <ak@linux.intel.com>
14960 * lto/lto.c (print_lto_report_1): Fix LTO report names.
14962 2013-04-22 Andi Kleen <ak@linux.intel.com>
14964 * lto/lto.c (print_lto_report_1): Declare early.
14965 (read_cgraph_and_symbols): Call print_lto_report_1 early.
14967 2013-04-22 Andi Kleen <ak@linux.intel.com>
14969 * common.opt (-flto-report-wpa): Add.
14970 * doc/invoke.texi (-flto-report-wpa): Add.
14971 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
14974 2013-04-22 Xinliang David Li <davidxl@google.com>
14976 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
14977 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
14978 * Makefile.in: New dependency
14980 David Daney <ddaney.cavm@gmail.com>
14982 * configure.ac (gcc_cv_as_micromips_support): Use the
14983 --fatal-warnings option.
14984 * configure: Regenerate.
14986 2013-04-22 Marek Polacek <polacek@redhat.com>
14989 * tsan.c (instrument_expr): Don't instrument expression
14990 in case its size is zero.
14992 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
14996 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14998 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15000 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15002 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15003 (gimple_stmt_iterator): New typedef.
15004 * gimple.h (gimple_stmt_iterator): Rename to...
15005 (gimple_stmt_iterator_d): ... This.
15006 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15007 trees be valid for GIMPLE and GENERIC.
15008 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15009 * gimple-fold.c (gimple_fold_call): Call target hook
15010 gimple_fold_builtin.
15011 * hooks.c (hook_bool_gsiptr_false): New.
15012 * hooks.h (hook_bool_gsiptr_false): New.
15013 * target.def (fold_stmt): New.
15014 * doc/tm.texi: Regenerate.
15016 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15019 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15020 a set sp if no stack realignment.
15022 2013-04-22 Nick Clifton <nickc@redhat.com>
15024 * config.gcc (tilegx-linux): Extend extra_objs rather than
15026 (tilepro-linux): Likewise.
15028 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15030 * config/aarch64/aarch64-builtins.c
15032 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15033 (VAR<1-12>): Add MAP parameter.
15034 (BUILTIN_*): Likewise.
15035 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15036 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15037 (aarch64_ushl_n<mode>): Likewise.
15038 (aarch64_sshr_n<mode>): Likewise.
15039 (aarch64_ushr_n<mode>): Likewise.
15040 (aarch64_<maxmin><mode>): Likewise.
15041 (aarch64_sqrt<mode>): Likewise.
15042 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15043 (vshr<q>_n_*): Likewise.
15045 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15047 * config/aarch64/aarch64-builtins.c
15048 (aarch64_simd_builtin_type_mode): Handle SF types.
15050 (BUILTIN_GPF): Define.
15051 (aarch64_init_simd_builtins): Handle SF types.
15052 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
15053 (frecps): Likewise.
15054 (frecpx): Likewise.
15055 * config/aarch64/aarch64-simd.md
15056 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
15057 (aarch64_frecpe<mode>): New.
15058 (aarch64_frecps<mode>): Likewise.
15059 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
15060 (v8type): Add frecp<esx>.
15061 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
15062 (aarch64_frecps<mode>): Likewise.
15063 * config/aarch64/iterators.md (FRECP): New.
15064 (frecp_suffix): Likewise.
15065 * config/aarch64/arm_neon.h
15066 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
15068 2013-04-22 Christian Bruel <christian.bruel@st.com>
15071 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
15072 (REG_CLASS_NAMES): Idem.
15073 (REG_CLASS_CONTENTS): Idem.
15074 (REGCLASS_HAS_FP_REG): Idem.
15075 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
15076 (sh_conditional_register_usage): Idem.
15078 2013-04-21 Jeff Law <law@redhat.com>
15080 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
15081 (ssa_forward_propagate_and_combine): Use it.
15083 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15085 * lra.c: Update the flow chart diagram.
15087 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15089 PR rtl-optimization/56847
15090 * lra-constraints.c (process_alt_operands): Discourage alternative
15091 with non-matche doffsettable memory constraint fro memory with
15094 2013-04-19 Richard Biener <rguenther@suse.de>
15096 PR tree-optimization/56982
15097 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
15099 * gimplify.c (gimplify_call_expr): Notice special calls.
15100 (gimplify_modify_expr): Likewise.
15101 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
15102 abnormal control flow receivers.
15103 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
15104 in the same way as cfun->has_nonlocal_labels.
15105 (gimple_purge_dead_abnormal_call_edges): Likewise.
15106 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
15107 receivers start a basic-block.
15109 2013-04-19 Richard Biener <rguenther@suse.de>
15111 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
15113 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
15114 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
15115 (SLP_TREE_LOAD_PERMUTATION): Add.
15116 (vect_transform_slp_perm_load): Adjust prototype.
15117 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
15118 (vect_free_slp_instance): Likewise.
15119 (vect_create_new_slp_node): Likewise.
15120 (vect_supported_slp_permutation_p): Remove.
15121 (vect_slp_rearrange_stmts): Adjust.
15122 (vect_supported_load_permutation_p): Likewise. Inline
15123 vect_supported_slp_permutation_p here.
15124 (vect_analyze_slp_instance): Compute load permutations per
15125 slp node instead of per instance.
15126 (vect_get_slp_defs): Adjust.
15127 (vect_transform_slp_perm_load): Likewise.
15128 (vect_schedule_slp_instance): Remove redundant code.
15129 (vect_schedule_slp): Remove hack for PR56270, add it ...
15130 * tree-vect-stmts.c (vectorizable_load): ... here, do not
15131 CSE loads for SLP. Adjust.
15133 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15135 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
15136 spelling in two comments.
15138 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15141 * config/arm/arm.c (load_multiple_sequence): Require SP
15142 as base register for loads if SP is in the register list.
15144 2013-04-19 Martin Jambor <mjambor@suse.cz>
15146 PR tree-optimization/56718
15147 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
15148 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
15149 and made public. Adjusted all callers.
15150 (ipa_intraprocedural_devirtualization): New function.
15151 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
15152 (ipa_intraprocedural_devirtualization): Likewise.
15153 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
15155 2013-04-19 Richard Biener <rguenther@suse.de>
15157 PR tree-optimization/57000
15158 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
15160 2013-04-19 Terry Guo <terry.guo@arm.com>
15162 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
15164 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
15165 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
15166 (cortex_m4_fmacs): Use new reservations.
15167 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
15169 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
15171 PR rtl-optimization/56999
15172 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
15174 (lra_coalesce): Remove split_origin_bitmap and related code.
15175 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
15176 ranges if necessary.
15178 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
15180 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
15182 (ix86_expand_call): Remove clobbered_registers array and use
15183 x86_64_ms_sysv_extra_clobbered_registers instead.
15184 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
15186 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
15188 * config/i386/i386.md (*call_rex64_ms_sysv): Use
15189 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
15190 (*call_value_rex64_ms_sysv): Ditto.
15192 2013-04-18 Cary Coutant <ccoutant@google.com>
15194 * dwarf2out.c (output_pubnames): Check die_perennial_p of
15195 parent instead of die_mark.
15197 2013-04-18 Diego Novillo <dnovillo@google.com>
15199 * gimple.c (create_gimple_tmp): New.
15200 (get_expr_type): New.
15201 (build_assign): New.
15202 (build_type_cast): New.
15203 * gimple.h (enum ssa_mode): Define.
15204 (gimple_seq_set_location): New.
15205 * asan.c (build_check_stmt): Change some gimple_build_* calls
15206 to use build_assign and build_type_cast.
15208 2013-04-18 Richard Biener <rguenther@suse.de>
15210 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
15211 handle negative step. Remove redundant checks.
15212 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
15213 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
15214 for negative step and grouped loads fail to vectorize.
15216 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15218 * emit-rtl.c (reset_insn_used_flags): New function.
15219 (reset_all_used_flags): Use it.
15220 (verify_insn_sharing): New function.
15221 (verify_rtl_sharing): Fix verification for SEQUENCEs.
15223 2013-04-18 Jakub Jelinek <jakub@redhat.com>
15225 PR tree-optimization/56984
15226 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
15227 and (x >> M) >= N don't register any assertion if N << M is the
15230 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15232 * lower-subreg.c (resolve_simple_move): If called self-recursive,
15233 do not delete_insn insns that have not yet been emitted, only
15234 unlink them with remove_insn.
15235 * df-scan.c (df_insn_delete): Revert r197492.
15237 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
15239 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
15240 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
15242 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15244 * config/arm/arm.md (movsicc_insn): Convert define_insn into
15245 define_insn_and_split.
15246 (and_scc,ior_scc,negscc): Likewise.
15247 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
15249 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15251 * config/arm/arm.c (use_return_insn): Return 0 for targets that
15252 can benefit from using a sequence of LDRD instructions in epilogue
15253 instead of a single LDM instruction.
15255 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
15258 * doc/extend.texi: Fix typo.
15260 2013-04-17 Richard Biener <rguenther@suse.de>
15262 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
15263 (vect_build_slp_tree): ... here.
15264 (vect_build_slp_tree_1): Compute which stmts of the SLP group
15265 match. Remove special-casing of mismatched complex loads.
15266 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
15267 re-try the match with swapped commutative operands.
15268 (vect_supported_load_permutation_p): Remove special-casing of
15269 mismatched complex loads.
15270 (vect_analyze_slp_instance): Adjust.
15272 2013-04-17 Richard Biener <rguenther@suse.de>
15274 PR rtl-optimization/56921
15275 * cfgloop.h (struct loop): Add simple_loop_desc member.
15276 (struct niter_desc): Mark with GTY(()).
15277 (simple_loop_desc): Do not use aux field but simple_loop_desc.
15278 * loop-iv.c (get_simple_loop_desc): Likewise.
15279 (free_simple_loop_desc): Likewise.
15282 2013-04-16 Richard Biener <rguenther@suse.de>
15284 PR rtl-optimization/56921
15285 * loop-init.c (pass_rtl_move_loop_invariants): Add
15286 TODO_do_not_ggc_collect to todo_flags_finish.
15287 (pass_rtl_unswitch): Same.
15288 (pass_rtl_unroll_and_peel_loops): Same.
15289 (pass_rtl_doloop): Same.
15291 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
15293 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
15294 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
15295 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
15297 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
15298 * tree-streamer.c (record_common_node): Adjust reference in comment.
15300 2013-04-17 Terry Guo <terry.guo@arm.com>
15302 * config/arm/cortex-m4.md: Add a new bypass.
15304 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15306 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
15308 (*subs_<optab><mode>_multp2): New pattern.
15309 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
15310 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
15312 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15314 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
15315 (*subs_mul_imm_<mode>): New pattern.
15317 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
15320 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
15321 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
15322 (vsx_movti_32bit): Change j->wa to O->wa.
15324 2013-04-16 Richard Biener <rguenther@suse.de>
15326 PR rtl-optimization/56921
15327 * loop-init.c (pass_rtl_move_loop_invariants): Add
15328 TODO_do_not_ggc_collect to todo_flags_finish.
15329 (pass_rtl_unswitch): Same.
15330 (pass_rtl_unroll_and_peel_loops): Same.
15331 (pass_rtl_doloop): Same.
15333 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
15335 * config/arm/arm.c (emit_multi_reg_push): New declaration
15336 for an existing function.
15337 (arm_emit_strd_push): New function.
15338 (arm_expand_prologue): Used here.
15339 (arm_emit_ldrd_pop): New function.
15340 (arm_expand_epilogue): Used here.
15341 (arm_get_frame_offsets): Update condition.
15342 (arm_emit_multi_reg_pop): Add a special case for load of a single
15343 register with writeback.
15345 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
15347 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
15350 2013-04-16 Richard Biener <rguenther@suse.de>
15352 PR tree-optimization/56756
15353 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
15354 (first_mem_ref_loc): New.
15355 (execute_sm): Place the load temporarily before a previous
15356 access instead of in the latch edge to ensure its SSA dependencies
15357 are defined at points dominating the load.
15359 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
15361 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
15362 correct fix by moving header and footer insn to the footer of
15363 the merged basic block. Clear BB_END of the merged-away block.
15365 PR middle-end/43631
15366 * emit-rtl.c (make_note_raw): New function.
15367 (link_insn_into_chain): New static inline function.
15368 (add_insn): Use it.
15369 (add_insn_before, add_insn_after): Factor insn chain linking code...
15370 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
15371 using link_insn_into_chain.
15372 (note_outside_basic_block_p): New helper function for emit_note_after
15373 and emit_note_before.
15374 (emit_note_after): Use nobb variant of add_insn_after if the note
15375 should not be contained in a basic block.
15376 (emit_note_before): Use nobb variant of add_insn_before if the note
15377 should not be contained in a basic block.
15378 (emit_note_copy): Use make_note_raw.
15379 (emit_note): Likewise.
15380 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
15381 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
15382 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
15383 the moved barrier the tail of the basic block it follows.
15384 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
15386 2013-04-15 Jakub Jelinek <jakub@redhat.com>
15388 PR tree-optimization/56962
15389 * gimple-ssa-strength-reduction.c (record_increment): Only set
15390 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
15391 either rhs1 or rhs2 is equal to c->base_expr.
15393 2013-04-15 Richard Biener <rguenther@suse.de>
15395 PR tree-optimization/56933
15396 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
15398 (GROUP_READ_WRITE_DEPENDENCE): Remove.
15399 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
15400 * tree-vect-data-refs.c (vect_analyze_group_access): Move
15401 dependence check ...
15402 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
15404 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
15405 GROUP_READ_WRITE_DEPENDENCE.
15407 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15409 * emit-rtl.c (reset_all_used_flags): New function.
15410 (verify_rtl_sharing): Call reset_all_used_flags before and after
15411 performing the checks.
15413 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15415 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
15416 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
15417 * config/arm/constraints.md (De): New constraint.
15418 * config/arm/neon.md (anddi3_neon): Delete.
15419 (neon_vand<mode>): Expand to standard anddi3 pattern.
15420 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
15421 Move earlier in the file.
15422 (neon_inv_logic_op2): Likewise.
15423 (arm_anddi_operand_neon): New predicate.
15425 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15427 * configure.ac (gcc_cv_ld_as_needed): Set
15428 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
15429 Use -z ignore, -z record on *-*-solaris2*.
15430 (HAVE_LD_AS_NEEDED): Update comment.
15431 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
15432 * configure: Regenerate.
15433 * config.in: Regenerate.
15434 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
15435 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
15436 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
15437 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
15438 equivalents. Fix markup.
15439 * doc/tm.texi: Regenerate.
15441 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
15443 * config/i386/i386.opt: New option mstack-protector-guard=.
15444 * config/i386/i386-opts.h: Add enum stack_protector_guard.
15445 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
15446 TARGET_SSP_TLS_GUARD.
15447 * config/i386/i386.c (ix86_option_override_internal): Set
15448 ix86_stack_protector_guard.
15449 * config/i386/i386.md (stack_protect_set): Enable for
15450 TARGET_SSP_TLS_GUARD only.
15451 (stack_protect_set_<mode>): Ditto.
15452 (stack_protect_test): Ditto.
15453 (stack_protect_test_<mode>): Ditto.
15454 * doc/invoke.texi (i386 Option): Document.
15456 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
15459 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
15460 (S_MODES): Set H_MODE bit.
15461 (SF_MODES): Set only S_MODE and SF_MODE bits.
15462 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
15463 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
15464 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
15465 <MODE_FLOAT>: Likewise.
15467 2013-04-15 Joey Ye <joey.ye@arm.com>
15469 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
15471 2013-04-15 Joey Ye <joey.ye@arm.com>
15473 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
15475 (thumb_far_jump_used_p): Count instruction size and set
15478 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
15480 * reorg.c (fill_simple_delay_slots): Reindent block of code.
15481 * resource.c (mark_target_live_regs): Reformat conditional block.
15483 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
15485 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
15486 notes, they are emitted only just before final.
15487 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
15489 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
15491 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
15492 * cfgrtl.c (delete_insn): Call it here instead.
15493 * lra-spills.c (lra_final_code_change): Use delete_insn.
15494 * haifa-sched.c (sched_remove_insn): Likewise.
15495 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
15496 returning to the nop pool.
15497 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
15498 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
15500 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
15502 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
15503 * doc/tm.texi: Regenerated.
15505 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
15507 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
15510 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
15512 * df-core.c (df_find_def): Compare register numbers.
15513 (df_find_use): Likewise.
15515 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
15518 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
15519 lra_in_progress for return.
15521 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
15523 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
15524 define_insn into define_insn_and_split and emit movsicc patterns.
15526 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
15528 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
15530 2013-04-12 Richard Biener <rguenther@suse.de>
15532 * tree-pass.h (TODO_do_not_ggc_collect): New.
15533 * passes.c (execute_one_ipa_transform_pass): Honor
15534 TODO_do_not_ggc_collect.
15535 (execute_one_pass): Likewise.
15538 2013-04-10 Richard Biener <rguenther@suse.de>
15540 * passes.c (init_optimization_passes): Remove reload pass.
15541 * ira.c (do_reload): Merge into ...
15543 (rest_of_handle_reload): Remove.
15544 (pass_reload): Likewise.
15545 * config/i386/i386.c (ix86_option_override): Refer to ira instead
15546 of reload for vzeroupper pass placement.
15548 2013-04-12 Jakub Jelinek <jakub@redhat.com>
15550 PR tree-optimization/56918
15551 PR tree-optimization/56920
15552 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
15553 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
15554 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
15555 use wide_mul_with_sign method.
15557 2013-04-12 Richard Biener <rguenther@suse.de>
15559 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
15560 not be considered a gimple constant.
15562 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
15564 * fold-const.c (const_binop): Handle vector shifts by a scalar.
15565 (fold_binary_loc): Call const_binop also for mixed vector-scalar
15568 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
15569 Jakub Jelinek <jakub@redhat.com>
15571 * opts.c: Include diagnostic-color.h.
15572 (common_handle_option): Handle OPT_fdiagnostics_color_.
15573 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
15574 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
15575 (diagnostic-color.o): New.
15576 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
15577 (diagnostic_color_rule): New enum.
15578 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
15579 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
15580 the location string.
15581 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
15582 either NULL, or color kind.
15583 * diagnostic-color.c: New file.
15584 * diagnostic-color.h: New file.
15585 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
15587 * doc/invoke.texi (-fdiagnostics-color): Document.
15588 * pretty-print.h (pp_show_color): Define.
15589 (struct pretty_print_info): Add show_color field.
15590 * diagnostic.c: Include diagnostic-color.h.
15591 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
15592 macros. Colorize error:, warning: etc. strings and also the location
15594 (diagnostic_show_locus): Colorize the caret line.
15595 * pretty-print.c: Include diagnostic-color.h.
15596 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
15597 inside of %< %> quotes or quoted through q format modifier.
15599 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15601 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
15603 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15605 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
15606 code in CC_NZ mode.
15607 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
15610 2013-04-11 Marek Polacek <polacek@redhat.com>
15612 PR tree-optimization/48184
15613 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
15615 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
15617 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
15618 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
15619 (skip_simple_arithmetic): Tidy up.
15620 * tree.h (skip_simple_constant_arithmetic): Declare.
15622 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15624 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
15626 2013-04-11 Richard Biener <rguenther@suse.de>
15628 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15629 generate vector constants.
15631 2013-04-11 Richard Biener <rguenther@suse.de>
15633 PR tree-optimization/56878
15634 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
15635 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
15637 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
15638 Prefer to align the DR with the most invariant base address.
15640 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
15642 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
15645 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
15647 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
15648 floating-point vector comparisons against 0.
15650 2013-04-11 Jakub Jelinek <jakub@redhat.com>
15652 PR tree-optimization/56899
15653 * fold-const.c (extract_muldiv_1): Apply distributive law
15654 only if TYPE_OVERFLOW_WRAPS (ctype).
15656 2013-04-11 Bin Cheng <bin.cheng@arm.com>
15659 * ira-costs.c (scan_one_insn): Check whether the source rtx of
15660 loading has side effect.
15662 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
15664 * config/sparc/sparc.c: Include tree-pass.h.
15665 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
15666 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
15667 head of file. Change return type. Split off gate function.
15668 (sparc_gate_work_around_errata): New function.
15669 (pass_work_around_errata): New pass definition.
15670 (insert_pass_work_around_errata) New pass insert definition to
15671 insert pass_work_around_errata just after delayed-branch scheduling.
15672 (sparc_option_override): Insert the pass.
15673 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
15675 2013-04-10 David S. Miller <davem@davemloft.net>
15677 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
15678 or -mcpu=hypersparc.
15680 * target.def (cstore_mode): New hook.
15681 * target.h: Include insn-codes.h
15682 * targhooks.c: Likewise.
15683 (default_cstore_mode): New function.
15684 * targhooks.h: Declare it.
15685 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
15686 * doc/tm.texi: Rebuild.
15687 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
15688 target hook, rather than inspecting the insn_data.
15689 * config/sparc/sparc.c (sparc_cstore_mode): New function.
15690 (TARGET_CSTORE_MODE): Redefine.
15691 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
15693 * config/sparc/predicates.md (cstore_result_operand): New special
15695 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
15696 Use it for operand 0.
15697 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
15698 (*snesi_special): Likewise.
15699 (*snesi_zero): Likewise.
15700 (*seqsi_zero): Likewise.
15701 (*sltu_insn): Likewise.
15702 (*sgeu_insn): Likewise.
15703 (*seqdi_special): Make operand 0 and comparison operation be of
15705 (*snedi_special): Likewise.
15706 (*snedi_special_vis3): Likewise.
15707 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
15708 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
15709 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
15710 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
15711 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
15712 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
15713 (*sltu_extend_sp64): Likewise.
15714 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
15715 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
15716 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
15717 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
15718 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
15720 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
15722 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
15723 (aarch64_start_file): Use the new function.
15725 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
15726 Jason Merrill <jason@redhat.com>
15728 * common.opt: Add -gdwarf.
15729 * opts.c (common_handle_option): Handle it.
15730 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
15732 2013-04-10 Richard Biener <rguenther@suse.de>
15734 * passes.c (execute_todo): Do not call ggc_collect conditional here.
15735 (execute_one_ipa_transform_pass): But unconditionally here.
15736 (execute_one_pass): And here.
15737 (init_optimization_passes): Remove reload pass.
15738 * tree-pass.h (TODO_ggc_collect): Remove.
15739 (pass_reload): Likewise.
15740 * ira.c (do_reload): Merge into ...
15742 (rest_of_handle_reload): Remove.
15743 (pass_reload): Likewise.
15744 * config/i386/i386.c (ix86_option_override): Refer to ira instead
15745 of reload for vzeroupper pass placement.
15746 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
15747 and todo_flags_finish of all passes.
15749 2013-04-10 Richard Biener <rguenther@suse.de>
15751 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
15752 first_const_oprnd field, rename first_def_type to first_op_type.
15753 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
15754 (vect_get_and_check_slp_defs): Always use the type of the
15755 operand. Allow mixed vect_external_def, vect_constant_def types.
15756 (vect_get_constant_vectors): Handle mixed vect_external_def,
15757 vect_constant_def types.
15759 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
15761 PR tree-optimization/55524
15762 * tree-ssa-math-opts.c
15763 (convert_mult_to_fma): Don't use an fms construct
15764 when we don't have an fms operation, but fnma, and it looks
15765 likely that we'll be able to use the latter.
15767 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
15769 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
15771 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
15772 inline fail caused by overwritable functions.
15774 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
15776 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
15777 unnecessary bits in the constant power of two case.
15779 2013-04-10 Richard Biener <rguenther@suse.de>
15781 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
15782 broken code swapping operands.
15783 (vect_build_slp_tree): Do not compute load permutations here.
15784 (vect_analyze_slp_instance): Compute load permutations here,
15785 after building the SLP tree.
15787 2013-04-09 Christian Bruel <christian.bruel@st.com>
15789 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
15790 of next/prev_real_insn.
15792 2013-04-09 Jan Hubicka <jh@suse.cz>
15794 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15795 Drop aliased parameter.
15796 (function_and_variable_visibility): Do not handle alias pairs.
15797 * cgraph.c (varpool_externally_visible_p): Update prototype.
15798 * varpool.c (varpool_add_new_variable): Update.
15800 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15802 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
15804 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
15806 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
15808 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
15810 2013-04-09 Marek Polacek <polacek@redhat.com>
15812 PR tree-optimization/48762
15813 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
15815 2013-04-09 Richard Biener <rguenther@suse.de>
15817 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
15819 (vect_build_slp_tree): Likewise.
15820 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
15821 calculating the cost of a SLP instance.
15822 (vect_analyze_slp_instance): Use it from here, after building
15825 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15827 PR middle-end/56883
15828 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
15829 expand_omp_for_static_chunk): Use simple_p = true in
15830 force_gimple_operand_gsi calls when assigning to addressable decls.
15832 2013-04-09 Jeff Law <law@redhat.com>
15834 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
15835 when the boolean was created by converting a wider object which
15836 had a boolean range.
15838 2013-04-09 Richard Biener <rguenther@suse.de>
15840 * tree-vectorizer.h (slp_void_p): Remove.
15841 (slp_tree): Typedef before _slp_tree declaration.
15842 (struct _slp_tree): Use a vector of slp_tree as children.
15843 (vect_get_place_in_interleaving_chain): Remove.
15844 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
15846 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
15848 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
15849 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
15850 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
15851 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
15852 Use slp_node instead of slp_void_p and adjust.
15854 2013-04-09 Richard Biener <rguenther@suse.de>
15856 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
15857 work that is not necessary.
15859 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15861 PR tree-optimization/56854
15862 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
15863 forward into clobber stmts if it would change MEM_REF lhs into
15866 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
15868 * tree.c (type_hash_lookup, type_hash_add): Make static.
15869 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
15871 2013-04-09 Richard Biener <rguenther@suse.de>
15873 * tree.h (unsave_expr_now): Remove.
15874 * tree-inline.c (mark_local_for_remap_r): Remove.
15875 (unsave_expr_1): Likewise.
15876 (unsave_r): Likewise.
15877 (unsave_expr_now): Likewise.
15878 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
15879 (propagate_tree_value): Likewise.
15881 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
15883 * doc/rtl.texi (sequence): Rewrite documentation to match the
15884 current use of SEQUENCE rtl objects.
15885 * rtl.def (SEQUENCE): Likewise.
15887 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
15888 Update documentation.
15889 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
15890 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
15892 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
15894 2013-04-08 Teresa Johnson <tejohnson@google.com>
15896 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
15897 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
15899 (estimate_edge_size_and_time): Add comment to suggest using rounding
15901 (estimate_node_size_and_time): Ditto.
15902 (remap_edge_change_prob): Use helper rounding divide methods.
15903 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
15904 (gimple_mod_pow2_value_transform): Ditto.
15905 (gimple_mod_subtract_transform): Ditto.
15906 (gimple_ic_transform): Ditto.
15907 (gimple_stringops_transform): Ditto.
15908 * stmt.c (conditional_probability): Ditto.
15909 (emit_case_dispatch_table): Ditto.
15910 * lto-cgraph.c (merge_profile_summaries): Ditto.
15911 * tree-optimize.c (execute_fixup_cfg): Ditto.
15912 * cfgcleanup.c (try_forward_edges): Ditto.
15913 * cfgloopmanip.c (scale_loop_profile): Ditto.
15915 (duplicate_loop_to_header_edge): Ditto.
15916 (lv_adjust_loop_entry_edge): Ditto.
15917 * tree-vect-loop.c (vect_transform_loop): Ditto.
15918 * profile.c (compute_branch_probabilities): Ditto.
15919 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
15920 * lto-streamer-in.c (input_cfg): Ditto.
15921 * gimple-streamer-in.c (input_bb): Ditto.
15922 * ipa-cp.c (update_profiling_info): Ditto.
15923 (update_specialized_profile): Ditto.
15924 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
15925 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
15927 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15928 (compute_trg_info): Ditto.
15929 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
15930 (purge_dead_edges): Ditto.
15931 * loop-unswitch.c (unswitch_loop): Ditto.
15932 * cgraphclones.c (cgraph_clone_edge): Ditto.
15933 (cgraph_clone_node): Ditto.
15934 * tree-inline.c (copy_bb): Ditto.
15935 (copy_edges_for_bb): Ditto.
15936 (initialize_cfun): Ditto.
15937 (copy_cfg_body): Ditto.
15938 (expand_call_inline): Ditto.
15940 2013-04-08 Kai Tietz <ktietz@redhat.com>
15942 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
15943 TARGET_CYGWIN64 by TARGET_64BIT.
15945 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15947 * config/epiphany/epiphany.md (GPR_1): New constant.
15948 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
15949 * config/epiphany/epiphany.c (gen_compare_reg):
15950 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
15951 is already in place.
15952 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
15953 Don't require being called during rtl expansion; If y operlaps r0,
15955 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
15956 (epiphany_expand_epilogue): Likewise.
15958 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
15959 Don't use CC_FPmode for ORDERED / UNORDERED.
15960 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
15962 * config/epiphany/constraints.md (CnL): New constraint.
15963 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
15964 * config/epiphany/predicates.md (add_operand): Allow 1024.
15966 * config/epiphany/epiphany.md (logical_op): New code iterator.
15967 (op_mnc): New code attribute.
15968 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
15969 (mov_f+1, mov_f+2): New peephole2 patterns.
15971 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
15972 (cstoresi4): Also allow re-use of zero result when doing a NE
15973 comparison to a non-zero operand.
15974 Use (clobber (scratch)) for first insn if the gpr output is not needed.
15976 * config/epiphany/epiphany.md (<insn_opname>v2si3):
15977 Use gen_addsi3_i / gen_subsi3_i.
15979 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15983 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
15984 contain anything but clobbers, at most one __builtin_stack_restore,
15985 optionally debug stmts and final resx, and if it has at least one
15986 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
15987 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
15988 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
15989 which isn't defaut definition, remove them.
15990 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
15991 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
15992 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
15993 with MEM_REF LHS with SSA_NAME address.
15995 2013-04-08 Jeff Law <law@redhat.com>
15997 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
15999 2013-04-08 Richard Biener <rguenther@suse.de>
16001 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16003 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16004 determined vector type.
16005 (vect_analyze_data_refs): Likewise.
16006 (vect_get_new_vect_var): Adjust.
16007 (vect_create_destination_var): Preserve SSA name versions.
16008 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16009 not dump anything here.
16011 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16013 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16014 Add member lr_slot_known.
16015 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16017 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16018 Remove code that sets lr_slot_offset according to what a previous
16019 version of epiphany_emit_save_restore used to do.
16020 (epiphany_emit_save_restore): When doing an lr save or restore,
16021 set/verify lr_slot_known and lr_slot_offset.
16023 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16026 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16029 2013-04-08 Richard Biener <rguenther@suse.de>
16031 * alias.c (find_base_term): Fix thinko in previous change.
16033 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16035 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16036 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16037 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16038 if possible to compute val.
16039 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16040 For QImode integers don't require anything about precision. Use
16041 const_with_all_bytes_same to find out if the constant doesn't have
16042 repeated bytes in it.
16044 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16046 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16049 2013-04-08 Marek Polacek <polacek@redhat.com>
16051 PR rtl-optimization/48182
16052 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
16055 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16058 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
16059 nuses, make sure we have a label.
16061 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16064 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
16065 (rs6000_emit_swdiv_low_precision): Remove.
16066 (rs6000_emit_swdiv): Rewrite to handle between one and four
16067 iterations of Newton-Raphson generally; modify required number of
16068 iterations for some cases.
16069 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
16071 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
16073 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
16074 set-but-unused variable.
16076 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
16077 basic blocks of released function bodies garbage-collectable.
16079 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
16080 (struct rtl_opt_pass): Add TODO_df_finish.
16082 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
16084 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16086 * config/arm/constraints.md (q): New constraint.
16087 * config/arm/ldrdstrd.md: New file.
16088 * config/arm/arm.md (ldrdstrd.md) New include.
16089 (arm_movdi): Use "q" instead of "r" constraint
16090 for double-word memory access.
16091 (movdf_soft_insn): Likewise.
16092 * config/arm/vfp.md (movdi_vfp): Likewise.
16093 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
16094 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
16095 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
16096 (mem_ok_for_ldrd_strd): Likewise.
16097 (output_move_double): Update assertion.
16099 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16101 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
16103 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16105 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
16106 define_insn_and_split.
16107 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
16109 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16111 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
16112 define_insn_and_split.
16113 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
16114 (shiftsi3_compare): New pattern.
16115 (rrx): New pattern.
16116 * config/arm/unspecs.md (UNSPEC_RRX): New.
16118 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16120 * config/arm/arm.md (negdi_extendsidi): New pattern.
16121 (negdi_zero_extendsidi): Likewise.
16123 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16125 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
16126 define_insn_and_split.
16127 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
16128 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
16130 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16132 * config/arm/arm.md (arm_subdi3): Convert define_insn into
16133 define_insn_and_split.
16134 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
16135 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
16137 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16139 * config/arm/arm.md (subsi3_carryin): New pattern.
16140 (subsi3_carryin_const): Likewise.
16141 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
16142 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
16144 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16146 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
16148 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16150 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
16151 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
16153 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16155 * config/arm/arm.c (arm_expand_builtin): Change fcode
16156 type to unsigned int.
16158 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16160 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
16162 2013-04-04 Ian Lance Taylor <iant@google.com>
16164 * doc/standards.texi (Standards): The Go frontend supports the Go 1
16167 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
16169 PR middle-end/56729
16170 * df-scan.c (df_insn_delete): Disable failing assert.
16172 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16174 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
16175 New function prototype.
16176 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
16177 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
16178 (arm_builtin_vectorized_function): New function.
16180 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16182 * config/arm/arm_neon_builtins.def: New file.
16183 * config/arm/arm.c (neon_builtin_data): Move contents to
16184 arm_neon_builtins.def.
16185 (enum arm_builtins): Include neon builtin definitions.
16186 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
16187 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
16189 2013-04-04 Marek Polacek <polacek@redhat.com>
16191 PR tree-optimization/48186
16192 * predict.c (maybe_hot_frequency_p): Return false if
16193 HOT_BB_FREQUENCY_FRACTION is 0.
16194 (cgraph_maybe_hot_edge_p): Likewise.
16196 2013-04-04 Richard Biener <rguenther@suse.de>
16198 PR tree-optimization/56826
16199 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
16202 2013-04-04 Richard Biener <rguenther@suse.de>
16204 PR tree-optimization/56213
16205 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
16206 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
16208 2013-04-04 Richard Biener <rguenther@suse.de>
16210 PR tree-optimization/56837
16211 * tree-loop-distribution.c (classify_partition): For non-zero
16212 values require that the value has the same precision as its
16213 mode to be useful as memset value.
16215 2013-04-03 Nick Clifton <nickc@redhat.com>
16217 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
16218 (fmssf4): Use fmsf.s on E3V5 architectures.
16219 (fnmasf4): Use fnmaf.s on E3V5 architectures.
16220 (fnmssf4): Use fnmsf.s on E3V5 architectures.
16222 2013-04-03 Jeff Law <law@redhat.com>
16224 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
16225 (lra-eliminations.o): Likewise.
16227 2013-04-03 Teresa Johnson <tejohnson@google.com>
16229 * gcov-io.c (compute_working_sets): Moved most of body of old
16230 compute_working_sets here from profile.c.
16231 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
16232 (gcov_working_set_t): Moved typedef here from basic-block.h
16233 (compute_working_set): Declare.
16234 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
16235 (get_working_sets): Renamed from compute_working_set,
16236 replace most of body with call to new compute_working_sets.
16237 (get_exec_counts): Replace call to compute_working_sets
16238 to get_working_sets.
16239 * profile.h (get_working_sets): Renamed from compute_working_set.
16240 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
16241 to get_working_sets.
16242 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
16243 * gcov-dump.c (dump_working_sets): New function.
16245 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
16247 * hwint.c (sext_hwi, zext_hwi): New functions.
16248 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
16249 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
16250 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
16251 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
16252 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
16253 (sext_hwi, zext_hwi): New functions.
16255 2013-04-03 Jeff Law <law@redhat.com>
16257 PR tree-optimization/56799
16258 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
16259 back test for widening conversion erroneously dropped in prior change.
16261 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16264 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
16265 instead of next_real_insn.
16267 2013-04-03 Marek Polacek <polacek@redhat.com>
16270 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
16272 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16275 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
16277 (thumb1_output_casesi): Likewise.
16278 (thumb2_output_casesi): Likewise.
16280 2013-04-03 Richard Biener <rguenther@suse.de>
16282 PR tree-optimization/56817
16283 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
16285 (tree_unroll_loops_completely_1): ... new function to manually
16286 walk the loop tree, properly defering outer loops of unrolled
16287 loops to later iterations.
16289 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16291 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
16292 (vectorizable_load): Likewise.
16293 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
16294 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
16296 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16298 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
16301 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16303 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
16305 2013-04-03 Bin Cheng <bin.cheng@arm.com>
16307 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
16309 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16311 PR tree-optimization/56790
16312 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
16315 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16317 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
16319 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
16320 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
16323 2013-04-03 Jakub Jelinek <jakub@redhat.com>
16326 * tree.h (force_folding_builtin_constant_p): New decl.
16327 * builtins.c (force_folding_builtin_constant_p): New variable.
16328 (fold_builtin_constant_p): Fold immediately also if
16329 force_folding_builtin_constant_p.
16331 2013-04-03 Richard Biener <rguenther@suse.de>
16333 PR tree-optimization/56812
16334 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
16335 DRs of the same interleaving chain are independent.
16337 2013-04-02 Jason Merrill <jason@redhat.com>
16339 * gdbinit.in (pbb): Use debug fn.
16341 2013-04-02 Lawrence Crowl <crowl@google.com>
16343 * sese.h (struct ivtype_map_elt_s): Remove unused.
16344 (extern debug_ivtype_map): Remove unused.
16345 (extern eq_ivtype_map_elts): Remove unused.
16346 * sese.c (debug_ivtype_map): Removed unused.
16347 (debug_ivtype_map_1): Removed unused.
16348 (debug_ivtype_elt): Remove unused.
16349 (eq_ivtype_map_elts): Remove unused.
16352 2013-04-02 Kai Tietz <ktietz@redhat.com>
16355 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
16356 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
16357 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
16359 (legitimize_pe_coff_symbol): Likewise.
16360 (is_imported_p): New helper-function.
16361 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
16362 for Windows x64 targets.
16363 (ix86_expand_prologue): Optimize for pe-coff targets.
16364 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
16365 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
16366 medium/large code-model.
16367 (legitimize_pic_address): Likewise.
16368 (legitimize_tls_address): Likewise.
16369 (ix86_expand_call): Likewise.
16370 (x86_output_mi_thunk): Likewise.
16371 (get_dllimport_decl): Add new beimport argument.
16372 (construct_plt_address): Don't assert for x64 pe-coff targets.
16373 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
16375 (SYMBOL_FLAG_STUBVAR): New macro.
16376 (SYMBOL_REF_STUBVAR_P): Likewise.
16377 * config/i386/winnt.c (stub_list): New structure.
16378 (stub_head): New local variable.
16379 (i386_pe_record_stub): New function.
16380 (i386_pe_file_end): Emit refptr-stubs.
16382 2013-04-02 Jakub Jelinek <jakub@redhat.com>
16384 PR rtl-optimization/56745
16385 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
16386 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
16389 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
16390 and both of them are MEM_REFs, just compare first argument for
16391 equality and attempt to deal even with differing offsets.
16394 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
16395 of gimple_clobber_p to be MEM_REF.
16396 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
16397 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
16398 after gimplification.
16399 * asan.c (get_mem_ref_of_assignment): Don't instrument
16400 gimple_clobber_p stmts.
16401 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
16402 gimple_clobber_p stmt if they have MEM_REF lhs and
16403 are dead because of another gimple_clobber_p stmt.
16404 * tree-ssa-live.c (clear_unused_block_pointer): Treat
16405 gimple_clobber_p stmts like debug stmts.
16406 (remove_unused_locals): Remove clobbers with MEM_REF lhs
16407 that refer to unused VAR_DECLs or uninitialized values.
16408 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
16409 gimple_clobber_p stmts if they refer to removed parameters.
16410 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
16413 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
16415 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
16416 using SWI48 mode attribute.
16418 2013-04-02 Wei Mi <wmi@google.com>
16420 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
16421 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
16422 *<rotate_insn><mode>3_mask in i386.md.
16424 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
16426 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
16428 2013-04-02 Richard Biener <rguenther@suse.de>
16430 PR tree-optimization/56778
16431 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
16432 Runtime alias tests are not supported for gather loads.
16433 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
16434 stmts referenced from SSA operands before updating SSA form.
16436 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
16437 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16439 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
16440 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
16441 * config/arm/cortex-a53.md: New file.
16442 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
16443 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
16444 * config/arm/arm.c (arm_issue_rate): Likewise.
16445 * config/arm/arm-tune.md: Regenerate
16446 * config/arm/arm-tables.opt: Regenerate.
16447 * config/arm/arm-cores.def: Add cortex-a53.
16449 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
16451 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
16454 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
16456 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
16457 scalar load/store operations using B/H registers.
16458 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
16460 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
16462 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
16464 * config/aarch64/aarch64.c
16465 (aarch64_simd_scalar_immediate_valid_for_move): New.
16466 * config/aarch64/aarch64-protos.h
16467 (aarch64_simd_scalar_immediate_valid_for_move): New.
16468 * config/aarch64/constraints.md (Dh, Dq): New.
16469 * config/aarch64/iterators.md (hq): New.
16471 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
16473 * reorg.c (get_branch_condition): Deal with conditional returns.
16474 (fill_simple_delay_slots): Remove dead code dealing with jumps.
16476 2013-04-01 Wei Mi <wmi@google.com>
16478 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
16479 Truncate operand 2 using %b asm operand modifier.
16480 (*<shift_insn><mode>3_mask): Ditto.
16481 (*<rotate_insn><mode>3_mask): Ditto.
16483 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
16485 PR middle-end/56798
16486 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
16488 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
16490 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
16492 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
16494 2013-03-30 Lawrence Crowl <crowl@google.com>
16496 * dse.c (clear_alias_sets): Remove never set.
16497 (disqualified_clear_alias_sets): Remove never set.
16498 (clear_alias_mode_pool): Remove never set.
16499 (dse_step0): Remove condition that is never true.
16500 (canon_address): Remove condition that is never true.
16501 (dse_step7): Remove condition that is never true.
16502 (rest_of_handle_dse): Remove condition that is never true.
16503 (rest_of_handle_dse::did_global): Remove never read from above.
16504 (dse_step2_spill): Remove never called from above.
16505 (dse_step5_spill): Remove never called from above.
16507 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
16509 * doc/md.texi (Standard Names) <casesi>: Update documentation for
16510 JUMP_TABLE_DATA changes.
16511 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
16512 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
16513 (Insns) <jump_table_data>: New entry.
16514 * doc/tm.texi: Regenerate.
16516 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
16518 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
16519 for table jump at the end of a basic block using tablejump_p.
16520 * targhooks.c (default_invalid_within_doloop): Likewise.
16521 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
16522 target hook implementation that is identical to the default hook.
16523 (rs6000_invalid_within_doloop): Remove.
16525 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
16526 unused variable from tablejump_p call.
16528 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
16529 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
16530 (INSN_DELETED_P): Likewise.
16531 (emit_jump_table_data): New prototype.
16532 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
16533 after 4th as unused.
16534 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
16535 * sched-vis.c (print_insn): Likewise.
16536 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
16537 insn for compatibility with back ends that use next_active_insn to
16538 identify jump table data.
16539 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
16540 (remove_insn): Likewise.
16541 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
16543 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
16544 (emit_jump_table_data): New function.
16546 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
16547 basic block, a JUMP_TABLE_DATA never is.
16548 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
16549 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
16550 off from code handling real insns.
16551 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
16552 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
16553 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
16554 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
16555 is not a NONDEBUG_INSN_P.
16556 * ira-costs.c (scan_one_insn): Likewise.
16557 * jump.c (mark_all_labels): Likewise.
16558 (mark_jump_label_1): Likewise.
16559 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
16560 * lra.c (get_insn_freq): Expect all insns reaching here to be in
16562 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
16563 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
16564 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
16565 JUMP_TABLE_DATA_P insns.
16566 (calculate_elim_costs_all_insns): Likewise.
16567 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
16568 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
16569 (delete_output_reload): Code style fixups.
16570 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
16571 insn flags on this non-insn.
16572 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
16573 as scheduling barriers, for pre-change compatibility.
16574 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
16575 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
16577 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
16578 redundant JUMP_TABLE_DATA_P test.
16579 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
16580 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
16581 (frv_for_each_packet): Likewise.
16582 * config/i386/i386.c (min_insn_size): Likewise.
16583 (ix86_avoid_jump_mispredicts): Likewise.
16584 * config/m32r/m32r.c (m32r_is_insn): Likewise.
16585 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
16586 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
16587 (mips16_insn_length): Robustify.
16588 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
16589 (mips16_split_long_branches): Likewise.
16590 * config/pa/pa.c (pa_combine_instructions): Likewise.
16591 * config/rs6000/rs6000.c (get_next_active_insn): Treat
16592 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
16593 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
16594 as contributing to pool range lengths.
16595 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
16596 Remove redundant JUMP_TABLE_DATA_P test.
16597 (sh_loop_align): Likewise.
16598 (split_branches): Likewise.
16599 (sh_insn_length_adjustment): Likewise.
16600 * config/spu/spu.c (get_branch_target): Likewise.
16602 2013-03-29 Jan Hubicka <jh@suse.cz>
16604 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
16605 gcov streaming; stream hot bb threshold to ltrans.
16606 * predict.c (get_hot_bb_threshold): Break out from ....
16607 (maybe_hot_count_p): ... here.
16608 (set_hot_bb_threshold): New function.
16609 * lto-section-in.c (lto_section_name): Add profile.
16610 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
16611 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
16612 and data-streamer.h
16613 (histogram_entry): New structure.
16614 (histogram, histogram_pool): New global vars.
16615 (histogram_hash): New structure.
16616 (histogram_hash::hash): New method.
16617 (histogram_hash::equal): Likewise.
16618 (account_time_size): New function.
16619 (cmp_counts): New function.
16620 (dump_histogram): New function.
16621 (ipa_profile_generate_summary): New function.
16622 (ipa_profile_write_summary): New function.
16623 (ipa_profile_read_summary): New function.
16624 (ipa_profile): Decide on threshold.
16625 (pass_ipa_profile): Add ipa_profile_write_summary and
16626 ipa_profile_read_summary.
16627 * Makefile.in (ipa.o): Update dependencies.
16628 * lto-streamer.h (LTO_section_ipa_profile): New section.
16630 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
16632 * tree.h (VAR_P): New.
16634 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
16637 * doc/invoke.texi ([-fwhole-program]): Fix typo.
16639 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
16641 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
16642 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
16643 (control_flow_insn_p): Likewise.
16644 * cfgrtl.c (duplicate_insn_chain): Likewise.
16645 * final.c (get_attr_length_1): Likewise.
16646 (shorten_branches): Likewise.
16647 (final_scan_insn): Likewise.
16648 * function.c (instantiate_virtual_regs): Likewise.
16649 * gcse.c (insert_insn_end_basic_block): Likewise.
16650 * ira-costs.c (scan_one_insn): Likewise.
16651 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
16652 * lra.c (check_rtl): Likewise.
16653 * reload1.c (elimination_costs_in_insn): Likewise.
16654 * reorg.c (follow_jumps): Likewise.
16656 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
16657 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
16658 (thumb_far_jump_used_p): Likewise.
16659 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
16660 (workaround_speculation): Likewise.
16661 (add_sched_insns_for_speculation): Likewise.
16662 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
16663 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
16664 (frv_for_each_packet): Likewise.
16665 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
16666 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
16667 (final_emit_insn_group_barriers): Likewise.
16668 * config/m32r/m32r.c (m32r_is_insn): Likewise.
16669 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
16670 (mips16_insn_length): Likewise.
16671 * config/pa/pa.c (pa_reorg): Likewise.
16672 (pa_combine_instructions): Likewise.
16673 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
16674 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
16675 (sh_reorg): Likewise.
16676 (split_branches): Likewise.
16677 * config/spu/spu.c (get_branch_target): Likewise.
16679 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
16682 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
16684 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
16685 Fix declaration name.
16687 2013-03-28 Lawrence Crowl <crowl@google.com>
16689 * graphds.h (struct graph.indicies): Remove unused.
16690 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
16691 (SCOP_ORIGINAL_PDDRS): Remove unused.
16692 * sese.h (extern insert_loop_close_phis): Removed unused.
16693 (extern insert_guard_phis): Removed unused.
16694 (extern ivtype_map_elt_info): Removed unused.
16695 (new_ivtype_map_elt): Removed unused.
16696 * sese.c (ivtype_map_elt_info): Removed unused.
16698 2013-03-28 Lawrence Crowl <crowl@google.com>
16700 * Makefile.in: Add several missing include dependences.
16702 (test-dump.o): New. This object is not added to any executable,
16703 but is present for ad-hoc testing.
16705 (debug (const bitmap_head_def &)): New.
16706 (debug (const bitmap_head_def *)): New.
16708 (extern debug (const bitmap_head_def &)): New.
16709 (extern debug (const bitmap_head_def *)): New.
16711 (debug (edge_def &)): New.
16712 (debug (edge_def *)): New.
16714 (debug (basic_block_def &)): New.
16715 (debug (basic_block_def *)): New.
16717 (dump_node (const_tree, int, FILE *)): Correct source file.
16719 (debug (die_struct &)): New.
16720 (debug (die_struct *)): New.
16722 (extern debug (die_struct &)): New.
16723 (extern debug (die_struct *)): New.
16724 * gimple-pretty-print.c
16725 (debug (gimple_statement_d &)): New.
16726 (debug (gimple_statement_d *)): New.
16727 * gimple-pretty-print.h
16728 (extern debug (gimple_statement_d &)): New.
16729 (extern debug (gimple_statement_d *)): New.
16731 (debug (ira_allocno_copy &)): New.
16732 (debug (ira_allocno_copy *)): New.
16733 (debug (ira_allocno &)): New.
16734 (debug (ira_allocno *)): New.
16736 (extern debug (ira_allocno_copy &)): New.
16737 (extern debug (ira_allocno_copy *)): New.
16738 (extern debug (ira_allocno &)): New.
16739 (extern debug (ira_allocno *)): New.
16741 (debug (live_range &)): New.
16742 (debug (live_range *)): New.
16744 (debug (lra_live_range &)): New.
16745 (debug (lra_live_range *)): New.
16747 (debug (lra_live_range &)): New.
16748 (debug (lra_live_range *)): New.
16750 (debug (omega_pb_d &)): New.
16751 (debug (omega_pb_d *)): New.
16753 (extern debug (omega_pb_d &)): New.
16754 (extern debug (omega_pb_d *)): New.
16756 (debug (const rtx_def &)): New.
16757 (debug (const rtx_def *)): New.
16759 (debug_tree (tree): Move within file.
16760 (debug_raw (const tree_node &)): New.
16761 (debug_raw (const tree_node *)): New.
16762 (dump_tree_via_hooks (const tree_node *, int)): New.
16763 (debug (const tree_node &)): New.
16764 (debug (const tree_node *)): New.
16765 (debug_verbose (const tree_node &)): New.
16766 (debug_verbose (const tree_node *)): New.
16767 (debug_head (const tree_node &)): New.
16768 (debug_head (const tree_node *)): New.
16769 (debug_body (const tree_node &)): New.
16770 (debug_body (const tree_node *)): New.
16771 (debug_vec_tree (tree): Move and reimplement in terms of dump.
16772 (debug (vec<tree, va_gc> &)): New.
16773 (debug (vec<tree, va_gc> *)): New.
16775 (extern debug (const rtx_def &)): New.
16776 (extern debug (const rtx_def *)): New.
16778 (debug_raw (simple_bitmap_def &)): New.
16779 (debug_raw (simple_bitmap_def *)): New.
16780 (debug (simple_bitmap_def &)): New.
16781 (debug (simple_bitmap_def *)): New.
16783 (extern debug (simple_bitmap_def &)): New.
16784 (extern debug (simple_bitmap_def *)): New.
16785 (extern debug_raw (simple_bitmap_def &)): New.
16786 (extern debug_raw (simple_bitmap_def *)): New.
16788 (debug (vinsn_def &)): New.
16789 (debug (vinsn_def *)): New.
16790 (debug_verbose (vinsn_def &)): New.
16791 (debug_verbose (vinsn_def *)): New.
16792 (debug (expr_def &)): New.
16793 (debug (expr_def *)): New.
16794 (debug_verbose (expr_def &)): New.
16795 (debug_verbose (expr_def *)): New.
16796 (debug (vec<rtx> &)): New.
16797 (debug (vec<rtx> *)): New.
16799 (extern debug (vinsn_def &)): New.
16800 (extern debug (vinsn_def *)): New.
16801 (extern debug_verbose (vinsn_def &)): New.
16802 (extern debug_verbose (vinsn_def *)): New.
16803 (extern debug (expr_def &)): New.
16804 (extern debug (expr_def *)): New.
16805 (extern debug_verbose (expr_def &)): New.
16806 (extern debug_verbose (expr_def *)): New.
16807 (extern debug (vec<rtx> &)): New.
16808 (extern debug (vec<rtx> *)): New.
16810 (_list_iter_cond_expr): Make inline instead of static.
16812 (debug (sreal &)): New.
16813 (debug (sreal *)): New.
16815 (extern debug (sreal &)): New.
16816 (extern debug (sreal *)): New.
16818 (extern debug_raw (const tree_node &)): New.
16819 (extern debug_raw (const tree_node *)): New.
16820 (extern debug (const tree_node &)): New.
16821 (extern debug (const tree_node *)): New.
16822 (extern debug_verbose (const tree_node &)): New.
16823 (extern debug_verbose (const tree_node *)): New.
16824 (extern debug_head (const tree_node &)): New.
16825 (extern debug_head (const tree_node *)): New.
16826 (extern debug_body (const tree_node &)): New.
16827 (extern debug_body (const tree_node *)): New.
16828 (extern debug (vec<tree, va_gc> &)): New.
16829 (extern debug (vec<tree, va_gc> *)): New.
16831 (debug (struct loop &)): New.
16832 (debug (struct loop *)): New.
16833 (debug_verbose (struct loop &)): New.
16834 (debug_verbose (struct loop *)): New.
16835 * tree-dump.c: Add header dependence.
16837 (extern debug (struct loop &)): New.
16838 (extern debug (struct loop *)): New.
16839 (extern debug_verbose (struct loop &)): New.
16840 (extern debug_verbose (struct loop *)): New.
16842 (debug (data_reference &)): New.
16843 (debug (data_reference *)): New.
16844 (debug (vec<data_reference_p> &)): New.
16845 (debug (vec<data_reference_p> *)): New.
16846 (debug (vec<ddr_p> &)): New.
16847 (debug (vec<ddr_p> *)): New.
16849 (extern debug (data_reference &)): New.
16850 (extern debug (data_reference *)): New.
16851 (extern debug (vec<data_reference_p> &)): New.
16852 (extern debug (vec<data_reference_p> *)): New.
16853 (extern debug (vec<ddr_p> &)): New.
16854 (extern debug (vec<ddr_p> *)): New.
16856 (debug (pt_solution &)): New.
16857 (debug (pt_solution *)): New.
16859 (extern debug (pt_solution &)): New.
16860 (extern debug (pt_solution *)): New.
16862 (debug (_var_map &)): New.
16863 (debug (_var_map *)): New.
16864 (debug (tree_live_info_d &)): New.
16865 (debug (tree_live_info_d *)): New.
16867 (extern debug (_var_map &)): New.
16868 (extern debug (_var_map *)): New.
16869 (extern debug (tree_live_info_d &)): New.
16870 (extern debug (tree_live_info_d *)): New.
16872 2013-03-28 Jan Hubicka <jh@suse.cz>
16874 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
16876 2013-03-28 Ian Bolton <ian.bolton@arm.com>
16878 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
16879 record only when desired or required.
16881 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
16883 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
16884 *vec_extractv2di_1_rex64. Use x64 isa attribute.
16886 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16888 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
16889 (*andsi3_compare0_uxtw): New pattern.
16890 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
16891 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
16893 2013-03-28 Jan Hubicka <jh@suse.cz>
16895 * data-streamer-in.c (streamer_read_gcov_count): New function.
16896 * gimple-streamer-out.c: Include value-prof.h.
16897 (output_gimple_stmt): Output histogram.
16898 (output_bb): Use streamer_write_gcov_count.
16899 * value-prof.c: Include data-streamer.h
16900 (dump_histogram_value): Add HIST_TYPE_MAX.
16901 (stream_out_histogram_value): New function.
16902 (stream_in_histogram_value): New function.
16903 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
16904 (stream_out_histogram_value, stream_in_histogram_value): Declare.
16905 * data-streamer-out.c (streamer_write_gcov_count): New function.
16906 (streamer_write_gcov_count_stream): New function.
16907 * lto-cgraph.c (lto_output_edge): Update counter streaming.
16908 (lto_output_node): Likewise.
16909 (input_node, input_edge): Likewise.
16910 * lto-streamer-out.c (output_cfg): Update streaming.
16911 * lto-streamer-in.c (input_cfg): Likewise.
16912 * data-streamer.h (streamer_write_gcov_count,
16913 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
16914 * gimple-streamer-in.c: Include value-prof.h
16915 (input_gimple_stmt): Input histograms.
16916 (input_bb): Update profile streaming.
16918 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16920 * genmodes.c (emit_max_int): New function.
16921 (emit_insn_modes_h): Added call to emit_max_function.
16922 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
16924 * machmode.def: Fixed comment.
16926 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16928 * combine.c (try_combine): Removed useless assert.
16929 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
16931 2013-03-28 Marek Polacek <polacek@redhat.com>
16932 Richard Biener <rguenther@suse.de>
16934 PR tree-optimization/56695
16935 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
16936 build signed result of a vector comparison.
16937 * tree-cfg.c (verify_gimple_comparison): Check that a result
16938 of a vector comparison has signed type.
16940 2013-03-28 Richard Biener <rguenther@suse.de>
16942 PR tree-optimization/37021
16943 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
16944 do not restrict gaps between groups.
16945 * tree-vect-stmts.c (vectorizable_load): Properly account for
16946 a gap between groups.
16948 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
16950 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
16951 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
16954 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
16956 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
16957 * doc/extend.texi (Named Address Spaces): Ditto.
16958 (Variable Attributes): Ditto.
16960 2013-03-27 Kai Tietz <ktietz@redhat.com>
16962 * config.build: Add support for cygwin x64 target.
16963 * config.gcc: Likewise.
16964 * config.host: Likewise.
16965 * configure.ac: Likewise
16966 * configure: Regenerated.
16968 2013-03-27 Kai Tietz <ktietz@redhat.com>
16970 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
16971 * config/i386/t-cygwin-w64: New file.
16972 * config/i386/cygwin-w64.h: New file.
16973 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
16974 and add support for x64-cygwin target.
16975 (CPP_SPEC): Likewise.
16976 (CXX_WRAP_SPEC_LIST): Undefine before define.
16977 (LIBGCJ_SONAME): Use 15 as version.
16979 2013-03-27 Richard Biener <rguenther@suse.de>
16981 PR tree-optimization/56716
16982 * tree-ssa-structalias.c (perform_var_substitution): Adjust
16983 dumping for ref nodes.
16985 2013-03-27 Martin Jambor <mjambor@suse.cz>
16987 PR tree-optimization/55334
16988 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
16989 restricted pointers to arrays.
16991 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
16993 * Makefile.in (.SUFFIXES): Add .cc.
16994 (.c.o): Apply same recipe for implicit rule .cc.o.
16996 2013-03-27 Richard Biener <rguenther@suse.de>
16998 PR tree-optimization/37021
16999 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17000 REALPART/IMAGPART_EXPRs around the supported refs.
17001 * tree-ssa-structalias.c (find_func_aliases): Assume that
17002 floating-point values are not used to transfer pointers.
17004 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17006 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17007 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17008 * doc/tm.texi: Regenerate.
17009 * targhooks.h (default_has_ifunc_p): New.
17010 * targhooks.c (default_has_ifunc_p): Ditto.
17011 * config/linux-protos.h: New file.
17012 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17013 hook for linux which disables support of indirect functions in android.
17014 * config/linux-android.c: New file.
17015 * config/t-linux-android.c: Ditto.
17016 * config.gcc: Added new object file linux-android.o.
17017 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17018 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17019 * varasm.c (do_assemble_alias): Likewise.
17020 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17021 doesn't support indirect functions.
17022 * configure: Regenerate.
17024 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17027 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17028 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17031 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17033 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17035 2013-03-27 Terry Guo <terry.guo@arm.com>
17037 * config/arm/arm-cores.def: Added core cortex-r7.
17038 * config/arm/arm-tune.md: Regenerated.
17039 * config/arm/arm-tables.opt: Regenerated.
17040 * doc/invoke.texi: Added entry for core cortex-r7.
17042 2013-03-27 Walter Lee <walt@tilera.com>
17044 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17045 double-decrement of next_scratch_regno.
17047 2013-03-27 Walter Lee <walt@tilera.com>
17049 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17051 (insn_v1mulus): Ditto.
17052 (insn_v2muls): Ditto.
17054 2013-03-27 Walter Lee <walt@tilera.com>
17056 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
17057 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
17059 2013-03-27 Walter Lee <walt@tilera.com>
17061 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
17062 (*sibcall_value): Ditto.
17064 2013-03-27 Walter Lee <walt@tilera.com>
17066 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
17067 (insn_mnz_v8qi): ... this ...
17068 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
17070 (insn_v<n>mnz): Replaced by ...
17071 (insn_v1mnz): ... this ...
17072 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
17074 (insn_mz_<mode>): Replaced by ...
17075 (insn_mz_v8qi): ... this ...
17076 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
17078 (insn_v<n>mz): Replaced by ...
17079 (insn_v1mz): ... this ...
17080 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
17083 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
17085 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
17087 2013-03-26 Roland McGrath <mcgrathr@google.com>
17089 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
17090 than fprintf with a non-constant, non-format string.
17092 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17094 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
17095 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
17096 operand 0 predicate.
17097 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
17098 attribute. Use general_x64nomem_operand as operand 1 predicate.
17099 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
17100 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17101 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
17102 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17103 (mov<mode>_insv_1): Remove expander. Merge insn with
17104 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
17105 Use general_x64nomem_operand as operand 1 predicate.
17106 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
17107 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
17108 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17109 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
17110 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
17111 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
17112 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17113 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
17114 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
17115 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
17116 (general_x64nomem_operand): Ditto.
17118 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
17120 * config/rtems.opt: Add -pthread option.
17122 2013-03-26 Richard Biener <rguenther@suse.de>
17124 * alias.c (find_base_term): Avoid redundant and not used recursion.
17125 (base_alias_check): Get the initial base term from the caller.
17126 (true_dependence_1): Compute and pass base terms to base_alias_check.
17127 (write_dependence_p): Likewise.
17128 (may_alias_p): Likewise.
17130 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
17132 * config/aarch64/aarch64.c (aarch64_classify_address): Support
17133 PC-relative load in SI modes and above only.
17135 2013-03-26 Xinyu Qi <xyqi@marvell.com>
17137 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
17138 * config/arm/iwmmxt.md (WCGR0): Update.
17139 (WCGR1, WCGR2, WCGR3): Likewise.
17141 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17143 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
17144 Use x64 and nox64 isa attributes.
17146 2013-03-26 Richard Biener <rguenther@suse.de>
17148 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
17149 alignment computations and rely on get_object_alignment_1
17150 for the !TYPE_P case.
17151 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
17153 2013-03-26 Walter Lee <walt@tilera.com>
17155 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
17156 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
17158 2013-03-25 Jeff Law <law@redhat.com>
17160 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
17161 check for INTEGRAL_TYPE_P that was missing due to checking in
17162 wrong version of prior patch.
17164 2013-03-25 Walter Lee <walt@tilera.com>
17166 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
17167 TILEGX_INSN_SHUFFLEBYTES1.
17168 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
17170 (tilegx_builtins): Ditto.
17171 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
17173 2013-03-25 Walter Lee <walt@tilera.com>
17175 * config/tilegx/tilegx.md (floatsisf2): New pattern.
17176 (floatunssisf2): New pattern.
17177 (floatsidf2): New pattern.
17178 (floatunssidf2): New pattern.
17180 2013-03-25 Walter Lee <walt@tilera.com>
17182 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
17183 tests for constraint J, K, N, P.
17185 2013-03-25 Walter Lee <walt@tilera.com>
17187 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
17188 Use indirect/pcrel encoding.
17189 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
17192 2013-03-25 Steve Ellcey <sellcey@mips.com>
17194 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
17195 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
17196 * config/mips/mips.c (mips_option_override): Set IMADD default.
17197 * config/mips/mips.h (PTF_AVOID_IMADD): New.
17198 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
17199 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
17200 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
17201 * doc/invoke.texi (-mimadd/-mno-imadd): New.
17203 2013-03-25 Jeff Law <law@redhat.com>
17205 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
17206 slightly to avoid creating and folding useless trees. Simplify
17207 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
17209 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
17211 * config/i386/i386.md (*zero_extendsidi2): Merge with
17212 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
17213 * config/i386/predicates.md (x86_64_zext_operand): Rename from
17214 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
17215 targets. Clarify comment.
17217 2013-03-25 Martin Jambor <mjambor@suse.cz>
17219 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
17220 pass-through jump functions differently.
17221 (ipa_read_jump_function): Likewise. Also use setter functions to set
17224 2013-03-25 Martin Jambor <mjambor@suse.cz>
17226 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
17227 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
17229 (ipa_get_indirect_edge_target): New function.
17230 (devirtualization_time_bonus): New parameter known_aggs, pass it to
17231 ipa_get_indirect_edge_target. Update all callers.
17232 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
17233 ipa_get_indirect_edge_target_1 instead of calling
17234 ipa_get_indirect_edge_target.
17235 (create_specialized_node): Pass aggvlas to
17236 ipcp_discover_new_direct_edges.
17238 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17240 * config/arm/arm.md (f_sels, f_seld): New types.
17241 (*cmov<mode>): New pattern.
17242 * config/arm/predicates.md (arm_vsel_comparison_operator): New
17245 2013-03-25 Kai Tietz <ktietz@redhat.com>
17247 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
17248 POSIX-printf for mingw-hosted builds.
17250 2013-03-25 Richard Biener <rguenther@suse.de>
17252 PR middle-end/56694
17253 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
17254 must-not-throw stmt location.
17256 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17258 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
17259 Emit load-acquire versions when acq is true.
17260 (arm_emit_store_exclusive): Add rel parameter.
17261 Emit store-release versions when rel is true.
17262 (arm_split_compare_and_swap): Use acquire-release instructions
17264 of barriers when appropriate.
17265 (arm_split_atomic_op): Likewise.
17266 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
17267 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
17268 (VUNSPEC_SLX): Likewise.
17269 (VUNSPEC_LDA): Likewise.
17270 (VUNSPEC_STL): Likewise.
17271 * config/arm/sync.md (atomic_load<mode>): New pattern.
17272 (atomic_store<mode>): Likewise.
17273 (arm_load_acquire_exclusive<mode>): Likewise.
17274 (arm_load_acquire_exclusivesi): Likewise.
17275 (arm_load_acquire_exclusivedi): Likewise.
17276 (arm_store_release_exclusive<mode>): Likewise.
17278 2013-03-25 Catherine Moore <clm@codesourcery.com>
17280 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
17281 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
17282 * config/mip/predicates.md (lwsp_swsp_operand,
17283 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
17284 sb16_operand, db4_operand, db7_operand, ib3_operand,
17285 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
17286 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
17287 andi16_operand): New predicates.
17288 * config/mips/mips.md (compression): New attribute.
17289 (enabled): New attribute.
17290 (length): Consider compression in computing length.
17291 (shift_compression): New code attribute.
17292 (*add<mode>3): New operands. Record compression.
17293 (sub<mode>3): Likewise.
17294 (one_cmpl<mode>2): Likewise.
17295 (*and<mode>3): Likewise.
17296 (*ior<mode>3): Likewise.
17297 (unnamed pattern for xor): Likewise.
17298 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
17299 (*<optab><mode>3): Likewise.
17300 (*mov<mode>_internal: Likewise.
17301 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
17302 (mips_unsigned_immediate_p): New.
17303 (umips_lwsp_swsp_address_p): New.
17304 (m16_based_address_p): New.
17305 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
17306 (mips_unsigned_immediate_p): New prototype.
17307 (lwsp_swsp_address_p): New prototype.
17308 (m16_based_address_p): New prototype.
17309 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
17310 (mips_signed_immediate_p): New function.
17311 (m16_based_address_p): New function.
17312 (lwsp_swsp_address_p): New function.
17313 (mips_print_operand_punctuation): Recognize short delay slot insns
17314 for microMIPS.add<mode>3"
17316 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17319 * config/arm/iterators.md (v_cmp_result): New mode attribute.
17320 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
17322 2013-03-25 Richard Biener <rguenther@suse.de>
17324 PR tree-optimization/56689
17325 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
17328 2013-03-25 Richard Biener <rguenther@suse.de>
17330 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
17332 (memory_references): Likewise.
17333 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
17334 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
17335 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
17336 (gather_mem_refs_in_loops): Fold into ...
17337 (analyze_memory_references): ... this. Move initialization
17338 to tree_ssa_lim_initialize.
17339 (fill_always_executed_in): Rename to ...
17340 (fill_always_executed_in_1): ... this.
17341 (fill_always_executed_in): Move contains_call computation to
17342 this new function from ...
17343 (tree_ssa_lim_initialize): ... here.
17344 (tree_ssa_lim): Call fill_always_executed_in.
17346 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17348 * postreload.c (reload_combine): Fix code detecting returns.
17350 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17352 * function.c (emit_use_return_register_into_block): On cc0 targets,
17353 do not emit the sequence between cc0 setter and user.
17355 2013-03-25 Kai Tietz <ktietz@redhat.com>
17357 * config/i386/predicates.md (local_symbolic_operand): Interpret
17358 dll-imported symbols as none-local.
17360 2013-03-25 Richard Biener <rguenther@suse.de>
17362 * tree-ssa-loop-im.c (struct depend): Remove.
17363 (struct lim_aux_data): Make depends a vec of gimples.
17364 (free_lim_aux_data): Adjust.
17365 (add_dependency): Likewise.
17366 (set_level): Likewise.
17368 2013-03-25 Richard Biener <rguenther@suse.de>
17370 PR middle-end/56434
17371 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
17372 the pointer returned by calls with ECF_MALLOC set.
17374 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
17376 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
17378 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
17380 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
17381 using MMXMODE mode iterator.
17382 (*move<mode>_internal): Merge with *movv2sf_internal and
17383 *movv2sf_internal_rex64 using MMXMODE mode iterator.
17385 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
17387 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
17388 (record_last_mem_set_info): Likewise.
17390 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
17391 of XNEWVEC followed by memset.
17392 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
17394 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
17396 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
17397 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
17398 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
17399 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
17400 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
17401 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
17402 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
17403 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
17404 BARRIER_P instead of GET_CODE.
17406 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17408 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
17409 inaccuracy in the probing code.
17411 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
17412 (ctrapdi4): Likewise.
17414 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17416 * calls.c (expand_call): Add missing guard to code handling return
17417 of non-BLKmode structures in MSB.
17418 * function.c (expand_function_end): Likewise.
17420 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17422 * combine.c (try_combine): Adjust comment. Do not add the set of
17423 insn #0 if the destination indirectly is set or dies in insn #2.
17424 Tidy up code to distribute a new note.
17426 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
17428 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
17429 also for alternatives 16 and 17.
17431 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
17433 * config/i386/sse.md (*mov<mode>_internal): Merge with
17434 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
17435 Emit insn template depending on type attribute. Use
17436 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
17437 movd instead of movq mnemonic for interunit moves. Rewrite mode
17438 attribute calculation. Remove unit attribute calculation.
17439 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
17440 Set prefix_data16 attribute for DImode ssemov types.
17441 Use Ym instead of y for SSE-MMX conversion alternatives.
17442 Reorder operand constraints.
17444 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
17446 * df.h (df_insn_delete): Adjust prototype.
17447 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
17448 and let it decide whether mark the basic block dirty.
17449 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
17450 * df-scan.c (df_insn_info_delete): New helper function, split
17451 off from df_insn_delete.
17452 (df_scan_free_bb_info): Use it.
17453 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
17455 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
17456 that the insn is actually an insn and it has a non-NULL basic block.
17457 Do not mark basic block dirty if only deleting a DEBUG_INSN.
17459 2013-03-22 Richard Biener <rguenther@suse.de>
17461 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
17463 (mem_ref_alloc): Do not allocate them.
17464 (refs_independent_p): Do not query or maintain a cache.
17466 2013-03-22 Richard Biener <rguenther@suse.de>
17468 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
17469 (gather_mem_refs_in_loops): Do not compute it.
17470 (analyze_memory_references): Do not allocate it.
17471 (tree_ssa_lim_finalize): Do not free it.
17472 (for_all_locs_in_loop): Do not query all_refs_in_loop.
17474 2013-03-22 Richard Biener <rguenther@suse.de>
17476 * is-a.h (as_a): Use gcc_checking_assert.
17478 2013-03-22 Ian Bolton <ian.bolton@arm.com>
17480 * config/aarch64/aarch64.c (aarch64_print_operand): New
17481 format specifier for printing a constant in hex.
17482 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
17483 format specifier for printing second operand.
17485 2013-03-22 Richard Biener <rguenther@suse.de>
17487 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
17489 (gather_mem_refs_in_loops): Perform store accumulation here.
17490 (create_vop_ref_mapping_loop): Remove.
17491 (create_vop_ref_mapping): Likewise.
17492 (analyze_memory_references): Initialize refs_stored_in_loop.
17493 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
17494 (record_indep_loop): Remove.
17495 (record_dep_loop): New function.
17496 (ref_indep_loop_p_1): Adjust to only walk over references
17497 in the loop, not its subloops.
17498 (ref_indep_loop_p): Rename to ...
17499 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
17500 maintaining a more fine-grained cache.
17501 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
17502 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
17504 2013-03-22 Richard Biener <rguenther@suse.de>
17506 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
17507 (struct mem_ref): Make accesses_in_loop a vec of a vec of
17508 aggregate mem_ref_loc.
17509 (free_mem_ref_locs): Inline into ...
17510 (memref_free): ... this and adjust.
17511 (mem_ref_alloc): Adjust.
17512 (mem_ref_locs_alloc): Remove.
17513 (record_mem_ref_loc): Adjust.
17514 (get_all_locs_in_loop): Rewrite into ...
17515 (for_all_locs_in_loop): ... this iterator.
17516 (rewrite_mem_ref_loc): New functor.
17517 (rewrite_mem_refs): Use for_all_locs_in_loop.
17518 (sm_set_flag_if_changed): New functor.
17519 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
17520 (ref_always_accessed): New functor.
17521 (ref_always_accessed_p): Use for_all_locs_in_loop.
17523 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17525 * tree-pass.h (PROP_gimple_lvec): New.
17526 * passes.c (dump_properties): Handle PROP_gimple_lvec.
17527 (init_optimization_passes): Move pass_lower_vector.
17528 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
17530 (pass_lower_vector): Provide PROP_gimple_lvec.
17531 (pass_lower_vector_ssa): Likewise.
17532 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
17534 2013-03-21 Mark Wielaard <mjw@redhat.com>
17536 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
17538 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
17540 * config/i386/i386.md (*movdi_internal): Disparage slightly
17541 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
17542 conversion alternatives.
17544 2013-03-21 Jakub Jelinek <jakub@redhat.com>
17546 PR middle-end/48087
17547 * diagnostic.def (DK_WERROR): New kind.
17548 * diagnostic.h (werrorcount): Define.
17549 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
17550 promoted to DK_ERROR, increment DK_WERROR counter instead of
17552 * toplev.c (toplev_main): Call print_ignored_options even if
17553 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
17554 even if just werrorcount is non-zero.
17557 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
17559 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
17560 (string_cst_pool_decl): New function.
17561 (optimize_one_addr_into_implicit_ptr): New function.
17562 (resolve_addr_in_expr): Optimize DWARF location expression
17563 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
17564 which doesn't live in memory, but has DW_AT_location or
17565 DW_AT_const_value, or refers to a string literal, into
17566 DW_OP_GNU_implicit_pointer.
17567 (optimize_location_into_implicit_ptr): New function.
17568 (resolve_addr): If removing DW_AT_location of a variable because
17569 it was DW_OP_addr of address of the variable, but the variable doesn't
17570 live in memory, try to emit const value attribute for the initializer.
17572 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17574 * tree.h (VECTOR_TYPE_P): New macro.
17575 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
17576 TYPE_MODE): Use it.
17577 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
17578 VEC_COND_EXPR cannot be lvalues.
17579 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
17581 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17583 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
17584 Restrict the transformation to equal modes.
17586 2013-03-21 Richard Biener <rguenther@suse.de>
17588 PR tree-optimization/39326
17589 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
17590 (MEM_ANALYZABLE): Adjust.
17591 (record_mem_ref_loc): Move bitmap ops ...
17592 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
17593 unanalyzable refs, do not record locations for it.
17594 (analyze_memory_references): Allocate ref zero as shared
17596 (refs_independent_p): Do not test for unanalyzed mems here.
17597 (ref_indep_loop_p_1): Special-case disambiguation against
17598 the unanalyzed ref.
17599 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
17601 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
17603 * config/arm/arm-protos.h (tune_params): Add
17604 prefer_neon_for_64bits field.
17605 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
17606 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
17607 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
17608 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
17609 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
17610 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
17611 (arm_option_override): Handle -mneon-for-64bits new option.
17612 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
17613 (prefer_neon_for_64bits): Declare new variable.
17614 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
17615 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
17616 (arch_enabled): Handle new arch types. Remove support for onlya8
17618 (one_cmpldi2): Use new arch names.
17619 (zero_extend<mode>di2, extend<mode>di2): Ditto.
17620 * config/arm/arm.opt (mneon-for-64bits): Add option.
17621 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
17622 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
17623 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
17625 * doc/invoke.texi (-mneon-for-64bits): Document.
17627 2013-03-21 Richard Biener <rguenther@suse.de>
17629 PR tree-optimization/39326
17630 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
17631 (sort_bbs_in_loop_postorder_cmp): New function.
17632 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
17634 2013-03-21 Richard Biener <rguenther@suse.de>
17636 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
17637 (vect_insert_into_interleaving_chain): Likewise.
17638 (vect_drs_dependent_in_basic_block): Inline ...
17639 (vect_slp_analyze_data_ref_dependence): ... here. New function,
17641 (vect_analyze_data_ref_dependence): ... here. Simplify.
17642 (vect_check_interleaving): Simplify.
17643 (vect_analyze_data_ref_dependences): Likewise. Split out ...
17644 (vect_slp_analyze_data_ref_dependences): ... this new function.
17645 (dr_group_sort_cmp): New function.
17646 (vect_analyze_data_ref_accesses): Compute data-reference groups
17647 here instead of in vect_analyze_data_ref_dependence. Use
17648 a more efficient algorithm.
17649 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
17650 vect_slp_analyze_data_ref_dependences. Call
17651 vect_analyze_data_ref_accesses earlier.
17652 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
17653 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
17654 (vect_slp_analyze_data_ref_dependences): New prototype.
17656 2013-03-21 Richard Biener <rguenther@suse.de>
17658 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
17659 ref is stored in the loop.
17660 (find_refs_for_sm): Walk only over all stores.
17661 (store_motion_loop): Allocate from lim_bitmap_obstack.
17662 (store_motion): Likewise.
17664 2013-03-21 Richard Biener <rguenther@suse.de>
17666 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
17667 Update virtual SSA form.
17669 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17671 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
17672 * configure: Regenerate.
17673 * config.in: Regenerate.
17674 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
17675 if !HAVE_LD_EH_FRAME_CIEV3.
17677 2013-03-21 Richard Biener <rguenther@suse.de>
17679 * tree-cfg.c (verify_expr_no_block): New function.
17680 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
17681 nor DECL_VALUE_EXPR have locations with associated blocks.
17682 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
17683 (clear_unused_block_pointer): Remove code dealing with
17684 blocks in DECL_DEBUG_EXPR locations.
17686 2013-03-21 Richard Biener <rguenther@suse.de>
17688 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
17689 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
17690 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
17691 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
17692 instead of DECL_DEBUG_EXPR_IS_FROM.
17693 * gimplify.c (gimplify_modify_expr): Likewise.
17694 * tree-cfg.c (verify_expr_location_1): Likewise.
17695 * tree-complex.c (create_one_component_var): Likewise.
17696 * tree-sra.c (create_access_replacement): Likewise.
17697 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
17698 (clear_unused_block_pointer): Likewise.
17699 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
17700 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
17701 * var-tracking.c (var_debug_decl): Likewise.
17702 (track_expr_p): Likewise.
17703 * tree-inline.c (add_local_variables): Likewise. Set
17704 DECL_HAS_DEBUG_EXPR_P after copying it.
17705 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
17706 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
17708 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
17711 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
17712 * configure: Regenerate.
17713 * config.in: Regenerate.
17714 * config/i386/i386.md (*movdf_internal): Use
17715 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
17716 movd instead of movq mnemonic for interunit moves.
17717 (*movdi_internal): Ditto.
17719 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17721 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
17722 (abd<mode>_3): New pattern.
17723 (aba<mode>_3): New pattern.
17724 (fabd<mode>_3): New pattern.
17726 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17728 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
17729 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
17730 occurrence of REGISTER_PREFIX as its empty string.
17732 2013-03-20 Jeff Law <law@redhat.com>
17734 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
17735 addititional equivalences for equality comparisons between an SSA_NAME
17736 and a constant where the SSA_NAME was set from a widening conversion.
17738 2013-03-20 Walter Lee <walt@tilera.com>
17740 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
17742 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17744 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
17745 depending on type attribute.
17746 (*movti_internal): Ditto.
17747 (*movtf_internal): Ditto.
17748 (*movxf_internal): Ditto.
17749 (*movdf_internal): Ditto.
17750 (*movsf_internal): Ditto.
17752 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17754 * config/i386/i386.md (*movti_internal): Set prefix attribute to
17755 maybe_vex for sselog1 and ssemov types.
17756 (*movdi_internal): Reorder operand constraints.
17757 (*movsi_internal): Ditto. Set prefix attribute to
17758 maybe_vex for sselog1 and ssemov types.
17759 (*movtf_internal): Set prefix attribute to maybe_vex
17760 for sselog1 and ssemov types.
17761 (*movdf_internal): Ditto. Set prefix_data16 attribute for
17762 DImode ssemov types. Reorder operand constraints.
17763 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
17764 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
17765 attribute for SImode ssemov types. Reorder operand constraints.
17767 2013-03-20 Martin Jambor <mjambor@suse.cz>
17769 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
17770 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
17772 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
17774 * config/rs6000/predicates.md (indexed_address, update_address_mem
17775 update_indexed_address_mem): New predicates.
17776 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
17777 attribute for load/store instructions.
17778 * config/rs6000/dfp.md (movsd_store): Likewise.
17779 (movsd_load): Likewise.
17780 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
17781 (unnamed HI->DI extend define_insn): Likewise.
17782 (unnamed SI->DI extend define_insn): Likewise.
17783 (unnamed QI->SI extend define_insn): Likewise.
17784 (unnamed QI->HI extend define_insn): Likewise.
17785 (unnamed HI->SI extend define_insn): Likewise.
17786 (unnamed HI->SI extend define_insn): Likewise.
17787 (extendsfdf2_fpr): Likewise.
17788 (movsi_internal1): Likewise.
17789 (movsi_internal1_single): Likewise.
17790 (movhi_internal): Likewise.
17791 (movqi_internal): Likewise.
17792 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
17793 attribute for load/store instructions.
17794 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
17796 (mov<mode>_softfloat): Likewise.
17797 (mov<mode>_hardfloat32): Likewise.
17798 (mov<mode>_hardfloat64): Likewise.
17799 (mov<mode>_softfloat64): Likewise.
17800 (movdi_internal32): Likewise.
17801 (movdi_internal64): Likewise.
17802 (probe_stack_<mode>): Likewise.
17804 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
17806 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
17807 floating point, and decimal floating point to reload iterator.
17809 * config/rs6000/constraints.md (wl constraint): New constraints to
17810 return FLOAT_REGS if certain options are used to reduce the number
17811 of separate patterns that exist in the file.
17812 (wx constraint): Likewise.
17813 (wz constraint): Likewise.
17815 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17816 -mdebug=reg, print wg, wl, wx, and wz constraints.
17817 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17818 Initialize the reload functions for 64-bit binary/decimal floating
17820 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17821 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17822 create the buffer on the stack to overcome not having a 32-bit
17824 (rs6000_emit_move): Likewise.
17825 (rs6000_secondary_memory_needed_rtx): Likewise.
17826 (rs6000_alloc_sdmode_stack_slot): Likewise.
17827 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17828 via xxlxor, just like DFmode 0.0.
17830 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
17831 define as 1 if we are running on a power7 or newer.
17832 (enum r6000_reg_class_enum): Add new constraints.
17834 * config/rs6000/dfp.md (movsd): Delete, combine with binary
17835 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17836 with other moves by using conditional constraits (wg). Use LFIWZX
17837 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
17838 (movsd splitter): Likewise.
17839 (movsd_hardfloat): Likewise.
17840 (movsd_softfloat): Likewise.
17842 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
17843 binary and decimal floating point moves.
17844 (fmove_ok): New attributes to combine binary and decimal floating
17845 point moves, and to combine power6x (mfpgpr) moves along normal
17847 (real_value_to_target): Likewise.
17848 (f32_lr): Likewise.
17849 (f32_lm): Likewise.
17850 (f32_li): Likewise.
17851 (f32_sr): Likewise.
17852 (f32_sm): Likewise.
17853 (f32_si): Likewise.
17854 (movsf): Combine binary and decimal floating point moves. Combine
17855 power6x (mfpgpr) moves with other moves by using conditional
17856 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
17857 (mov<mode> for SFmode/SDmode); Likewise.
17858 (SFmode/SDmode splitters): Likewise.
17859 (movsf_hardfloat): Likewise.
17860 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
17861 (movsf_softfloat): Likewise.
17862 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
17864 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
17865 wx and wz constraints.
17867 * config/rs6000/constraints.md (wg constraint): New constraint to
17868 return FLOAT_REGS if -mmfpgpr (power6x) was used.
17870 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
17873 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17874 -mdebug=reg, print wg, wl, wx, and wz constraints.
17875 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17876 Initialize the reload functions for 64-bit binary/decimal floating
17878 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17879 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17880 create the buffer on the stack to overcome not having a 32-bit
17882 (rs6000_emit_move): Likewise.
17883 (rs6000_secondary_memory_needed_rtx): Likewise.
17884 (rs6000_alloc_sdmode_stack_slot): Likewise.
17885 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17886 via xxlxor, just like DFmode 0.0.
17888 * config/rs6000/dfp.md (movdd): Delete, combine with binary
17889 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17890 with other moves by using conditional constraits (wg). Use LFIWZX
17891 and STFIWX for loading SDmode on power7.
17892 (movdd splitters): Likewise.
17893 (movdd_hardfloat32): Likewise.
17894 (movdd_softfloat32): Likewise.
17895 (movdd_hardfloat64_mfpgpr): Likewise.
17896 (movdd_hardfloat64): Likewise.
17897 (movdd_softfloat64): Likewise.
17899 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
17900 64-bit binary and decimal floating point moves.
17901 (FMOVE64X): Likewise.
17902 (movdf): Combine 64-bit binary and decimal floating point moves.
17903 Combine power6x (mfpgpr) moves with other moves by using
17904 conditional constraits (wg).
17905 (mov<mode> for DFmode/DDmode): Likewise.
17906 (DFmode/DDmode splitters): Likewise.
17907 (movdf_hardfloat32): Likewise.
17908 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
17909 (movdf_softfloat32): Likewise.
17910 (movdf_hardfloat64_mfpgpr): Likewise.
17911 (movdf_hardfloat64): Likewise.
17912 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
17913 (movdf_softfloat64): Likewise.
17914 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
17915 (reload_<mode>_load): Move to later in the file so they aren't in
17916 the middle of the floating point move insns.
17917 (reload_<mode>_store): Likewise.
17919 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
17922 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
17923 constraint if -mdebug=reg.
17924 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
17925 Enable using dd reload support if needed.
17927 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
17928 binary and decimal floating point moves in rs6000.md.
17929 (movtd_internal): Likewise.
17931 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
17932 decimal floating point moves.
17934 (movtf_internal): Likewise.
17935 (mov<mode>_internal, TDmode/TFmode): Likewise.
17936 (movtf_softfloat): Likewise.
17937 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
17939 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
17940 movdi_internal64, using wg constraint for move direct operations.
17941 (movdi_internal64): Likewise.
17943 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
17944 MODES_TIEABLE_P for selected modes. Print the numerical value of
17945 the various virtual registers. Use GPR/FPR first/last values,
17946 instead of hard coding the register numbers. Print which modes
17947 have reload functions registered.
17948 (rs6000_option_override_internal): If -mdebug=reg, trace the options
17949 settings before/after setting cpu, target and subtarget settings.
17950 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
17951 and for secondary reload failures in rs6000_secondary_reload_inner.
17952 (rs6000_secondary_reload_fail): Likewise.
17953 (rs6000_secondary_reload_inner): Likewise.
17955 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
17956 macros for first/last GPR and FPR registers.
17957 (LAST_GPR_REGNO): Likewise.
17958 (FIRST_FPR_REGNO): Likewise.
17959 (LAST_FPR_REGNO): Likewise.
17961 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
17962 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
17963 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
17964 (vcond<mode><mode>): Likewise.
17965 (vcondu<mode><mode>): Likewise.
17966 (vector_gtu<mode>): Likewise.
17967 (vector_gte<mode>): Likewise.
17968 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
17969 to prevent the compiler from converting DImode operations to TImode.
17970 (ior<mode>3): Likewise.
17971 (and<mode>3): Likewise.
17972 (one_cmpl<mode>2): Likewise.
17973 (nor<mode>3): Likewise.
17974 (andc<mode>3): Likewise.
17976 * config/rs6000/constraints.md (wt constraint): New constraint
17977 that returns VSX_REGS if TImode is allowed in VSX registers.
17979 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
17980 constant under VSX.
17982 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
17983 similar to TImode, but it is restricted to being in the GPRs.
17985 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
17986 TImode to occupy a single VSX register.
17988 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
17989 -mvsx-timode for power7/power8.
17990 (power7 cpu): Likewise.
17991 (power8 cpu): Likewise.
17993 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
17994 sure that TFmode/TDmode take up two registers if they are ever
17995 allowed in the upper VSX registers.
17996 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
17998 (rs6000_init_hard_regno_mode_ok): Likewise.
17999 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18000 constraint. Print if LRA is turned on.
18001 (rs6000_option_override_internal): Give an error if -mvsx-timode
18002 and VSX is not enabled.
18003 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18004 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18005 to reg+offset addressing. Use PTImode when checking offset
18006 addresses for validity.
18007 (reg_offset_addressing_ok_p): Likewise.
18008 (rs6000_legitimate_offset_address_p): Likewise.
18009 (rs6000_legitimize_address): Likewise.
18010 (rs6000_legitimize_reload_address): Likewise.
18011 (rs6000_legitimate_address_p): Likewise.
18012 (rs6000_eliminate_indexed_memrefs): Likewise.
18013 (rs6000_emit_move): Likewise.
18014 (rs6000_secondary_reload): Likewise.
18015 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18016 reloads to fpr registers to continue to use reg+offset addressing,
18017 but 64-bit reloads to altivec registers need reg+reg addressing.
18018 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18019 it. Treat LO_SUM like a PLUS operation.
18020 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18021 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18022 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18023 registers to share a register with a smaller sized type, since VSX
18024 puts scalars in the upper 64-bits.
18025 (print_operand): Add support for PTImode.
18026 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18027 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18028 registers, but don't have arithmetic support.
18029 (rs6000_memory_move_cost): Add test for VSX.
18030 (rs6000_opt_masks): Add -mvsx-timode.
18032 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18035 (VSr): Use wt constraint for TImode.
18036 (VSv): Drop TImode support.
18037 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18038 (vsx_movti_64bit): Likewise.
18039 (vsx_movti_32bit): Likewise.
18040 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18041 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18042 one '?' on the appropriate output constraint. Do not allow TImode
18043 logical operations on 32-bit systems.
18044 (vsx_ior<mode>3): Likewise.
18045 (vsx_xor<mode>3): Likewise.
18046 (vsx_one_cmpl<mode>2): Likewise.
18047 (vsx_nor<mode>3): Likewise.
18048 (vsx_andc<mode>3): Likewise.
18049 (vsx_concat_<mode>): Likewise.
18050 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18052 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
18053 OPTION_MASK_VSX_TIMODE.
18054 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
18055 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
18057 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
18058 (TI2 iterator): New iterator for TImode, PTImode.
18059 (wd mode attribute): Add values for vector types.
18060 (movti_string): Replace TI move operations with operations for TImode
18061 and PTImode. Add support for TImode being allowed in VSX registers.
18062 (mov<mode>_string, TImode/PTImode): Likewise.
18063 (movti_ppc64): Likewise.
18064 (mov<mode>_ppc64, TImode/PTImode): Likewise.
18065 (TI mode splitters): Likewise.
18067 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
18070 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
18072 PR tree-optimization/56355
18073 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
18074 Also handle integers with undefined overflow.
18076 2013-03-20 Catherine Moore <clm@codesourcery.com>
18077 Maciej W. Rozycki <macro@codesourcery.com>
18078 Tom de Vries <tom@codesourcery.com>
18079 Nathan Sidwell <nathan@codesourcery.com>
18080 Iain Sandoe <iain@codesourcery.com>
18081 Nathan Froyd <froydnj@codesourcery.com>
18082 Chao-ying Fu <fu@mips.com>
18084 * doc/extend.texi: (micromips, nomicromips, nocompression):
18085 Document new function attributes.
18086 * doc/invoke.texi (minterlink-compressed, mmicromips,
18087 m14k, m14ke, m14kec): Document new options.
18088 (minterlink-mips16): Update documentation.
18089 * doc/md.texi (ZC, ZD): Document new constraints.
18090 * configure.ac (gcc_cv_as_micromips): Check if linker
18091 supports the .set micromips directive.
18092 * configure: Regenerate.
18093 * config.in: Regenerate.
18094 * config/mips/mips-tables.opt: Regenerate.
18095 * config/mips/micromips.md: New file.
18096 * constraints.md (ZC, ZD): New constraints.
18097 * config/mips/predicates.md (movep_src_register): New predicate.
18098 (movep_src_operand): New predicate.
18099 (non_volatile_mem_operand): New predicate.
18100 * config/mips/mips.md (multimem): New type.
18101 (length): Differentiate between 17-bit and 18-bit branch offsets.
18102 (MOVEP1, MOVEP2): New mode iterator.
18103 (mov_<load>l): Use ZC constraint.
18104 (mov_<load>r): Likewise.
18105 (mov_<store>l): Likewise.
18106 (mov_<store>r): Likewise.
18107 (*branch_equality<mode>_inverted): Add microMIPS support.
18108 (*branch_equality<mode>): Likewise.
18109 (*jump_absolute): Likewise.
18110 (indirect_jump_<mode>): Likewise.
18111 (tablejump_<mode>): Likewise.
18112 (<optab>_internal): Likewise.
18113 (sibcall_internal): Likewise.
18114 (sibcall_value_internal): Likewise.
18115 (prefetch): Use constraint ZD.
18116 * config/mips/mips.opt (minterlink-compressed): New option.
18117 (minterlink-mips16): Now an alias for minterlink-compressed.
18118 (mmicromips): New option.
18119 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18120 (compare_and_swap_12): Likewise.
18121 (sync_add<mode>): Likewise.
18122 (sync_<optab>_12): Likewise.
18123 (sync_old_<optab>_12): Likewise.
18124 (sync_new_<optab>_12): Likewise.
18125 (sync_nand_12): Likewise.
18126 (sync_old_nand_12): Likewise.
18127 (sync_new_nand_12): Likewise.
18128 (sync_sub<mode>): Likewise.
18129 (sync_old_add<mode>): Likewise.
18130 (sync_old_sub<mode>): Likewise.
18131 (sync_new_add<mode>): Likewise.
18132 (sync_new_sub<mode>): Likewise.
18133 (sync_<optab><mode>): Likewise.
18134 (sync_old_<optab><mode>): Likewise.
18135 (sync_new_<optab><mode>): Likewise.
18136 (sync_nand<mode>): Likewise.
18137 (sync_old_nand<mode>): Likewise.
18138 (sync_new_nand<mode>): Likewise.
18139 (sync_lock_test_and_set<mode>): Likewise.
18140 (test_and_set_12): Likewise.
18141 (atomic_compare_and_swap<mode>): Likewise.
18142 (atomic_exchange<mode>_llsc): Likewise.
18143 (atomic_fetch_add<mode>_llsc): Likewise.
18144 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18145 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18146 (umips_save_restore_pattern_p): Likewise.
18147 (umips_load_store_pair_p): Likewise.
18148 (umips_output_load_store_pair): Likewise.
18149 (umips_movep_target_p): Likewise.
18150 (umips_12bit_offset_address_p): Likewise.
18151 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18152 (mips_base_mips16): Rename this...
18153 (mips_base_compression_flags): ...to this. Update all uses.
18154 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18155 (mips_mips16_decl_p): Delete.
18156 (mips_nomips16_decl_p): Delete.
18157 (mips_get_compress_on_flags): New function.
18158 (mips_get_compress_off_flags): New function.
18159 (mips_get_compress_mode): New function.
18160 (mips_get_compress_on_name): New function.
18161 (mips_get_compress_off_name): New function.
18162 (mips_insert_attributes): Support multiple compression types.
18163 (mips_merge_decl_attributes): Likewise.
18164 (umips_12bit_offset_address_p): New function.
18165 (mips_start_function_definition): Emit .set micromips directive.
18166 (mips_call_may_need_jalx_p): New function.
18167 (mips_function_ok_for_sibcall): Add microMIPS support.
18168 (mips_print_operand_punctuation): Support short delay slots and
18170 (umips_swm_mask, umips_swm_encoding): New.
18171 (umips_build_save_restore): New function.
18172 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18173 (was_mips16_p): Remove.
18174 (old_compression_mode): New.
18175 (mips_set_compression_mode): New function.
18176 (mips_set_current_function): Add microMIPS support.
18177 (mips_option_override): Likewise.
18178 (umips_save_restore_pattern_p): New function.
18179 (umips_output_save_restore): New function.
18180 (umips_load_store_pair_p_1): New function.
18181 (umips_load_store_pair_p): New function.
18182 (umips_output_load_store_pair_1): New function.
18183 (umips_output_load_store_pair): New function.
18184 (umips_movep_target_p) New function.
18185 (mips_prepare_pch_save): Add microMIPS support.
18186 * config/mips/mips.h (TARGET_COMPRESSION): New.
18187 (TARGET_CPU_CPP_BUILTINS): Update macro
18188 to use new compression flags and to support microMIPS.
18189 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18190 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18191 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18192 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18193 (ASM_SPEC): Support mmicromips and mno-micromips.
18194 (M16STORE_REG_P): New macro.
18195 (MIPS_CALL): Support TARGET_MICROMIPS.
18196 (MICROMIPS_J): New macro.
18197 (mips_base_mips16): Rename this...
18198 (mips_base_compression_flags): ...to this.
18199 (UMIPS_12BIT_OFFSET_P): New macro.
18200 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18201 (MULTILIB_DIRNAMES): Likewise.
18202 2013-03-20 Richard Biener <rguenther@suse.de>
18204 PR tree-optimization/56661
18205 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
18206 the result does not have to be distinct.
18208 2013-03-20 Richard Biener <rguenther@suse.de>
18210 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
18213 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18214 Steven Bosscher <steven@gcc.gnu.org>
18216 PR rtl-optimization/56605
18217 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
18219 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18222 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
18223 that require movd instead of movq.
18225 2013-03-20 Richard Biener <rguenther@suse.de>
18227 * tree-ssa-structalias.c (struct variable_info): Add pointer
18228 to the first field of an aggregate with sub-vars. Make
18229 this and the pointer to the next subfield its ID.
18230 (vi_next): New function.
18231 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
18232 storedanything_id, integer_id): Increment by one.
18233 (new_var_info, get_call_vi, lookup_call_clobber_vi,
18234 get_call_clobber_vi): Adjust.
18235 (solution_set_expand): Simplify and speedup.
18236 (solution_set_add): Inline into ...
18237 (set_union_with_increment): ... this. Adjust accordingly.
18238 (do_sd_constraint): Likewise.
18239 (do_ds_constraint): Likewise.
18240 (do_complex_constraint): Simplify.
18241 (build_pred_graph): Adjust.
18242 (solve_graph): Likewise. Simplify and speedup.
18243 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
18244 get_constraint_for_component_ref, get_constraint_for_1,
18245 first_vi_for_offset, first_or_preceding_vi_for_offset,
18246 create_function_info_for, create_variable_info_for_1,
18247 create_variable_info_for, intra_create_variable_infos): Adjust.
18248 (init_base_vars): Push NULL for ID zero.
18249 (compute_points_to_sets): Adjust.
18251 2013-03-20 Richard Biener <rguenther@suse.de>
18253 * cfgloop.c (verify_loop_structure): Streamline and avoid
18254 ICEing on corrupt loop tree.
18255 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
18258 2013-03-20 Richard Biener <rguenther@suse.de>
18260 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
18261 check whether an SSA update is needed.
18263 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
18265 * config/mips/constraints.md (T): Rename to...
18269 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
18270 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
18272 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18274 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
18275 (*subsi3_carryin_uxtw): Likewise.
18277 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18279 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
18280 (*rorsi3_insn_uxtw): Likewise.
18282 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18284 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
18285 (*extrsi5_insn_uxtw): Likewise.
18287 2013-03-19 Richard Biener <rguenther@suse.de>
18289 PR tree-optimization/56273
18290 * passes.c (init_optimization_passes): Move second VRP after DOM.
18292 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
18294 * config/i386/i386.md (*movti_internal): Merge from
18295 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
18296 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
18297 nox64 isa attributes.
18299 2013-03-18 Richard Biener <rguenther@suse.de>
18301 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
18303 (merge_node_constraints): Likewise.
18304 (build_succ_graph): Likewise.
18305 (valid_graph_edge): Inline into single caller.
18306 (unify_nodes): Likewise. Use bitmap_set_bit return value
18308 (scc_visit): Fix formatting and variable use.
18309 (do_sd_constraint): Use gcc_checking_assert.
18310 (do_ds_constraint): Likewise.
18311 (do_complex_constraint): Likewise.
18312 (condense_visit): Likewise. Cleanup.
18313 (dump_pred_graph): New function.
18314 (perform_var_substitution): Dump the pred-graph before
18315 variable substitution.
18316 (find_equivalent_node): Use gcc_checking_assert.
18317 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
18319 2013-03-18 Richard Biener <rguenther@suse.de>
18321 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
18322 Remove cond_expr_stmt_list argument and do not gimplify the
18324 (vect_loop_versioning): Adjust.
18325 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
18326 Cleanup to use less temporaries.
18327 (vect_create_data_ref_ptr): Cleanup.
18329 2013-03-18 Jakub Jelinek <jakub@redhat.com>
18331 PR tree-optimization/56635
18332 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
18333 require types_compatible_p types.
18335 2013-03-18 Nick Clifton <nickc@redhat.com>
18337 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
18338 spurious backslash.
18340 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
18341 Add missing line to comment describing function.
18343 2013-03-18 Richard Biener <rguenther@suse.de>
18345 PR tree-optimization/56210
18346 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18347 Handle string / character search functions.
18348 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
18350 2013-03-18 Richard Biener <rguenther@suse.de>
18352 PR middle-end/56483
18353 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
18354 and implement properly.
18355 * gimple.h (gimple_cond_single_var_p): Remove.
18357 2013-03-18 Richard Biener <rguenther@suse.de>
18359 * tree-data-ref.h (find_data_references_in_loop): Declare.
18360 * tree-data-ref.c (get_references_in_stmt): Use a stack
18361 vector pre-allocated in the callers.
18362 (find_data_references_in_stmt): Adjust.
18363 (graphite_find_data_references_in_stmt): Likewise.
18364 (create_rdg_vertices): Likewise.
18365 (find_data_references_in_loop): Export.
18366 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
18367 Compute dependences here...
18368 (vect_analyze_data_refs): ...not here. When we encounter
18369 a non-vectorizable data reference in basic-block vectorization
18370 truncate the data reference vector. Do not bother to
18371 fixup data-dependence information for gather loads.
18372 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
18373 of data references, as reported.
18375 2013-03-18 Richard Biener <rguenther@suse.de>
18377 PR tree-optimization/3713
18378 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
18379 has_constants and expr.
18380 (stmt_has_constants): Properly valueize SSA names when deciding
18381 whether the stmt has constants.
18383 2013-03-18 Richard Biener <rguenther@suse.de>
18385 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
18386 whole function when there is nothing to do.
18387 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
18388 * tree-vectorizer.c (vectorize_loops): Update virtual and
18389 loop-closed SSA once.
18390 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
18392 2013-03-18 Richard Biener <rguenther@suse.de>
18394 PR middle-end/56113
18395 * domwalk.c (bb_postorder): New global static.
18396 (cmp_bb_postorder): New function.
18397 (walk_dominator_tree): Replace scheme imposing an order for
18398 visiting dominator sons by one sorting them at the time they
18399 are pushed on the stack.
18401 2013-03-18 Richard Biener <rguenther@suse.de>
18403 PR tree-optimization/39326
18404 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
18405 (struct mem_ref): Replace mem member with ao_ref typed member.
18406 (MEM_ANALYZABLE): Adjust.
18407 (memref_eq): Likewise.
18408 (mem_ref_alloc): Likewise.
18409 (gather_mem_refs_stmt): Likewise.
18410 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
18411 (execute_sm_if_changed_flag_set): Adjust.
18412 (execute_sm): Likewise.
18413 (ref_always_accessed_p): Likewise.
18414 (refs_independent_p): Likewise.
18415 (can_sm_ref_p): Likewise.
18417 2013-03-18 Jakub Jelinek <jakub@redhat.com>
18420 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
18421 return 1 even for !unsignedp.
18423 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18425 * config/i386/i386.md (isa): Add x64 and nox64.
18426 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
18427 (*pushtf): Enable *roF alternative for x64 isa only.
18428 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
18429 mode attribute of integer alternatives to DImode for TARGET_64BIT.
18430 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
18431 (*movtf_internal): Merge from *movtf_internal_rex64 and
18432 *movtf_internal_sse. Use x64 and nox64 isa attributes.
18433 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
18434 nox64 isa attributes.
18435 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
18436 nox64 isa attributes.
18437 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
18439 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18441 * config/alpha/alpha.c (TARGET_LRA_P): New define.
18443 2013-03-17 Jakub Jelinek <jakub@redhat.com>
18446 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
18447 class names. Remove trailing comma after "ALL_REGS".
18449 2013-03-16 Jan Hubicka <jh@suse.cz>
18451 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
18452 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
18453 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
18454 of cgraph_get_create_node.
18455 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
18457 2013-03-16 Jason Merrill <jason@redhat.com>
18460 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
18461 with DW_AT_default_value.
18463 2013-03-16 Jakub Jelinek <jakub@redhat.com>
18465 * BASE-VER: Set to 4.9.0.
18467 2013-03-14 Andi Kleen <ak@linux.intel.com>
18470 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
18471 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
18472 Document _x* TSX intrinsics.
18474 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
18475 David Holsgrove <david.holsgrove@xilinx.com>
18477 * configure.ac: Add MicroBlaze TLS support detection.
18478 * configure: Regenerate.
18479 * config/microblaze/microblaze-protos.h
18480 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
18481 symbol_mentioned_p, label_mentioned_p): Add prototypes.
18482 * config/microblaze/microblaze.c (microblaze_address_type): Add
18483 ADDRESS_TLS and tls_reloc address types.
18484 (microblaze_address_info): Add tls_reloc.
18485 (TARGET_HAVE_TLS): Define.
18486 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
18487 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
18488 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
18489 load_tls_operand, microblaze_call_tls_get_addr,
18490 microblaze_legitimize_tls_address): New functions.
18491 (microblaze_classify_unspec): Handle UNSPEC_TLS.
18492 (get_base_reg): Use microblaze_tls_symbol_p.
18493 (microblaze_classify_address): Handle TLS.
18494 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
18495 label_mentioned_p and microblaze_tls_referenced_p.
18496 (microblaze_legitimize_address): Handle TLS.
18497 (microblaze_address_insns): Handle ADDRESS_TLS.
18498 (pic_address_needs_scratch): Handle TLS.
18499 (print_operand_address): Handle TLS.
18500 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
18501 (microblaze_expand_move): Handle TLS.
18502 (microblaze_legitimate_constant_p): Check
18503 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
18504 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
18505 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
18506 (PIC_OFFSET_TABLE_REGNUM): Set.
18507 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
18508 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
18509 (addsi3, movsi_internal2, movdf_internal): Update constraints
18510 * config/microblaze/predicates.md (arith_plus_operand): Define
18511 (move_operand): Redefine as move_src_operand,
18512 check microblaze_tls_referenced_p.
18514 2013-03-14 Ian Bolton <ian.bolton@arm.com>
18516 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
18517 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
18519 2013-03-14 Ian Bolton <ian.bolton@arm.com>
18521 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
18524 2013-03-14 Jakub Jelinek <jakub@redhat.com>
18526 PR tree-optimization/53265
18527 * common.opt (Waggressive-loop-optimizations): New option.
18528 * tree-ssa-loop-niter.c: Include tree-pass.h.
18529 (do_warn_aggressive_loop_optimizations): New function.
18530 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
18531 if number_of_latch_executions returned constant.
18532 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
18533 early. If number_of_latch_executions returned constant, set
18534 nb_iterations_upper_bound back to it.
18535 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
18537 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
18538 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
18540 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
18541 (MULTILIB_OSDIRNAMES): Set.
18542 * genmultilib: If defaultosdirname doesn't start with :: , set
18543 defaultosdirname2 instead, clear it and emit two . multilib_raw
18544 entries instead of just one.
18546 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
18548 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
18549 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
18550 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
18551 (SUBTARGET_OVERRIDE_OPTIONS): New.
18553 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
18556 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
18557 (musermode): Convert to Var(TARGET_USERMODE).
18558 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
18559 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
18560 * config/sh/sh.c (sh_option_override): Use
18561 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
18562 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
18564 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
18566 (udivsi3_i4_single, divsi3_i4_single): Use
18567 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
18569 2013-03-13 Dave Korn <dave.korn.cygwin@....>
18571 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
18574 2013-03-13 Richard Biener <rguenther@suse.de>
18576 PR tree-optimization/56608
18577 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
18578 calls when vectorizing basic-blocks.
18580 2013-03-13 Jakub Jelinek <jakub@redhat.com>
18583 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
18586 2013-03-12 Jakub Jelinek <jakub@redhat.com>
18588 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
18590 2013-03-11 Jan Hubicka <jh@suse.cz>
18593 * lto-streamer-out.c (output_symbol_p): Skip references from
18594 constructors of external variables.
18596 2013-03-11 Jan Hubicka <jh@suse.cz>
18598 PR middle-end/56571
18599 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
18601 * emit-rtl.c (verify_rtx_sharing): Likewise.
18602 (copy_insn_1): Likewise.
18603 * rtl.c (copy_rtx): Likewise.
18605 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
18608 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
18609 output_operand_lossage message.
18611 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
18614 * arm.c (shift_op): Validate RTL pattern on the fly.
18615 (arm_print_operand, case 'S'): Don't use shift_operator to validate
18618 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18621 * config/pa/pa.md (call_value): Check for calls to powf and direct to
18622 new call patterns that clobber %fr12.
18623 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
18624 split and postreload patterns.
18625 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
18626 registers %fr12 and %fr12R as call used.
18628 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
18630 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
18631 (canon_address, record_store, replace_read, check_mem_read_rtx,
18632 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
18633 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
18634 rest_of_handle_dse): Likewise.
18636 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
18638 PR middle-end/56524
18639 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
18641 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
18642 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
18643 (save_optabs_if_changed): Replace with...
18644 (init_tree_optimization_optabs): ...this.
18645 * optabs.c (save_optabs_if_changed): Rename to...
18646 (init_tree_optimization_optabs): ...this. Take the optimization node
18647 as argument. Do nothing if the base optabs are already correct.
18648 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
18649 to recompute optabs.
18650 * function.h (function): Remove optabs field.
18651 * function.c (invoke_set_current_function_hook): Call
18652 init_tree_optimization_optabs. Use the result to initialize
18655 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
18657 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
18658 if GTMA_HAS_NO_INSTRUMENTATION.
18659 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
18660 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
18661 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
18662 * gimple-pretty-print.c (dump_gimple_transaction): Handle
18663 GTMA_HAS_NO_INSTRUMENTATION.
18665 2013-03-08 Jakub Jelinek <jakub@redhat.com>
18667 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
18670 2013-03-08 Marek Polacek <polacek@redhat.com>
18671 Jakub Jelinek <jakub@redhat.com>
18673 PR tree-optimization/56478
18674 * predict.c (is_comparison_with_loop_invariant_p): Change the
18675 type of loop_step to tree.
18676 (predict_loops): Adjust.
18677 (predict_iv_comparison): Perform the computations on double_ints.
18679 2013-03-08 Richard Biener <rguenther@suse.de>
18681 PR tree-optimization/56570
18682 * tree-cfg.c (verify_expr_location_1): Verify locations for
18684 * tree-sra.c (create_access_replacement): Strip locations
18685 from DECL_DEBUG_EXPRs.
18687 2013-03-08 Richard Biener <rguenther@suse.de>
18689 * tree-inline.c (expand_call_inline): Do not associate
18690 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
18691 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
18693 2013-03-08 Richard Biener <rguenther@suse.de>
18695 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
18696 or block changes with -Og. Fix for location / block encoding
18697 changes and PHI arguments with locations.
18699 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
18701 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
18703 (struct output_info): Likewise.
18704 (register_overhead): Remove bad gcc_assert.
18705 (bitmap_find_bit): If there is only a single bitmap element, do not
18706 count a miss as a search.
18707 (print_statistics): Update for counter type changes.
18708 (dump_bitmap_statistics): Likewise. Print headers such that they
18709 are properly lined up with the printed counters.
18711 2013-03-07 Jakub Jelinek <jakub@redhat.com>
18713 PR tree-optimization/56559
18714 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
18715 check that it has only a single use.
18717 2013-03-07 Richard Biener <rguenther@suse.de>
18719 * doc/invoke.texi (fwhole-program): Discourage use in combination
18722 2013-03-06 Jakub Jelinek <jakub@redhat.com>
18724 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
18726 PR tree-optimization/56539
18727 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
18728 instead of GSI_CONTINUE_LINKING as last argument to
18729 force_gimple_operand_gsi. Adjust function comment.
18731 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
18734 PR middle-end/56548
18735 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
18736 promoted mode, convert the result back to the original mode.
18738 2013-03-06 Richard Biener <rguenther@suse.de>
18740 PR middle-end/56294
18741 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
18742 (insert_updated_phi_nodes_compare_uids): New function.
18743 (update_ssa): Sort symbols_to_rename after UID before
18744 traversing it to insert PHI nodes.
18746 2013-03-06 Richard Biener <rguenther@suse.de>
18748 PR middle-end/50494
18749 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
18750 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
18753 2013-02-13 Richard Biener <rguenther@suse.de>
18756 * varasm.c (output_constant_def_1): Get the decl representing
18757 the constant as argument.
18758 (output_constant_def): Wrap output_constant_def_1.
18759 (make_decl_rtl): Use output_constant_def_1 with the decl
18760 representing the constant.
18761 (build_constant_desc): Optionally re-use a decl already
18762 representing the constant.
18763 (tree_output_constant_def): Adjust.
18765 2013-03-06 Joey Ye <joey.ye@arm.com>
18768 * gcc.c (convert_white_space): New function.
18769 (main): Handles white space in function name.
18771 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
18774 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
18775 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
18776 to SH_DIV_CALL_TABLE for TARGET_SH2.
18777 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
18779 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
18780 call-table options.
18782 2013-03-05 Sterling Augustine <saugustine@google.com>
18783 Cary Coutant <ccoutant@google.com>
18786 * dwarf2out.c (resolve_addr): Don't call
18787 remove_loc_list_addr_table_entries a second time for the same
18790 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18793 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
18794 (avoid_complex_debug_insns): New function.
18795 (expand_debug_locations): Call it.
18797 PR rtl-optimization/56484
18798 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
18799 lifetimes of hard registers on small register class machines.
18801 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18803 * config/microblaze/microblaze-protos.h: Rename
18804 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
18805 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
18807 (microblaze_fast_interrupt_function_p): New function.
18808 (microblaze_is_interrupt_handler): Rename to
18809 microblaze_is_interrupt_variant and add fast_interrupt check.
18810 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
18811 (save_restore_insns): Likewise.
18812 (compute_frame_size): Likewise.
18813 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
18814 (microblaze_globalize_label): Likewise.
18815 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
18816 * config/microblaze/microblaze.md: Use wrapper
18817 microblaze_is_interrupt_variant.
18819 2013-03-05 Kai Tietz <ktietz@redhat.com>
18821 * sdbout.c (sdbout_one_type): Switch to current function's section
18822 supporting cold/hot.
18824 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18826 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
18829 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18831 PR middle-end/56461
18832 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
18833 if VALGRIND_GET_VBITS is defined, temporarily make object
18834 memory all defined, and restore previous valgrind addressability
18835 and definability afterwards. Free this_object at the end.
18837 PR middle-end/56461
18838 * lra.c (lra): Call lra_clear_live_ranges if live_p,
18839 right before calling lra_create_live_ranges, also call it
18840 when clearing live_p. Only call lra_clear_live_ranges
18841 at the end if live_p.
18843 PR middle-end/56461
18844 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
18846 2013-03-05 Richard Biener <rguenther@suse.de>
18848 PR tree-optimization/56521
18849 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
18852 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
18855 * except.h (remove_unreachable_eh_regions): New prototype.
18856 * except.c (remove_eh_handler_splicer): New function, split out
18857 of remove_eh_handler.
18858 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
18859 warning about running it on many EH regions one at a time.
18860 (remove_unreachable_eh_regions_worker): New function, walk the
18861 EH tree in depth-first order and remove non-marked regions.
18862 (remove_unreachable_eh_regions): New function.
18863 * tree-eh.c (mark_reachable_handlers): New function, split out
18864 from remove_unreachable_handlers.
18865 (remove_unreachable_handlers): Use mark_reachable_handlers and
18866 remove_unreachable_eh_regions.
18867 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
18868 and remove_unreachable_eh_regions.
18870 2013-03-05 Richard Biener <rguenther@suse.de>
18872 PR middle-end/56525
18873 * loop-init.c (fix_loop_structure): Remove loops in two stages,
18874 not freeing them until the end.
18876 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18878 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
18880 2013-03-05 Richard Biener <rguenther@suse.de>
18882 PR tree-optimization/56270
18883 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
18884 of loads after scheduling an SLP instance.
18886 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18888 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
18890 (check_gcc_parallelize): Run guality.exp as a separate job from
18891 vect.exp with unsorted.exp and $(dg_target_exps) separately from
18892 struct-layout-1.exp with stackalign.exp.
18894 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
18896 PR middle-end/56461
18897 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
18898 load_index sbitmap even if some bit in it isn't set.
18900 PR middle-end/56461
18901 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
18902 (discover_iteration_bound_by_body_walk): Change queues to
18903 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
18904 spelling in comment. Call safe_push on queues[bound_index] directly.
18905 Release queues[queue_index] in every iteration unconditionally.
18906 Release bounds vector.
18908 PR middle-end/56461
18909 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
18910 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
18911 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
18914 2013-03-05 Richard Biener <rguenther@suse.de>
18917 * tree-inline.c (remap_blocks_to_null): New function.
18918 (expand_call_inline): When expanding a call stmt without
18919 an associated block inline remap all callee blocks to NULL.
18921 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18923 PR rtl-optimization/56494
18924 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
18925 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
18926 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
18928 PR middle-end/56461
18929 * sel-sched-ir.c (free_sched_pools): Release
18930 succs_info_pool.stack[succs_info_pool.max_top] vectors too
18931 if succs_info_pool.max_top isn't -1.
18934 * opts.c (opts_obstack, opts_concat): Moved to...
18935 * opts-common.c (opts_obstack, opts_concat): ... here.
18937 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18939 PR middle-end/56461
18940 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
18942 2013-03-04 Martin Jambor <mjambor@suse.cz>
18944 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
18945 all appropriate places.
18947 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
18949 PR tree-optimization/56424
18950 * ipa-split.c (split_function): Do not set the RSO flag if result is
18951 not by reference and its type is a register type.
18953 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
18955 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
18956 (microblaze_legitimate_pic_operand): Likewise
18957 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
18958 new function microblaze_legitimate_pic_operand
18959 * config/microblaze/microblaze-protos.h
18960 (microblaze_legitimate_pic_operand): Declare.
18962 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18964 * config/microblaze/predicates.md (call_insn_simple_operand):
18965 New predicate for supported rtx code types.
18966 * config/microblaze/microblaze.md (call_internal1): Use
18967 call_insn_simple_operand predicate.
18969 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18971 PR middle-end/56461
18972 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
18973 partitions.ordered_remove.
18975 PR middle-end/56461
18976 * tree-vect-stmts.c (vectorizable_conversion): Don't call
18977 vec_oprnds0.create (1) for modifier == NONE.
18979 PR middle-end/56461
18980 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
18981 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
18982 vec_oprnds1 right before pushing anything to it for
18985 PR middle-end/56461
18986 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
18987 set nbbs to 0 instead of having separate code path.
18988 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
18989 instead of false as last argument if returning NULL.
18991 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
18993 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
18994 the attribute is now called "target" instead of "option".
18995 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
18996 * doc/tm.texi.in (Target Attributes): Likewise document the correct
18997 attribute/pragma name for TARGET_OPTION_VALID_P and
18998 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
18999 * doc/tm.texi: Regenerated.
19001 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19003 * config/microblaze/microblaze.c:
19004 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19005 * config/microblaze/microblaze.h: Add -mxl-reorder to
19007 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19008 instructions emitted if TARGET_REORDER.
19009 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19010 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19013 2013-03-01 Xinliang David Li <davidxl@google.com>
19015 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19018 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19020 PR middle-end/56461
19021 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19022 vector even when returning true. Fix up function comment formatting.
19024 PR middle-end/56461
19025 * ira-build.c (ira_loop_nodes_count): New variable.
19026 (create_loop_tree_nodes): Initialize it.
19027 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19029 PR middle-end/56461
19030 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19031 method on dr_chain and result_chain.
19032 * tree-vect-stmts.c (vectorizable_store): Only call
19033 result_chain.create if j == 0.
19035 PR middle-end/56461
19036 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19037 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19038 before overwriting it.
19040 2013-03-01 Tobias Burnus <burnus@net-b.de>
19042 * doc/extended.texi (C Extensions): Change order in @menu
19044 (Other MIPS Built-in Functions): Move last MIPS entry before
19045 "picoChip Built-in Functions".
19046 (SH Built-in Functions): Move after RX Built-in Functions.
19047 * doc/gcc.texi (Introduction): Change order in @menu
19049 * doc/md.texi (Constraints): Ditto.
19050 * gty.texi (Type Information): Ditto.
19051 (User-provided marking routines for template types): Make
19053 * doc/invoke.texi (AArch64 Options): Move before
19054 "Adapteva Epiphany Options".
19056 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
19057 Jakub Jelinek <jakub@redhat.com>
19060 * asan.c (gate_asan): Lookup no_sanitize_address instead of
19061 no_address_safety_analysis attribute.
19062 * doc/extend.texi (no_address_safety_attribute): Rename to
19063 no_sanitize_address attribute, mention no_address_safety_analysis
19064 attribute as deprecated alias.
19066 2013-02-28 Jakub Jelinek <jakub@redhat.com>
19068 PR middle-end/56461
19069 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
19070 type to vec<vec<tree> > *.
19071 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
19072 to be vec<tree> instead of vec<tree> *, set vec_defs
19073 to vNULL and call vec_defs.create (number_of_vects), adjust other
19075 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
19076 vectorizable_condition): Adjust vect_get_slp_defs callers.
19078 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19080 * config/aarch64/aarch64.c
19081 (aarch64_float_const_representable): Remove unused variable.
19083 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19085 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
19087 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19089 * config/aarch64/aarch64-builtins.c
19090 (aarch64_init_simd_builtins): Make static.
19092 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19094 * config/aarch64/aarch64.c
19095 (aarch64_simd_make_constant): Make static.
19097 2013-02-28 Martin Jambor <mjambor@suse.cz>
19099 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
19100 with no initialization to the RHS of debug statements.
19102 2013-02-28 Martin Jambor <mjambor@suse.cz>
19104 PR tree-optimization/56294
19105 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
19107 (get_access_replacement): Do not call create_access_replacement.
19108 Assert a replacement exists.
19109 (get_repl_default_def_ssa_name): Create the replacement declaration
19112 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19114 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
19115 final_end_function.
19117 2013-02-28 Marek Polacek <polacek@redhat.com>
19119 PR rtl-optimization/56466
19120 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
19121 if we're changing a loop.
19122 (peel_loops_completely): Likewise.
19124 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
19127 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
19129 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19132 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
19133 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
19134 INTX_FTYPE_FX, FX_FTYPE_INTX.
19135 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
19137 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19139 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
19140 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
19141 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
19142 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
19143 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
19144 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
19145 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
19146 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
19147 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
19148 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
19149 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
19150 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
19151 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
19152 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
19153 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
19154 (avrxmega6): Increase max flash segments from 5 to 6.
19155 * config/avr/t-multilib: Regenerate.
19156 * config/avr/avr-tables.opt: Regenerate.
19157 * doc/avr-mmcu.texi: Regenerate.
19159 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19161 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
19162 (avr_device_to_arch): Rename to avr_device_to_ld.
19163 (avr_device_to_as): New prototype.
19164 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
19165 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
19166 * config/avr/driver-avr.c (avr_device_to_as): New.
19167 (avr_device_to_arch): Rename to avr_device_to_ld.
19169 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19171 PR middle-end/56461
19172 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
19173 method on dr_chain and result_chain.
19175 PR middle-end/56461
19176 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
19177 pointer_set_destroy on not_executed_last_iteration.
19179 PR middle-end/56461
19180 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
19182 PR middle-end/56461
19183 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
19184 FOR_EACH_DEFINED_FUNCTION when freeing state.
19186 PR middle-end/56461
19187 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
19189 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
19192 PR middle-end/56461
19193 * ipa-cp.c (decide_whether_version_node): Call vec_free on
19194 known_aggs[i].items and release known_aggs vector.
19196 PR middle-end/56461
19197 * ipa-reference.c (propagate): Free node_info even for alias nodes.
19199 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19201 * config/microblaze/microblaze.c (microblaze_emit_compare):
19202 Use xor for EQ/NE comparisions.
19203 * config/microblaze/microblaze.md (cstoresf4): Add constraints
19204 (cbranchsf4): Adjust operator to comparison_operator.
19206 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19208 PR middle-end/56461
19209 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
19211 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
19212 vec_safe_push, always update *slot.
19213 (redirect_edge_var_map_clear): Use vec_free.
19214 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
19215 (free_var_map_entry): Use vec_free.
19216 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
19217 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
19219 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
19221 PR middle-end/45472
19222 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
19223 when the may_trap_p bit of the exprs being merged differs.
19224 Reorder tests for speculativeness in the logical and operator.
19226 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19228 * incpath.c (add_standard_paths): Use reconcat instead of concat
19229 where appropriate and avoid leaking memory.
19231 * opts.h: Include obstack.h.
19232 (opts_concat): New prototype.
19233 (opts_obstack): New declaration.
19234 * opts.c (opts_concat): New function.
19235 (opts_obstack): New variable.
19236 (init_options_struct): Call gcc_init_obstack on opts_obstack.
19237 (finish_options): Use opts_concat instead of concat
19238 and XOBNEWVEC instead of XNEWVEC.
19239 * opts-common.c (generate_canonical_option, decode_cmdline_option,
19240 generate_option): Likewise.
19241 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
19242 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
19245 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
19246 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
19248 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19250 PR middle-end/56461
19251 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
19253 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19255 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
19256 (arm_block_move_unaligned_straight): Likewise.
19257 (arm_adjust_block_mem): Likewise.
19259 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19262 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
19263 temp, cond and label.
19264 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
19267 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
19268 * config/c6x/c6x.h (dbx_register_map): Update declaration.
19271 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
19272 of prologue/epilogue functions.
19275 * config/tilegx/tilegx.c (tilegx_expand_prologue):
19276 Remove unused variable cfa_offset.
19277 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
19280 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
19281 type promotion to unsigned.
19284 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
19285 for HOST_WIDE_INT of 32 bit / same size as int.
19286 (arm_block_move_unaligned_straight): Likewise.
19287 (arm_adjust_block_mem): Likewise.
19290 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
19293 2013-02-26 Marek Polacek <polacek@redhat.com>
19295 PR tree-optimization/56426
19296 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
19298 2013-02-26 Richard Biener <rguenther@suse.de>
19301 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
19302 unused variable loops.
19304 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19306 PR tree-optimization/56448
19307 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
19308 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
19309 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
19310 later operands of the references, or even first operand for
19311 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
19313 PR tree-optimization/56443
19314 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
19315 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
19316 to type_for_mode langhook.
19318 2013-02-25 Matt Turner <mattst88@gmail.com>
19320 * doc/invoke.texi: Document r4700.
19322 2013-02-25 Richard Biener <rguenther@suse.de>
19324 PR tree-optimization/56175
19325 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
19327 (simplify_bitwise_binary): ... here. Also guard the conversion
19328 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
19330 2013-02-25 Catherine Moore <clm@codesourcery.com>
19333 2013-02-24 Catherine Moore <clm@codesourcery.com>
19334 Maciej W. Rozycki <macro@codesourcery.com>
19335 Tom de Vries <tom@codesourcery.com>
19336 Nathan Sidwell <nathan@codesourcery.com>
19337 Iain Sandoe <iain@codesourcery.com>
19338 Nathan Froyd <froydnj@codesourcery.com>
19339 Chao-ying Fu <fu@mips.com>
19341 * doc/extend.texi: (micromips, nomicromips, nocompression):
19342 Document new function attributes.
19343 * doc/invoke.texi (minterlink-compressed, mmicromips,
19344 m14k, m14ke, m14kec): Document new options.
19345 (minterlink-mips16): Update documentation.
19346 * doc/md.texi (ZC, ZD): Document new constraints.
19347 * configure.ac (gcc_cv_as_micromips): Check if linker
19348 supports the .set micromips directive.
19349 * configure: Regenerate.
19350 * config.in: Regenerate.
19351 * config/mips/mips-tables.opt: Regenerate.
19352 * config/mips/micromips.md: New file.
19353 * constraints.md (ZC, AD): New constraints.
19354 * config/mips/predicates.md (movep_src_register): New predicate.
19355 (movep_src_operand): New predicate.
19356 (non_volatile_mem_operand): New predicate.
19357 * config/mips/mips.md (multimem): New type.
19358 (length): Differentiate between 17-bit and 18-bit branch offsets.
19359 (MOVEP1, MOVEP2): New mode iterator.
19360 (mov_<load>l): Use ZC constraint.
19361 (mov_<load>r): Likewise.
19362 (mov_<store>l): Likewise.
19363 (mov_<store>r): Likewise.
19364 (*branch_equality<mode>_inverted): Add microMIPS support.
19365 (*branch_equality<mode>): Likewise.
19366 (*jump_absolute): Likewise.
19367 (indirect_jump_<mode>): Likewise.
19368 (tablejump_<mode>): Likewise.
19369 (<optab>_internal): Likewise.
19370 (sibcall_internal): Likewise.
19371 (sibcall_value_internal): Likewise.
19372 (prefetch): Use constraint ZD.
19373 * config/mips/mips.opt (minterlink-compressed): New option.
19374 (minterlink-mips16): Now an alias for minterlink-compressed.
19375 (mmicromips): New option.
19376 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19377 (compare_and_swap_12): Likewise.
19378 (sync_add<mode>): Likewise.
19379 (sync_<optab>_12): Likewise.
19380 (sync_old_<optab>_12): Likewise.
19381 (sync_new_<optab>_12): Likewise.
19382 (sync_nand_12): Likewise.
19383 (sync_old_nand_12): Likewise.
19384 (sync_new_nand_12): Likewise.
19385 (sync_sub<mode>): Likewise.
19386 (sync_old_add<mode>): Likewise.
19387 (sync_old_sub<mode>): Likewise.
19388 (sync_new_add<mode>): Likewise.
19389 (sync_new_sub<mode>): Likewise.
19390 (sync_<optab><mode>): Likewise.
19391 (sync_old_<optab><mode>): Likewise.
19392 (sync_new_<optab><mode>): Likewise.
19393 (sync_nand<mode>): Likewise.
19394 (sync_old_nand<mode>): Likewise.
19395 (sync_new_nand<mode>): Likewise.
19396 (sync_lock_test_and_set<mode>): Likewise.
19397 (test_and_set_12): Likewise.
19398 (atomic_compare_and_swap<mode>): Likewise.
19399 (atomic_exchange<mode>_llsc): Likewise.
19400 (atomic_fetch_add<mode>_llsc): Likewise.
19401 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19402 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19403 (umips_save_restore_pattern_p): Likewise.
19404 (umips_load_store_pair_p): Likewise.
19405 (umips_output_load_store_pair): Likewise.
19406 (umips_movep_target_p): Likewise.
19407 (umips_12bit_offset_address_p): Likewise.
19408 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19409 (mips_base_mips16): Rename this...
19410 (mips_base_compression_flags): ...to this. Update all uses.
19411 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19412 (mips_mips16_decl_p): Delete.
19413 (mips_nomips16_decl_p): Delete.
19414 (mips_get_compress_on_flags): New function.
19415 (mips_get_compress_off_flags): New function.
19416 (mips_get_compress_mode): New function.
19417 (mips_get_compress_on_name): New function.
19418 (mips_get_compress_off_name): New function.
19419 (mips_insert_attributes): Support multiple compression types.
19420 (mips_merge_decl_attributes): Likewise.
19421 (umips_12bit_offset_address_p): New function.
19422 (mips_start_function_definition): Emit .set micromips directive.
19423 (mips_call_may_need_jalx_p): New function.
19424 (mips_function_ok_for_sibcall): Add microMIPS support.
19425 (mips_print_operand_punctuation): Support short delay slots and
19427 (umips_swm_mask, umips_swm_encoding): New.
19428 (umips_build_save_restore): New function.
19429 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19430 (was_mips16_p): Remove.
19431 (old_compression_mode): New.
19432 (mips_set_compression_mode): New function.
19433 (mips_set_current_function): Add microMIPS support.
19434 (mips_option_override): Likewise.
19435 (umips_save_restore_pattern_p): New function.
19436 (umips_output_save_restore): New function.
19437 (umips_load_store_pair_p_1): New function.
19438 (umips_load_store_pair_p): New function.
19439 (umips_output_load_store_pair_1): New function.
19440 (umips_output_load_store_pair): New function.
19441 (umips_movep_target_p) New function.
19442 (mips_prepare_pch_save): Add microMIPS support.
19443 * config/mips/mips.h (TARGET_COMPRESSION): New.
19444 (TARGET_CPU_CPP_BUILTINS): Update macro
19445 to use new compression flags and to support microMIPS.
19446 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19447 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19448 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19449 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19450 (ASM_SPEC): Support mmicromips and mno-micromips.
19451 (M16STORE_REG_P): New macro.
19452 (MIPS_CALL): Support TARGET_MICROMIPS.
19453 (MICROMIPS_J): New macro.
19454 (mips_base_mips16): Rename this...
19455 (mips_base_compression_flags): ...to this.
19456 (UMIPS_12BIT_OFFSET_P): New macro.
19457 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19458 (MULTILIB_DIRNAMES): Likewise.
19460 2013-02-25 Tom de Vries <tom@codesourcery.com>
19462 PR rtl-optimization/56131
19463 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
19464 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
19465 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
19467 2013-02-25 Tobias Burnus <burnus@net-b.de>
19469 * doc/invoke.texi (-fsanitize=): Move from optimization
19470 to debugging options.
19472 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
19474 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
19476 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
19477 Alexander Monakov <amonakov@ispras.ru>
19479 PR middle-end/56077
19480 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
19481 flush pending lists also on non-jumps. Adjust comment.
19483 2013-02-24 Catherine Moore <clm@codesourcery.com>
19484 Maciej W. Rozycki <macro@codesourcery.com>
19485 Tom de Vries <tom@codesourcery.com>
19486 Nathan Sidwell <nathan@codesourcery.com>
19487 Iain Sandoe <iain@codesourcery.com>
19488 Nathan Froyd <froydnj@codesourcery.com>
19489 Chao-ying Fu <fu@mips.com>
19491 * doc/extend.texi: (micromips, nomicromips, nocompression):
19492 Document new function attributes.
19493 * doc/invoke.texi (minterlink-compressed, mmicromips,
19494 m14k, m14ke, m14kec): Document new options.
19495 (minterlink-mips16): Update documentation.
19496 * doc/md.texi (ZC, ZD): Document new constraints.
19497 * configure.ac (gcc_cv_as_micromips): Check if linker
19498 supports the .set micromips directive.
19499 * configure: Regenerate.
19500 * config.in: Regenerate.
19501 * config/mips/mips-tables.opt: Regenerate.
19502 * config/mips/micromips.md: New file.
19503 * constraints.md (ZC, AD): New constraints.
19504 * config/mips/predicates.md (movep_src_register): New predicate.
19505 (movep_src_operand): New predicate.
19506 (non_volatile_mem_operand): New predicate.
19507 * config/mips/mips.md (multimem): New type.
19508 (length): Differentiate between 17-bit and 18-bit branch offsets.
19509 (MOVEP1, MOVEP2): New mode iterator.
19510 (mov_<load>l): Use ZC constraint.
19511 (mov_<load>r): Likewise.
19512 (mov_<store>l): Likewise.
19513 (mov_<store>r): Likewise.
19514 (*branch_equality<mode>_inverted): Add microMIPS support.
19515 (*branch_equality<mode>): Likewise.
19516 (*jump_absolute): Likewise.
19517 (indirect_jump_<mode>): Likewise.
19518 (tablejump_<mode>): Likewise.
19519 (<optab>_internal): Likewise.
19520 (sibcall_internal): Likewise.
19521 (sibcall_value_internal): Likewise.
19522 (prefetch): Use constraint ZD.
19523 * config/mips/mips.opt (minterlink-compressed): New option.
19524 (minterlink-mips16): Now an alias for minterlink-compressed.
19525 (mmicromips): New option.
19526 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19527 (compare_and_swap_12): Likewise.
19528 (sync_add<mode>): Likewise.
19529 (sync_<optab>_12): Likewise.
19530 (sync_old_<optab>_12): Likewise.
19531 (sync_new_<optab>_12): Likewise.
19532 (sync_nand_12): Likewise.
19533 (sync_old_nand_12): Likewise.
19534 (sync_new_nand_12): Likewise.
19535 (sync_sub<mode>): Likewise.
19536 (sync_old_add<mode>): Likewise.
19537 (sync_old_sub<mode>): Likewise.
19538 (sync_new_add<mode>): Likewise.
19539 (sync_new_sub<mode>): Likewise.
19540 (sync_<optab><mode>): Likewise.
19541 (sync_old_<optab><mode>): Likewise.
19542 (sync_new_<optab><mode>): Likewise.
19543 (sync_nand<mode>): Likewise.
19544 (sync_old_nand<mode>): Likewise.
19545 (sync_new_nand<mode>): Likewise.
19546 (sync_lock_test_and_set<mode>): Likewise.
19547 (test_and_set_12): Likewise.
19548 (atomic_compare_and_swap<mode>): Likewise.
19549 (atomic_exchange<mode>_llsc): Likewise.
19550 (atomic_fetch_add<mode>_llsc): Likewise.
19551 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19552 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19553 (umips_save_restore_pattern_p): Likewise.
19554 (umips_load_store_pair_p): Likewise.
19555 (umips_output_load_store_pair): Likewise.
19556 (umips_movep_target_p): Likewise.
19557 (umips_12bit_offset_address_p): Likewise.
19558 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19559 (mips_base_mips16): Rename this...
19560 (mips_base_compression_flags): ...to this. Update all uses.
19561 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19562 (mips_mips16_decl_p): Delete.
19563 (mips_nomips16_decl_p): Delete.
19564 (mips_get_compress_on_flags): New function.
19565 (mips_get_compress_off_flags): New function.
19566 (mips_get_compress_mode): New function.
19567 (mips_get_compress_on_name): New function.
19568 (mips_get_compress_off_name): New function.
19569 (mips_insert_attributes): Support multiple compression types.
19570 (mips_merge_decl_attributes): Likewise.
19571 (umips_12bit_offset_address_p): New function.
19572 (mips_start_function_definition): Emit .set micromips directive.
19573 (mips_call_may_need_jalx_p): New function.
19574 (mips_function_ok_for_sibcall): Add microMIPS support.
19575 (mips_print_operand_punctuation): Support short delay slots and
19577 (umips_swm_mask, umips_swm_encoding): New.
19578 (umips_build_save_restore): New function.
19579 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19580 (was_mips16_p): Remove.
19581 (old_compression_mode): New.
19582 (mips_set_compression_mode): New function.
19583 (mips_set_current_function): Add microMIPS support.
19584 (mips_option_override): Likewise.
19585 (umips_save_restore_pattern_p): New function.
19586 (umips_output_save_restore): New function.
19587 (umips_load_store_pair_p_1): New function.
19588 (umips_load_store_pair_p): New function.
19589 (umips_output_load_store_pair_1): New function.
19590 (umips_output_load_store_pair): New function.
19591 (umips_movep_target_p) New function.
19592 (mips_prepare_pch_save): Add microMIPS support.
19593 * config/mips/mips.h (TARGET_COMPRESSION): New.
19594 (TARGET_CPU_CPP_BUILTINS): Update macro
19595 to use new compression flags and to support microMIPS.
19596 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19597 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19598 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19599 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19600 (ASM_SPEC): Support mmicromips and mno-micromips.
19601 (M16STORE_REG_P): New macro.
19602 (MIPS_CALL): Support TARGET_MICROMIPS.
19603 (MICROMIPS_J): New macro.
19604 (mips_base_mips16): Rename this...
19605 (mips_base_compression_flags): ...to this.
19606 (UMIPS_12BIT_OFFSET_P): New macro.
19607 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19608 (MULTILIB_DIRNAMES): Likewise.
19610 2013-02-24 Jakub Jelinek <jakub@redhat.com>
19613 * target-globals.c (save_target_globals): For init_reg_sets and
19614 target_reinit remporarily set this_fn_optabs to this_target_optabs.
19616 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
19618 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
19619 * config/aarch64/t-aarch64
19620 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
19622 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
19624 PR inline-asm/56148
19625 * lra-constraints.c (process_alt_operands): Reload operand
19626 conflicting with earlier clobber only if no more other conflicting
19629 2013-02-22 Jakub Jelinek <jakub@redhat.com>
19632 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
19633 if not linking a shared library.
19635 2013-02-22 Seth LaForge <sethml@google.com>
19637 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
19639 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
19641 * config/arm/arm.md (split for extendsidi): Update condition.
19642 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
19643 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
19644 (qhs_zextenddi_cstr): Likewise.
19646 2013-02-21 Jakub Jelinek <jakub@redhat.com>
19648 PR middle-end/56420
19649 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
19650 avoid signed wrapping.
19651 (expand_mult): Handle properly multiplication by
19652 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
19653 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
19654 in the compiler if coeff is HOST_WIDE_INT_MIN.
19655 (expand_divmod): Don't make ext_op1 static, change it's type to
19656 uhwi. Avoid undefined behavior in -INTVAL (op1).
19658 PR rtl-optimization/50339
19659 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
19661 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
19662 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
19663 into splitting_ashiftrt field.
19664 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
19666 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
19669 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
19671 PR middle-end/56108
19672 * trans-mem.c (execute_tm_mark): Do not expand transactions that
19673 are sure to go irrevocable.
19675 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
19677 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
19678 scalars are valid operands.
19680 2013-02-21 Martin Jambor <mjambor@suse.cz>
19682 PR tree-optimization/56310
19683 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
19684 only matching indices and non-negative final offsets.
19685 (intersect_aggregates_with_edge): Pass src_idx to
19686 agg_replacements_to_vector. Pass src_idx insstead of index to
19687 intersect_with_agg_replacements.
19689 2013-02-21 Martin Jambor <mjambor@suse.cz>
19691 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
19692 instead of hard-wired defaults.
19694 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
19696 * doc/invoke.texi (MIPS Options): Update documentation of the
19697 floating-point multiply-accumulate instruction restrictions.
19699 2013-02-21 Kostya Serebryany <kcc@google.com>
19701 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
19702 asan_shadow_offset on x86_64 linux.
19704 2013-02-21 Richard Biener <rguenther@suse.de>
19706 PR tree-optimization/56415
19708 2013-02-11 Richard Biener <rguenther@suse.de>
19710 PR tree-optimization/56273
19711 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19714 2013-02-21 Jakub Jelinek <jakub@redhat.com>
19717 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
19720 PR inline-asm/56405
19721 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
19722 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
19724 2013-02-20 Jan Hubicka <jh@suse.cz>
19726 PR tree-optimization/56265
19727 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
19728 when target is referenced for first time.
19730 2013-02-20 Richard Biener <rguenther@suse.de>
19732 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
19733 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
19734 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
19735 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
19736 not return anything.
19737 (rename_ssa_copies): Do not remove unused locals.
19738 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
19739 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
19740 * passes.c (execute_function_todo): Do not schedule unused locals
19741 removal if cleanup_tree_cfg did something.
19742 * tree-ssa-live.c (remove_unused_locals): Dump statistics
19743 about the number of removed locals.
19745 2013-02-20 Richard Biener <rguenther@suse.de>
19747 PR tree-optimization/56398
19748 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
19750 2013-02-20 Martin Jambor <mjambor@suse.cz>
19752 PR tree-optimization/55334
19753 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
19754 restricted pointers to arrays.
19756 2013-02-20 Richard Biener <rguenther@suse.de>
19757 Jakub Jelinek <jakub@redhat.com>
19759 PR tree-optimization/56396
19760 * tree-ssa-ccp.c (n_const_val): New static variable.
19761 (get_value): Return NULL for SSA names we don't have a lattice
19763 (ccp_initialize): Initialize n_const_val.
19764 * tree-ssa-copy.c (n_copy_of): New static variable.
19765 (init_copy_prop): Initialize n_copy_of.
19766 (get_value): Return NULL_TREE for SSA names we don't have a
19769 2013-02-20 Martin Jambor <mjambor@suse.cz>
19771 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
19773 2013-02-20 Richard Biener <rguenther@suse.de>
19775 * genpreds.c (write_lookup_constraint): Do not compare first
19776 letter of the constraint again.
19778 2013-02-20 Richard Biener <rguenther@suse.de>
19780 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
19782 (get_use_iv_cost): Terminate hashtable walk when coming across
19785 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
19787 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
19788 reassociation for avx2 targets.
19790 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19792 * config/microblaze/microblaze.c: microblaze_has_clz = 0
19793 Add version check for v8.10.a to enable microblaze_has_clz
19794 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
19795 version and TARGET_PATTERN_COMPARE check
19796 * config/microblaze/microblaze.md: New clzsi2 instruction
19798 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19800 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
19801 function before branching.
19803 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19805 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
19807 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
19809 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19811 PR middle-end/55889
19812 * sel-sched.c: Include ira.h.
19813 (implicit_clobber_conflict_p): New function.
19814 (moveup_expr): Use it.
19815 * Makefile.in (sel-sched.o): Depend on ira.h.
19817 2013-02-19 Richard Biener <rguenther@suse.de>
19819 PR tree-optimization/56384
19820 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
19821 (vn_hash_type): Split out from ...
19822 (vn_hash_constant_with_type): ... here.
19823 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
19824 (vn_phi_eq): Compare types from vn_phi_s structure.
19825 (vn_phi_lookup): Populate vn_phi_s type.
19826 (vn_phi_insert): Likewise.
19828 2013-02-19 Jakub Jelinek <jakub@redhat.com>
19830 PR tree-optimization/56350
19831 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
19832 if haven't found reduction or nested cycle operand, rather than
19833 asserting we must find it.
19835 PR tree-optimization/56381
19836 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
19839 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
19840 Jakub Jelinek <jakub@redhat.com>
19843 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
19844 (swap_optab_enable): Same.
19845 (init_all_optabs): Use argument instead of global.
19846 * tree.h (struct tree_optimization_option): New field target_optabs.
19847 * expr.h (init_all_optabs): Add argument to prototype.
19848 (TREE_OPTIMIZATION_OPTABS): New.
19849 (save_optabs_if_changed): Protoize.
19850 * optabs.h: Declare this_fn_optabs.
19851 * optabs.c (save_optabs_if_changed): New.
19852 Declare this_fn_optabs.
19853 (init_optabs): Add argument to init_all_optabs() call.
19854 * function.c (invoke_set_current_function_hook): Handle per
19856 * function.h (struct function): New field optabs.
19857 * config/mips/mips.c (mips_set_mips16_mode): Handle when
19858 optimization_current_node has changed.
19859 * target-globals.h (save_target_globals_default_opts): Protoize.
19860 * target-globals.c (save_target_globals_default_opts): New.
19862 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19865 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
19866 registers %fr12 and %fr12R as call used.
19869 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
19870 and HImode, require all displacements to be an integer multiple of
19872 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
19873 only allow QImode and HImode when reload is in progress and strict is
19874 true. Likewise for symbolic addresses. Use base14_operand to check
19875 displacements in REG+BASE addresses.
19877 2013-02-18 Richard Biener <rguenther@suse.de>
19879 PR tree-optimization/56366
19880 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19881 handle sign-conversion of outer-loop initial induction value.
19883 2013-02-18 Richard Biener <rguenther@suse.de>
19885 PR middle-end/56349
19886 * cfghooks.c (merge_blocks): If we merge a latch into another
19887 block adjust references to it.
19888 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
19889 (verify_loop_structure): Verify that a recorded latch is in fact
19892 2013-02-18 Richard Biener <rguenther@suse.de>
19894 PR tree-optimization/56321
19895 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
19896 order SSA name release and virtual operand unlinking.
19898 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19900 * config/microblaze/microblaze.md (save_stack_block): Define.
19901 (restore_stack_block): Likewise.
19903 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19905 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
19906 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
19907 * config/microblaze/microblaze.c (microblaze_option_override):
19908 Bail out early for PIC modes when target does not support PIC.
19910 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19912 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
19913 Replace with a microblaze version.
19914 (microblaze_trampoline_init): Adapt for microblaze.
19915 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
19918 2013-02-16 Jakub Jelinek <jakub@redhat.com>
19919 Dodji Seketeli <dodji@redhat.com>
19922 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
19923 (instrument_mem_region_access): Do not forget to always put
19924 instrumentation of the of 'base' and 'base + len' in a "if (len !=
19925 0) statement, even for cases where either 'base' or 'base + len'
19926 are not instrumented -- because they have been previously
19927 instrumented. Simplify the logic by putting all the statements
19928 instrument 'base + len' inside a sequence, and then insert that
19929 sequence right before the current insertion point. Then, to
19930 instrument 'base + len', just get an iterator on that statement.
19931 And do not forget to update the pointer to iterator the function
19932 received as argument.
19934 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
19936 PR rtl-optimization/56348
19937 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
19939 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
19941 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
19942 (clean_graph_dump_file): Pass base to start_graph_dump.
19944 2013-02-14 Richard Henderson <rth@redhat.com>
19947 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
19949 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
19951 * collect2-aix.h: Define F_LOADONLY.
19953 2013-02-14 Richard Biener <rguenther@suse.de>
19956 * varasm.c (output_constant_def_1): Get the decl representing
19957 the constant as argument.
19958 (output_constant_def): Wrap output_constant_def_1.
19959 (make_decl_rtl): Use output_constant_def_1 with the decl
19960 representing the constant.
19961 (build_constant_desc): Optionally re-use a decl already
19962 representing the constant.
19963 (tree_output_constant_def): Adjust.
19965 2013-02-14 Dodji Seketeli <dodji@redhat.com>
19968 * asan.c (instrument_builtin_call): Really put the length of the
19969 second source argument into src1_len.
19971 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19973 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
19974 argument. If it is false, don't create edge from then_bb to
19976 (insert_if_then_before_iter): Pass true to it.
19977 (build_check_stmt): Pass false to it.
19978 (transform_statements): Flush hash table only on extended basic
19979 block boundaries, rather than at the beginning of every bb.
19980 Don't flush hash table on nonfreeing_call_p calls.
19981 * tree-flow.h (nonfreeing_call_p): New prototype.
19982 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
19984 2013-02-13 David S. Miller <davem@davemloft.net>
19986 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
19988 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
19991 * ira.c (max_regno_before_ira): Move from ...
19993 (fix_reg_equiv_init): Use max_regno_before_ira instead of
19996 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19998 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20000 2013-02-13 Richard Biener <rguenther@suse.de>
20003 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20004 globals in MEM_REFs.
20006 2013-02-13 Richard Biener <rguenther@suse.de>
20008 * loop-init.c (loop_optimizer_init): Clear loop state when
20009 re-initializing preserved loops.
20010 * loop-unswitch.c (unswitch_single_loop): Return whether
20011 we unswitched the loop. Do not verify loop state here.
20012 (unswitch_loops): When we unswitched a loop discover new loops.
20014 2013-02-13 Kostya Serebryany <kcc@google.com>
20016 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20018 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20020 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20022 Avoid instrumenting duplicated memory access in the same basic block
20023 * Makefile.in (asan.o): Add new dependency on hash-table.h
20024 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20025 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20026 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20027 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20028 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20029 (get_mem_ref_of_assignment): New functions.
20030 (get_mem_refs_of_builtin_call): Extract from
20031 instrument_builtin_call and tweak a little bit to make it fit with
20033 (instrument_builtin_call): Use the new
20034 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20035 of is_gimple_builtin_call.
20036 (instrument_derefs, instrument_mem_region_access): Insert the
20037 instrumented memory reference into the hash table.
20038 (maybe_instrument_assignment): Renamed instrument_assignment into
20039 this, and change it to advance the iterator when instrumentation
20040 actually happened and return true in that case. This makes it
20041 homogeneous with maybe_instrument_assignment, and thus give a
20042 chance to callers to be more 'regular'.
20043 (transform_statements): Clear the memory reference hash table
20044 whenever we enter a new BB, when we cross a function call, or when
20045 we are done transforming statements. Use
20046 maybe_instrument_assignment instead of instrumentation. No more
20047 need to special case maybe_instrument_assignment and advance the
20048 iterator after calling it; it's now handled just like
20049 maybe_instrument_call. Update comment.
20051 2013-02-13 Richard Biener <rguenther@suse.de>
20053 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
20054 Fix loop discovery code.
20056 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
20058 PR inline-asm/56148
20059 * lra-constraints.c (process_alt_operands): Match early clobber
20060 operand with itself. Check conflicts with earlyclobber only if
20061 the operand is not reloaded. Prefer to reload conflicting operand
20062 if earlyclobber and matching operands are the same.
20064 2013-02-12 Richard Biener <rguenther@suse.de>
20067 * lto-streamer-out.c (write_symbol): Do not output symbols
20068 for hard register variables.
20070 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
20073 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
20074 (umulsidi3_insn, mulsidi3_insn): New insns.
20076 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
20078 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
20079 (struct tune_params): Add vec_costs field.
20080 * config/arm/arm.c (arm_builtin_vectorization_cost)
20081 (arm_add_stmt_cost): New functions.
20082 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
20083 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
20084 (arm_default_vec_cost): New struct of type cpu_vec_costs.
20085 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
20086 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
20087 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
20088 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
20090 2013-02-12 Richard Biener <rguenther@suse.de>
20093 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
20094 decls again if possible.
20096 2013-02-12 Richard Biener <rguenther@suse.de>
20098 PR middle-end/56288
20099 * tree-ssa.c (verify_ssa_name): Fix check, move
20100 SSA_NAME_IN_FREE_LIST check up.
20102 2013-02-12 Jakub Jelinek <jakub@redhat.com>
20103 Steven Bosscher <steven@gcc.gnu.org>
20105 PR rtl-optimization/56151
20106 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
20107 equal to op0 or op1, and last_insn pattern is CODE operation
20108 with MEM dest and one of the operands matches that MEM.
20110 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
20112 * doc/extend.texi: Document Function Multiversioning and "default"
20113 parameter string to target attribute.
20114 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
20115 target attribute parameter is "default".
20116 (ix86_compare_version_priority): Remove checks for target attribute.
20117 (ix86_mangle_function_version_assembler_name): Change error to sorry.
20118 Remove check for target attribute equal to NULL. Add assert.
20119 (ix86_generate_version_dispatcher_body): Change error to sorry.
20121 2013-02-11 Iain Sandoe <iain@codesourcery.com>
20122 Jack Howarth <howarth@bromo.med.uc.edu>
20123 Patrick Marlier <patrick.marlier@gmail.com>
20126 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
20127 define ENDFILE_SPEC as TM_DESTRUCTOR.
20128 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
20130 2013-02-11 Alexander Potapenko <glider@google.com>
20131 Jack Howarth <howarth@bromo.med.uc.edu>
20132 Jakub Jelinek <jakub@redhat.com>
20135 * config/darwin.c (cdtor_record): Rename ctor_record.
20136 (sort_cdtor_records): Rename sort_ctor_records.
20137 (finalize_dtors): New routine to sort destructors by
20138 priority before use in assemble_integer.
20139 (machopic_asm_out_destructor): Use finalize_dtors if needed.
20141 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
20143 PR rtl-optimization/56275
20144 * simplify-rtx.c (avoid_constant_pool_reference): Check that
20145 offset is non-negative and less than cmode size before
20146 calling simplify_subreg.
20148 2013-02-11 Richard Biener <rguenther@suse.de>
20150 PR tree-optimization/56264
20151 * cfgloop.h (fix_loop_structure): Adjust prototype.
20152 * loop-init.c (fix_loop_structure): Return the number of
20153 newly discovered loops.
20154 * tree-cfgcleanup.c (repair_loop_structures): When new loops
20155 are discovered, do a full loop-closed SSA rewrite.
20157 2013-02-11 Richard Biener <rguenther@suse.de>
20159 PR tree-optimization/56273
20160 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20162 (check_array_ref): Fix missing newline in dumps.
20163 (search_for_addr_array): Likewise.
20165 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
20167 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
20169 2013-02-09 Jakub Jelinek <jakub@redhat.com>
20172 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
20174 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
20176 PR rtl-optimization/56246
20177 * lra-constraints.c (simplify_operand_subreg): Try to reuse
20179 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
20180 constraints are satisfied.
20182 2013-02-08 Jeff Law <law@redhat.com>
20185 * emit-rtl.c (reg_is_parm_p): New function.
20186 * regs.h (reg_is_parm_p): New prototype.
20187 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
20188 callee-clobbered registers.
20190 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
20193 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
20194 If there is no implicit builtin declaration, just return NULL.
20196 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20198 * config/i386/sse.md (FMAMODEM): New mode iterator.
20199 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
20200 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
20202 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20204 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
20205 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
20206 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
20208 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20210 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
20211 (microblaze*-*-elf): Likewise.
20212 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
20214 * config/microblaze/microblaze-c.c: Add builtin defines for
20215 _LITTLE_ENDIAN and _BIG_ENDIAN.
20216 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
20217 add to TARGET_DEFAULT flags.
20218 Expand ASM_SPEC and LINK_SPEC.
20219 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
20220 * config/microblaze/microblaze.md: Update extendsidi2 and
20221 movdi_internal instructions to use low-order / high-order reg
20223 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
20224 options and inversemask / mask of LITTLE_ENDIAN.
20225 * config/microblaze/t-microblaze: Expand multilib options to
20226 include mlittle-endian (le) and update exceptions patterns.
20228 2013-02-08 Jakub Jelinek <jakub@redhat.com>
20230 PR rtl-optimization/56195
20231 * lra-constraints.c (get_reload_reg): Don't reuse regs
20232 if they have smaller mode than requested, if they have
20233 wider mode than requested, try to return a SUBREG.
20235 PR tree-optimization/56250
20236 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
20237 if type is unsigned and code isn't MULT_EXPR.
20239 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20241 PR tree-optimization/56064
20242 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
20243 bits according to mode.
20244 * fixed-value.h (fixed_from_double_int)
20245 (const_fixed_from_double_int): Adjust comments.
20247 2013-02-08 Richard Biener <rguenther@suse.de>
20250 * lto-streamer.h (struct data_in): Remove current_file, current_line
20251 and current_col members.
20252 * lto-streamer-out.c (lto_output_location): Stream changed bits
20253 en-block for efficiency.
20254 * lto-streamer-in.c (clear_line_info): Remove.
20255 (lto_input_location): Cache current file, line and column
20256 globally via local statics. Read changed bits en-block.
20257 (input_function): Do not call clear_line_info.
20258 (lto_read_body): Likewise.
20259 (lto_input_toplevel_asms): Likewise.
20261 2013-02-08 Michael Matz <matz@suse.de>
20263 PR tree-optimization/52448
20264 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
20265 (nt_call_phase): New static.
20266 (add_or_mark_expr): Only mark accesses with newer phase than any
20268 (nonfreeing_call_p): New.
20269 (nt_init_block): Update nt_call_phase, mark blocks as visited.
20270 (nt_fini_block): Keep blocks marked as visited.
20271 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
20273 2013-02-08 Richard Biener <rguenther@suse.de>
20275 * ira.c (ira): Free broken dominator information.
20277 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20279 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
20281 2013-02-08 Marek Polacek <polacek@redhat.com>
20283 * cfgloop.c (verify_loop_structure): Add more checking of headers.
20285 2013-02-08 Richard Biener <rguenther@suse.de>
20287 PR middle-end/56181
20288 * cfgloop.h (flow_loops_find): Adjust.
20289 (bb_loop_header_p): Declare.
20290 * cfgloop.c (bb_loop_header_p): New function split out from ...
20291 (flow_loops_find): ... here. Adjust function signature,
20292 support incremental loop structure update.
20293 (verify_loop_structure): Cleanup. Verify a loop is a loop.
20294 * cfgloopmanip.c (fix_loop_structure): Move ...
20295 * loop-init.c (fix_loop_structure): ... here.
20296 (apply_loop_flags): Split out from ...
20297 (loop_optimizer_init): ... here.
20298 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
20299 in incremental mode, only remove dead loops here.
20301 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20304 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
20305 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
20306 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
20307 (*round<mode>3.libgcc): New insns for fixed-modes.
20308 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
20309 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
20310 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
20311 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
20312 implementations. Define to __builtin_avr_absFX,
20313 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
20314 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
20315 __builtin_avr_countlsFX, respectively.
20316 * config/avr/avr-c.c (target.h): Include it.
20317 (enum avr_builtin_id): New enum.
20318 (avr_resolve_overloaded_builtin): New static function.
20319 (avr_register_target_pragmas): Use it to set
20320 targetm.resolve_overloaded_builtin.
20321 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
20322 tree nodes used by DEF_BUILTIN.
20323 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
20324 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
20325 <AVR_BUILTIN_xxBITS>: Same.
20327 2013-02-08 Richard Biener <rguenther@suse.de>
20329 * cfgloop.c (verify_loop_structure): Properly handle
20330 a loop exiting to another loop header.
20331 * ira-int.h (ira_loops): Remove.
20332 * ira.c (ira_loops): Remove.
20333 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
20334 (do_reload): Use loop_optimizer_finalize.
20335 * ira-build.c (create_loop_tree_nodes): Use get_loops and
20336 number_of_loops to access the loop tree.
20337 (more_one_region_p): Likewise.
20338 (finish_loop_tree_nodes): Likewise.
20339 (rebuild_regno_allocno_maps): Likewise.
20340 (mark_loops_for_removal): Likewise.
20341 (mark_all_loops_for_removal): Likewise.
20342 (remove_unnecessary_regions): Likewise.
20343 (ira_build): Likewise.
20344 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
20346 2013-02-08 Richard Biener <rguenther@suse.de>
20348 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
20349 * ipa-pure-const.c (analyze_function): Avoid calling
20350 mark_irreducible_loops twice.
20351 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
20353 2013-02-07 David S. Miller <davem@davemloft.net>
20355 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
20357 * var-tracking.c (vt_add_function_parameter): Test the presence of
20358 HAVE_window_save properly and do not remap argument registers when
20359 we have a leaf function.
20361 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
20364 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
20366 * config/i386/i386.c (ix86_print_operand): Ditto.
20368 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
20370 * lra-constraints.c (process_alt_operands): Fix recently added comment.
20372 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
20374 PR rtl-optimization/56225
20375 * lra-constraints.c (process_alt_operands): Check that reload hard
20376 reg can hold value for strict_low_part.
20378 2013-02-07 Jakub Jelinek <jakub@redhat.com>
20381 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
20382 dwarf2out_end_function.
20383 (in_first_function_p, maybe_at_text_label_p,
20384 first_loclabel_num_not_at_text_label): New variables.
20385 (dwarf2out_var_location): In the first function find out
20386 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
20387 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
20390 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
20392 PR rtl-optimization/56178
20393 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
20394 SUBREG of a register. Tidy up related block of code.
20395 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
20396 note if the source is a register or a SUBREG of a register.
20398 2013-02-07 Jakub Jelinek <jakub@redhat.com>
20401 * config/rs6000/rs6000.md (ptrm): New mode attr.
20402 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
20403 call_value_indirect_aix<pttrsize>,
20404 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
20407 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20409 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
20410 if -bnortl. Convert to strcmp and strncmp.
20412 2013-02-07 Alan Modra <amodra@gmail.com>
20415 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
20416 addresses won't wrap when offsetting.
20417 (rs6000_secondary_reload): Provide secondary reloads needed for
20418 wrapping LO_SUM addresses.
20420 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
20422 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
20423 MACH, just __MACH__.
20425 2013-02-06 Richard Biener <rguenther@suse.de>
20427 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
20428 instead of calling fix_loop_structure.
20430 2013-02-06 Jakub Jelinek <jakub@redhat.com>
20432 PR middle-end/56217
20433 * omp-low.c (use_pointer_for_field): Return false if
20434 lower_send_shared_vars doesn't generate any copy-out code.
20436 2013-02-06 Tom de Vries <tom@codesourcery.com>
20438 PR rtl-optimization/56131
20439 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
20440 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
20441 of the label is NULL. Add comment.
20443 2013-02-05 Jakub Jelinek <jakub@redhat.com>
20445 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
20448 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
20449 (STATIC_LIBTSAN_LIBS): Likewise.
20450 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
20451 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
20452 is defined, don't add anything else beyond that.
20453 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
20454 (LINK_COMMAND_SPEC): Use them.
20456 PR tree-optimization/56205
20457 * tree-stdarg.c (check_all_va_list_escapes): Return true if
20458 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
20459 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
20461 2013-02-05 Richard Biener <rguenther@suse.de>
20463 PR tree-optimization/53342
20464 PR tree-optimization/53185
20465 * tree-vectorizer.h (vect_check_strided_load): Remove.
20466 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
20467 not disallow peeling for vectorized strided loads.
20468 (vect_check_strided_load): Make static and simplify.
20469 (vect_analyze_data_refs): Adjust.
20470 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
20471 correctly when vectorizing strided loads.
20473 2013-02-05 Richard Biener <rguenther@suse.de>
20475 * doc/install.texi: Refer to ISL, not PPL.
20477 2013-02-05 Jan Hubicka <jh@suse.cz>
20479 PR tree-optimization/55789
20480 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
20482 2013-02-05 Jan Hubicka <jh@suse.cz>
20484 PR tree-optimization/55789
20485 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
20486 the dead call anyway.
20488 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
20491 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
20493 2013-02-04 Alexander Potapenko <glider@google.com>
20494 Jack Howarth <howarth@bromo.med.uc.edu>
20495 Jakub Jelinek <jakub@redhat.com>
20498 * config/darwin.c (sort_ctor_records): Stabilized qsort
20499 on constructor priority by using original position.
20500 (finalize_ctors): New routine to sort constructors by
20501 priority before use in assemble_integer.
20502 (machopic_asm_out_constructor): Use finalize_ctors if needed.
20504 2013-02-04 Jakub Jelinek <jakub@redhat.com>
20507 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
20508 about visibility on artificial decls.
20509 * config/sol2.c (solaris_assemble_visibility): Likewise.
20511 2013-02-04 Kai Tietz <ktietz@redhat.com>
20514 * config/i386/i386.c (function_value_ms_64): Add additional valtype
20515 argument and improve checking of return-argument types for 16-byte
20517 (ix86_function_value_1): Add additional valtype argument on call
20518 of function_value_64.
20519 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
20520 handling infunction_value_64 function.
20522 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
20524 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
20526 2013-02-04 Richard Biener <rguenther@suse.de>
20528 PR tree-optimization/56188
20529 * tree-ssa-structalias.c (label_visit): Consider case with
20530 initially non-empty points-to set.
20531 (perform_var_substitution): Dump node mapping and clean up.
20533 2013-02-04 Richard Guenther <rguenther@suse.de>
20536 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
20537 node prevail as last resort.
20538 (lto_symtab_merge_decls): Remove guard on LTRANS here.
20539 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
20541 2013-02-04 Richard Biener <rguenther@suse.de>
20543 PR tree-optimization/56113
20544 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
20546 (equiv_class_lookup_or_add): ... this.
20547 (label_visit): Adjust and fix error in previous patch.
20548 (perform_var_substitution): Adjust.
20550 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
20552 * config/sh/divtab.c: Fix formatting and comments throughout the file.
20553 * config/sh/sh4-300.md: Likewise.
20554 * config/sh/sh4a.md: Likewise.
20555 * config/sh/constraints.md: Likewise.
20556 * config/sh/sh.md: Likewise.
20557 * config/sh/netbsd-elf.h: Likewise.
20558 * config/sh/predicates.md: Likewise.
20559 * config/sh/sh-protos.h: Likewise.
20560 * config/sh/ushmedia.h: Likewise.
20561 * config/sh/linux.h: Likewise.
20562 * config/sh/sh.c: Likewise.
20563 * config/sh/superh.h: Likewise.
20564 * config/sh/elf.h: Likewise.
20565 * config/sh/sh4.md: Likewise.
20566 * config/sh/sh.h: Likewise.
20568 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20570 * config/pa/constraints.md: Adjust unused letters. Change "T"
20571 constraint to match_test floating_point_store_memory_operand().
20572 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
20573 (base14_operand): New.
20574 (floating_point_store_memory_operand): New.
20575 (integer_store_memory_operand): Revise to use base14_operand and
20576 reg_plus_base_memory_operand.
20577 (move_dest_operand): Allow symbolic_memory_operands.
20578 (symbolic_memory_operand): Check for LO_SOM.
20579 (symbolic_operand): Change default case to break.
20580 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
20581 CONST_DOUBLE values to be reloaded by putting them into memory when
20582 the destination is a floating point register.
20583 (movdf): Remove code to handle CONST_DOUBLE.
20585 (reload_indf_r1): New.
20586 (reload_insf_r1): New.
20587 Consistently use "Q" and "T" constraints with integer and floating
20588 point move instructions, respectively.
20589 (movdi): Remove FAIL.
20590 Change predicate for source operand unamed DImode move from
20591 general_operand to move_src_operand.
20592 (umulsidi3): Change predicate for destination operand to
20594 Likewise for similar unamed patterns.
20595 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
20596 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
20597 (hppa_legitimize_address): Simplify mask calculation.
20598 (pa_emit_move_sequence): Revised handling of secondary reloads from
20599 REG+D addresses for floating point loads and stores. Directly handle
20600 loading CONST0_RTX (mode) to a floating point register.
20601 (pa_secondary_reload): Handle reloading DF and SFmode constant values
20602 to floating point registers. Don't restrict secondary reloads to
20603 floating point registers to integer modes. Revise some comments and
20605 (TARGET_LEGITIMATE_ADDRESS_P): Define.
20606 (pa_legitimate_address_p): New.
20607 (pa_legitimize_reload_address): New.
20608 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
20609 (STRICT_REG_OK_FOR_BASE_P): New.
20610 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
20611 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
20613 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
20614 Andrew Dixie <andrewd@gentrack.com>
20616 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
20619 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
20621 * expmed.c (extract_bit_field_1): Pass the full width of the
20622 structure to get_best_reg_extraction_insn.
20624 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
20627 * configure.ac (use_cxa_atexit): Add AIX.
20628 * configure: Regenerate.
20630 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
20632 2013-02-01 Jakub Jelinek <jakub@redhat.com>
20635 * final.c (need_profile_function): New variable.
20636 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
20637 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
20638 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
20639 notes, targetm.asm_out.function_prologue doesn't emit anything,
20640 HAVE_prologue and profiler should be emitted before prologue,
20641 set need_profile_function instead of emitting it.
20642 (final_scan_insn): If need_profile_function, emit
20643 profile_function on the first NOTE_INSN_BASIC_BLOCK or
20644 NOTE_INSN_FUNCTION_BEG note.
20646 2013-02-01 Richard Henderson <rth@redhat.com>
20648 * config/rs6000/rs6000.md (smulditi3): New.
20651 * config/alpha/alpha.md (umulditi3): New.
20653 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
20655 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
20656 (ASM_OUTPUT_ALIGNED_LOCAL): New.
20658 2013-02-01 Richard Biener <rguenther@suse.de>
20660 PR tree-optimization/56113
20661 * tree-ssa-structalias.c (label_visit): Reduce work for
20662 single-predecessor nodes.
20664 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
20666 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
20667 range isn't testing for zero.
20669 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
20671 PR middle-end/56113
20672 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
20674 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
20675 Nick Clifton <nickc@redhat.com>
20677 * config/v850/constraints.md (Q): Define as a memory constraint.
20678 * config/v850/predicates.md (label_ref_operand): New predicate.
20679 (e3v5_shift_operand): New predicate.
20680 (ior_operator): New predicate.
20681 * config/v850/t-v850: Add e3v5 multilib.
20682 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
20683 (v850_gen_movdi): Prototype.
20684 * config/v850/v850.c: Add support for e3v5 architecture.
20685 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
20687 (construct_save_jarl): Add e3v5 long JARL support.
20688 (v850_adjust_insn_length): New function. Adjust length of call
20689 insns when using e3v5 instructions.
20690 (v850_gen_movdi): New function: Generate instructions to move a
20692 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
20693 (CPP_SPEC): Define __v850e3v5__ as appropriate.
20694 (TARGET_USE_FPU): Enable for e3v5.
20695 (CONST_OK_FOR_W): New macro.
20696 (ADJUST_INSN_LENGTH): Define.
20697 * config/v850/v850.md (UNSPEC_LOOP): Define.
20698 (attr cpu): Add v850e3v5.
20699 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
20700 (movdi): New pattern.
20701 (movdi_internal): New pattern.
20702 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
20703 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
20704 (cstoresf4): Likewise.
20705 (cstoredf4): Likewise.
20706 (insv): New pattern.
20707 (rotlso3_a): New pattern.
20708 (rotlsi3_b): New pattern
20709 (rotlsi3_v850e3v5): New pattern.
20710 (doloop_begin): New pattern.
20711 (fix_loop_counter): New pattern.
20712 (doloop_end): New pattern.
20713 (branch_normal): Add e3v5 long branch support.
20714 (branch_invert): Likewise.
20715 (branch_z_normal): Likewise.
20716 (branch_z_invert): Likewise.
20717 (branch_nz_normal): Likewise.
20718 (branch_nz_invert): Likewise.
20719 (call_internal_short): Add e3v5 register-indirect JARL support.
20720 (call_internal_long): Likewise.
20721 (call_value_internal_short): Likewise.
20722 (call_value_internal_long): Likewise.
20723 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
20724 (mloop): New option.
20725 * config.gcc: Add support for configuring v840e3v5 target.
20726 * doc/invoke.texi: Document new v850 specific command line options.
20728 2013-01-31 Paul Koning <ni1d@arrl.net>
20732 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
20733 children if parent is a class.
20734 (prune_unused_types_prune): Don't add DW_AT_declaration.
20736 2013-01-31 Richard Biener <rguenther@suse.de>
20738 PR tree-optimization/56157
20739 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
20740 match up operand with SLP child.
20742 2013-01-31 Jason Merrill <jason@redhat.com>
20745 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
20746 parameters the first time.
20747 (gen_scheduled_generic_parms_dies): Check completeness here.
20749 2013-01-31 Richard Biener <rguenther@suse.de>
20751 PR middle-end/53073
20752 * common.opt (faggressive-loop-optimizations): New flag,
20753 enabled by default.
20754 * doc/invoke.texi (faggressive-loop-optimizations): Document.
20755 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
20756 infer_loop_bounds_from_undefined by it.
20758 2013-01-31 Richard Biener <rguenther@suse.de>
20760 PR tree-optimization/56150
20761 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
20762 visit virtual operands.
20763 (find_uses_to_rename_bb): Likewise.
20765 2013-01-31 Richard Biener <rguenther@suse.de>
20767 PR tree-optimization/56150
20768 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
20769 mixed store non-store stmts.
20771 2013-01-30 Jakub Jelinek <jakub@redhat.com>
20774 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
20775 LIBASAN_EARLY_SPEC is defined.
20776 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
20777 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
20779 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
20782 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
20783 invalid args instead of ICEing on it.
20784 (ix86_valid_target_attribute_tree): Return error_mark_node if
20785 ix86_valid_target_attribute_inner_p failed.
20786 (ix86_valid_target_attribute_p): Return false only if
20787 ix86_valid_target_attribute_tree returned error_mark_node. Allow
20788 target("default") attribute.
20789 (sorted_attr_string): Change argument from const char * to tree,
20790 merge in all target attribute arguments rather than just one.
20791 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
20792 instead of free. Avoid using strcat.
20793 (ix86_mangle_function_version_assembler_name): Mangle
20794 target("default") as if no target attribute is present. Adjust
20795 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
20796 instead of xmalloc and XDELETEVEC instead of free.
20797 (ix86_function_versions): Don't return true if one of the decls
20798 doesn't have target attribute. If they don't and one of the decls
20799 is DECL_FUNCTION_VERSIONED, report an error. Adjust
20800 sorted_attr_string caller. Use XDELETEVEC instead of free.
20801 (ix86_supports_function_versions): Remove.
20802 (make_name): Fix up formatting.
20803 (make_dispatcher_decl): Remove resolver_name and its initialization.
20804 Avoid leaking memory.
20805 (is_function_default_version): Return true if there is
20806 target("default") attribute rather than no target attribute at all.
20807 (make_resolver_func): Avoid leaking memory.
20808 (ix86_generate_version_dispatcher_body): Likewise.
20809 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
20810 * target.def (supports_function_versions): Remove.
20811 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
20812 * doc/tm.texi: Regenerated.
20814 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
20816 PR rtl-optimization/56144
20817 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
20818 for values with side effects.
20820 2013-01-30 Richard Biener <rguenther@suse.de>
20822 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
20823 (sparseset_pop): Likewise.
20824 * cfganal.c (compute_idf): Likewise. Increase work-stack size
20825 to be able to use quick_push in the worker loop.
20827 2013-01-30 Marek Polacek <polacek@redhat.com>
20829 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
20831 2013-01-30 Richard Biener <rguenther@suse.de>
20834 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
20836 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
20838 PR tree-optimization/56064
20839 * fixed-value.c (fixed_from_double_int): New function.
20840 * fixed-value.h (fixed_from_double_int): New prototype.
20841 (const_fixed_from_double_int): New static inline function.
20842 * fold-const.c (native_interpret_fixed): New static function.
20843 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
20844 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
20845 (native_encode_fixed): New static function.
20846 (native_encode_expr) <FIXED_CST>: Use it.
20847 (native_interpret_int): Move double_int worker code to...
20848 * double-int.c (double_int::from_buffer): ...this new static method.
20849 * double-int.h (double_int::from_buffer): Prototype it.
20851 2013-01-30 Richard Biener <rguenther@suse.de>
20853 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
20854 New pointer-map and obstack.
20855 (init_alias_vars): Allocate pointer-map and obstack.
20856 (delete_points_to_sets): Free them.
20857 (find_what_var_points_to): Cache result.
20858 (find_what_p_points_to): Adjust for changed interface of
20859 find_what_var_points_to.
20860 (compute_points_to_sets): Likewise.
20861 (ipa_pta_execute): Likewise.
20863 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20865 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
20866 * configure: Regenerate.
20867 * config.in: Regenerate.
20868 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
20869 #nobits/#progbits if supported.
20871 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
20874 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
20875 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
20876 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
20878 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20880 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
20881 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
20883 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20885 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
20886 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
20888 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20890 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
20892 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
20893 * config/arm/cortex-a7.md: New bypasses using
20894 arm_mac_accumulator_is_result.
20896 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20898 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
20899 (cortex_a7_neon_mla): Likewise.
20900 (cortex_a7_fpfmad): New reservation.
20901 (cortex_a7_fpmacs): Use ffmas and update required units.
20902 (cortex_a7_fpmuld): Update required units and latency.
20903 (cortex_a7_fpmacd): Likewise.
20904 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
20905 (cortex_a7_neon). Likewise.
20906 (bypass) Update participating units.
20908 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20910 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
20911 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
20913 * config/arm/vfp11.md (vfp_farith): Use ffmas.
20914 (vfp_fmul): Use ffmad.
20915 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
20916 (cortex_r4_fmacd): Use ffmad.
20917 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
20918 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
20919 (cortex_a9_fmacd): Use ffmad.
20920 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
20921 (cortex_a8_vfp_macd): Use ffmad.
20922 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
20923 (cortex_a5_fpmacd): Use ffmad.
20924 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
20925 (cortex_a15_vfp_macd): Use ffmad.
20926 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
20928 2013-01-29 Jason Merrill <jason@redhat.com>
20931 * varasm.c (default_assemble_visibility): Don't warn about
20932 visibility on artificial decls.
20934 2013-01-29 Richard Biener <rguenther@suse.de>
20936 PR tree-optimization/56113
20937 * tree-ssa-structalias.c (equiv_class_lookup): Also return
20939 (label_visit): Free duplicate bitmaps and record the leader instead.
20940 (perform_var_substitution): Adjust.
20942 2013-01-29 Richard Biener <rguenther@suse.de>
20944 PR tree-optimization/55270
20945 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
20946 the CFG, schedule loops for fixup.
20948 2013-01-29 Nick Clifton <nickc@redhat.com>
20950 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
20953 2013-01-28 Leif Ekblad <leif@rdos.net>
20955 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
20956 * config/i386/i386.h (TARGET_RDOS): New macro.
20957 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
20958 * config/i386/i386.c (ix86_option_override_internal): For 64bit
20959 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
20960 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
20961 DEFAULT_LARGE_SECTION_THRESHOLD.
20962 * config/i386/i386.md (R14_REG, R15_REG): New constants.
20963 * config/i386/rdos.h: New file.
20964 * config/i386/rdos64.h: New file.
20966 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
20969 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
20972 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20974 PR rtl-optimization/56117
20975 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
20976 call cselib_lookup_from_insn on the MEM before calling
20977 add_insn_mem_dependence.
20979 2013-01-28 Richard Biener <rguenther@suse.de>
20981 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
20982 to a stmt that didn't have one.
20983 (copy_phis_for_bb): Likewise for PHI arguments.
20984 (copy_debug_stmt): Likewise for debug stmts.
20986 2013-01-28 Richard Biener <rguenther@suse.de>
20988 PR tree-optimization/56034
20989 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
20990 (partition_builtin_p): Adjust.
20991 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
20992 it is the last partition.
20993 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
20994 up the vertex for the definition.
20995 (classify_partition): Classify whether a partition is a
20996 PKIND_REDUCTION, thus has uses outside of the loop.
20997 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
20998 Merge all PKIND_REDUCTION partitions into the last partition.
20999 (tree_loop_distribution): Seed partitions from reductions as well.
21001 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21003 PR tree-optimization/56125
21004 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21005 pow(x,c) into sqrt(x) * powi(x, n/2) or
21006 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21007 optimizing for size.
21008 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21009 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21012 PR tree-optimization/56094
21013 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21014 to UNKNOWN_LOCATION while gimplifying expr.
21016 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21019 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21020 operand 0 in movabs insn template for -masm=intel asm alternative.
21021 (*movabs<mode>_2): Ditto for operand 1.
21023 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21026 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21029 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21031 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21034 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21036 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21037 Undef to avoid warning.
21039 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21041 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21042 * configure: Regenerate.
21044 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21046 PR tree-optimization/56098
21047 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21048 for stmts with volatile ops.
21049 (cond_store_replacement): Don't optimize if assign has volatile ops.
21050 (cond_if_else_store_replacement_1): Don't optimize if either
21051 then_assign or else_assign have volatile ops.
21052 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
21055 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21057 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
21059 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21061 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
21062 missing ':' in asm example.
21064 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
21066 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
21067 entries into lane and laneq entries.
21068 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
21069 Remove AdvSIMD scalar modes.
21070 (aarch64_sq<r>dmulh_laneq<mode>): New.
21071 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
21073 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
21074 builtin implementations to relfect changes in RTL in aarch64-simd.md.
21075 * config/aarch64/iterators.md (VCOND): New.
21078 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21081 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
21082 Add NULL LIBNAME argument to existing definitions.
21083 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
21084 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
21085 * config/avr/avr.c (DEF_BUILTIN): Same.
21086 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
21087 (avr_expand_builtin): Expand to a vanilla call if a libgcc
21088 implementation is available (DECL_ASSEMBLER_NAME is set).
21089 (avr_fold_absfx): New static function.
21090 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
21091 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
21092 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
21093 AVR_BUILTIN_ABSLLK.
21094 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
21095 (abshk, absk, abslk, absllk): Provide as static inline functions.
21097 2013-01-25 Marek Polacek <polacek@redhat.com>
21099 PR tree-optimization/56035
21100 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
21102 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
21104 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
21105 (*movtf_internal_rex64): Add (!o,C) alternative
21106 (*movxf_internal_rex64): Ditto.
21107 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
21109 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
21111 * doc/invoke.texi: fix typo.
21112 * doc/objc.texi: fix typo.
21114 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
21116 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
21117 for the first two alternatives.
21119 2013-01-24 Diego Novillo <dnovillo@google.com>
21121 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
21122 (ggc-zone.o): Remove.
21123 * configure.ac: Remove option --with-gc.
21124 * configure: Re-generate.
21125 * doc/install.texi: Remove documentation for --with-gc.
21126 * gengtype.c (write_enum_defn): Remove. Update all users.
21127 (write_Types_process_field): Remove generation of gt_e_* argument.
21128 (output_type_enum): Remove. Update all users.
21129 (write_enum_defn): Remove. Update all users.
21130 (enum alloc_zone): Remove. Update all users.
21131 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
21132 * ggc-common.c (ggc_splay_alloc): Remove first argument.
21133 Update all callers.
21134 (struct ptr_data): Remove field TYPE. Update all users.
21135 (gt_pch_note_object): Remove argument TYPE. Update all users.
21136 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
21138 * ggc-none.c (ggc_alloc_typed_stat): Remove.
21139 (struct alloc_zone): Remove.
21140 (ggc_internal_alloc_zone_stat): Remove.
21141 (ggc_internal_cleared_alloc_zone_stat): Remove.
21142 * ggc-page.c (ggc_alloc_typed_stat): Remove.
21143 (ggc_pch_count_object): Remove last argument. Update all users.
21144 (ggc_pch_alloc_object): Remove last argument. Update all users.
21145 (struct alloc_zone): Remove.
21146 * ggc-zone.c: Remove.
21147 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
21148 (struct alloc_zone): Remove.
21149 (ggc_alloc_typed_stat): Remove.
21150 (ggc_alloc_typed): Remove.
21151 (ggc_splay_alloc): Remove first argument.
21152 (rtl_zone): Remove. Update all users.
21153 (tree_zone): Remove. Update all users.
21154 (tree_id_zone): Remove. Update all users.
21155 (ggc_internal_zone_alloc_stat): Remove. Update all users.
21156 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
21157 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
21158 * tree-ssanames.c: Remove references to zone allocator in comments.
21160 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
21162 * config/avr/avr.c (avr_out_fract): Make register numbers that
21163 might be outside of source operand signed.
21165 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
21167 * config/i386/constraints.md (Yf): New constraint.
21168 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
21169 of f constraint to conditionaly disable x87 register preferences.
21170 (*movdf_internal): Ditto.
21171 (*movsf_internal): Ditto.
21173 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
21175 PR inline-asm/55934
21176 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
21177 that have operands with impossible constraints.
21178 Add a FIXME for a speed-up opportunity.
21179 * lra-constraints.c (process_alt_operands): Verify that a class
21180 selected from constraints on asms is valid for the operand mode.
21181 (curr_insn_transform): Remove incorrect comment.
21183 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
21185 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
21186 TOC operand is a valid symbol ref in the constant pool.
21188 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21190 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
21192 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
21195 * config/avr/stdfix.h: New file.
21196 * t-avr (stdfix-gcc.h): New rule to build it.
21197 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
21199 2013-01-23 Kostya Serebryany <kcc@google.com>
21201 * config/darwin.h: remove dependency on
21202 CoreFoundation (asan on Mac OS).
21204 2013-01-23 Jakub Jelinek <jakub@redhat.com>
21207 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
21208 instead of cmpdi_operand for first comparison operand.
21209 Don't assert that comparison operands aren't both constants.
21211 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21213 * doc/install.texi (Downloading the Source): Update references to
21214 downloading separate components.
21216 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21218 * doc/extend.texi (__int128): Improve grammar.
21220 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
21223 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
21224 alternative to (o,r).
21225 (*movdi_internal_rex64): Remove (!o,n) alternative.
21226 (DImode immediate->memory splitter): Remove.
21227 (DImode immediate->memory peephole2): Remove.
21228 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
21229 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
21230 alternative to (!o,*r).
21231 (*movtf_internal_sse): New pattern.
21232 (*movxf_internal_rex64): New pattern.
21233 (*movxf_internal): Disable for TARGET_64BIT.
21234 (*movdf_internal_rex64): Remove (!o,F) alternative.
21236 2013-01-22 Jakub Jelinek <jakub@redhat.com>
21238 PR middle-end/56074
21239 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
21240 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
21241 * tree-vect-loop-manip.c (find_loop_location): Also ignore
21242 stmt locations where LOCATION_LOCUS of the stmt location is
21243 UNKNOWN_LOCATION or BUILTINS_LOCATION.
21246 * config/i386/i386.md (UNSPEC_STOS): New.
21247 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
21248 *strsetqi_1): Add UNSPEC_STOS.
21250 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
21253 * doc/invoke.texi: Remove left over -Wsynth example.
21255 2013-01-21 Jakub Jelinek <jakub@redhat.com>
21257 PR tree-optimization/56051
21258 * fold-const.c (fold_binary_loc): Don't fold
21259 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
21260 a narrowing conversion, or widening conversion from signed
21263 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
21265 PR rtl-optimization/56023
21266 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
21267 dependent on debug instruction.
21269 2013-01-21 Martin Jambor <mjambor@suse.cz>
21271 PR middle-end/56022
21272 * function.c (allocate_struct_function): Call
21273 invoke_set_current_function_hook earlier.
21275 2013-01-21 Jakub Jelinek <jakub@redhat.com>
21277 * reload1.c (init_reload): Only initialize reload_obstack
21278 during the first call.
21280 2013-01-21 Marek Polacek <polacek@redhat.com>
21282 * cfgloop.c (verify_loop_structure): Fix up grammar.
21284 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
21286 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
21287 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
21289 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21292 * config/arm/marvell-pj4.md: Update copyright year.
21293 Fix up use of alu to alu_reg and simple_alu_imm.
21295 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
21297 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
21299 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
21302 * lra-constraints.c (curr_insn_transform): Don't reuse original
21303 insn for secondary memory move when memory mode should be different.
21305 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21307 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
21308 atomic_storedi_1): New patterns.
21310 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21312 btver2 pipeline descriptions.
21313 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
21315 * config/i386/i386.md (btver2_decode): New type attributes.
21316 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
21318 * config/i386/btver2.md: New file describing btver2 pipelines.
21320 2013-01-19 Andrew Pinski <apinski@cavium.com>
21322 PR tree-optimization/52631
21323 * tree-ssa-sccvn (visit_use): Before looking up the original
21324 statement, try looking up the simplified expression.
21326 2013-01-19 Anthony Green <green@moxielogic.com>
21328 * config/moxie/moxie.c (moxie_expand_prologue): Set
21329 current_function_static_stack_size.
21331 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21333 PR tree-optimization/56029
21334 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
21335 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
21337 2013-01-18 Sharad Singhai <singhai@google.com>
21339 PR tree-optimization/55995
21340 * dumpfile.c (dump_loc): Print location only if available.
21341 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
21343 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
21346 * lra-constraints.c (curr_insn_transform): Reuse original insn for
21347 secondary memory move.
21348 (inherit_reload_reg): Use rclass instead of cl for
21349 check_secondary_memory_needed_p.
21351 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21353 PR middle-end/56015
21354 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
21355 the case where writing real complex part of target modifies op1.
21357 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
21359 * config/aarch64/aarch64-simd.md
21360 (aarch64_vcond_internal<mode>): Handle unordered cases.
21361 * config/aarch64/iterators.md (v_cmp_result): New.
21363 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
21364 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21366 * config/arm/marvell-pj4.md: New file.
21367 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
21368 * config/arm/arm.md (generic_sched): Add marvell_pj4.
21369 (generic_vfp): Likewise.
21370 * config/arm/arm-cores.def: Add marvell-pj4.
21371 * config/arm/arm-tune.md: Regenerate.
21372 * config/arm/arm-tables.opt: Regenerate.
21373 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
21374 * doc/invoke.texi: Document marvell-pj4.
21376 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
21378 * config/aarch64/arm_neon.h: Map scalar types to standard types.
21380 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21385 * var-tracking.c (negative_power_of_two_p): New.
21386 (global_get_addr_cache, local_get_addr_cache): New.
21387 (get_addr_from_global_cache, get_addr_from_local_cache): New.
21388 (vt_canonicalize_addr): Rewrite using the above. Adjust the
21390 (vt_stack_offset_p): Remove.
21391 (vt_canon_true_dep): Always canonicalize loc's address.
21392 (clobber_overlapping_mems): Make sure we have a MEM.
21393 (local_get_addr_clear_given_value): New.
21394 (val_reset): Clear local cached entries.
21395 (compute_bb_dataflow): Create and release the local cache.
21396 Disable duplicate MEMs clobbering.
21397 (emit_notes_in_bb): Clobber MEMs likewise.
21398 (vt_emit_notes): Create and release the local cache.
21399 (vt_initialize, vt_finalize): Create and release the global
21400 cache, respectively.
21401 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
21403 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21405 PR libmudflap/53359
21406 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
21407 not found in the symtab.
21409 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21412 PR rtl-optimization/55547
21413 PR rtl-optimization/53827
21416 * alias.c (offset_overlap_p): New, factored out of...
21417 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
21418 the conservative special case for symbolic constants. Don't
21419 adjust zero sizes on alignment.
21421 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
21423 PR rtl-optimization/52573
21424 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
21425 REG_UNUSED for the same register.
21427 2013-01-17 Richard Biener <rguenther@suse.de>
21428 Marek Polacek <polacek@redhat.com>
21430 PR rtl-optimization/55833
21431 * loop-unswitch.c (unswitch_loops): Move loop verification...
21432 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
21433 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
21434 Set it to true when we're removing a loop from hierarchy tree in
21435 an irreducible region.
21436 (fix_bb_placements): Adjust caller.
21437 (fix_loop_placements): Likewise.
21439 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
21441 * config/avr/builtins.def (DEF_BUILTIN): Factor out
21442 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
21443 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
21444 Remove ID. Adjust comments.
21445 * config/avr/avr-c.c (avr_builtin_name): Remove.
21446 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
21447 * config/avr/avr.c (avr_tolower): New static function.
21448 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
21449 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
21450 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
21453 2013-01-17 Jan Hubicka <jh@suse.cz>
21455 PR tree-optimization/55273
21456 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
21458 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
21461 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
21462 store through atomic_store<mode>_1.
21463 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
21465 2013-01-17 Martin Jambor <mjambor@suse.cz>
21467 PR tree-optimizations/55264
21468 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
21469 for virtual methods.
21470 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
21471 virtual methods before inlining is over.
21472 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
21474 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
21477 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
21479 PR rtl-optimization/56005
21480 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
21481 pending reads for prefetch.
21483 2013-01-16 Ian Bolton <ian.bolton@arm.com>
21485 * config/aarch64/aarch64.md
21486 (*cstoresi_neg_uxtw): New pattern.
21487 (*cmovsi_insn_uxtw): New pattern.
21488 (*<optab>si3_uxtw): New pattern.
21489 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
21490 (*<optab>si3_insn_uxtw): New pattern.
21491 (*bswapsi2_uxtw): New pattern.
21493 2013-01-16 Richard Biener <rguenther@suse.de>
21495 * tree-inline.c (tree_function_versioning): Remove set but
21496 never used variable.
21498 2013-01-16 Richard Biener <rguenther@suse.de>
21500 PR tree-optimization/55964
21501 * tree-flow.h (rename_variables_in_loop): Remove.
21502 (rename_variables_in_bb): Likewise.
21503 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
21504 (copy_loop_before): Adjust and delete update-ssa status.
21505 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
21506 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
21507 (rename_variables_in_loop): Remove.
21508 (slpeel_update_phis_for_duplicate_loop): Likewise.
21509 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
21510 use available cfg machinery instead of duplicating it.
21511 Update PHI nodes and perform poor-mans SSA update here.
21512 (slpeel_tree_peel_loop_to_edge): Adjust.
21514 2013-01-16 Richard Biener <rguenther@suse.de>
21516 PR tree-optimization/54767
21517 PR tree-optimization/53465
21518 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
21519 (vrp_visit_phi_node): For PHI arguments coming via backedges
21520 drop all symbolical range information.
21521 (execute_vrp): Compute backedges.
21523 2013-01-16 Richard Biener <rguenther@suse.de>
21525 * doc/install.texi: Update CLooG and ISL requirements to
21528 2013-01-16 Christian Bruel <christian.bruel@st.com>
21531 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
21532 (broken_move): Handle UNSPECV_SP_SWITCH_B.
21533 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
21535 2013-01-16 DJ Delorie <dj@redhat.com>
21537 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
21538 (UNSPECV_SP_SWITCH_E): New.
21539 (sp_switch_1): Change to an unspec.
21540 (sp_switch_2): Change to an unspec. Don't use post-inc when we
21543 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
21545 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
21546 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
21547 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
21548 (expand_mem_thread_fence): Ditto.
21549 (expand_mem_signal_fence): Ditto.
21550 (expand_atomic_load): Ditto.
21551 (expand_atomic_store): Ditto.
21553 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
21555 PR rtl-optimization/55547
21556 PR rtl-optimization/53827
21559 * alias.c (memrefs_conflict_p): Set sizes to negative after
21562 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21565 * function.c (thread_prologue_and_epilogue_insns): Always
21566 add crtl->drap_reg to set_up_by_prologue.set, even if
21567 stack_realign_drap is false.
21569 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
21571 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
21572 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
21573 *call): Fix indention.
21575 2013-01-15 Tom de Vries <tom@codesourcery.com>
21578 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
21581 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
21583 PR rtl-optimization/55153
21584 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
21586 2013-01-15 Martin Jambor <mjambor@suse.cz>
21588 PR tree-optimization/55920
21589 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
21590 accesses as grp_to_be_debug_replaced.
21592 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21594 PR tree-optimization/55920
21595 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
21596 there is non-useless type conversion needed from debug rhs to lhs,
21597 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
21599 2013-01-15 Joseph Myers <joseph@codesourcery.com>
21600 Mikael Pettersson <mikpe@it.uu.se>
21603 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
21605 (ASM_OUTPUT_CASE_LABEL): Remove.
21606 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
21607 * final.c (shorten_branches): Update alignment of labels before
21608 jump tables if CASE_VECTOR_SHORTEN_MODE.
21610 2013-01-15 Richard Biener <rguenther@suse.de>
21613 * system.h: Do not include gmp.h for building host tools.
21615 2013-01-15 Richard Biener <rguenther@suse.de>
21617 PR middle-end/55882
21618 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
21619 account for bitpos when computing alignment.
21621 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
21623 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
21624 (ix86_target_macros_internal): Likewise.
21626 * config/i386/i386.c (m_CORE2I7): Removed.
21627 (m_CORE_HASWELL): New macro.
21628 (m_CORE_ALL): Likewise.
21629 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
21630 (initial_ix86_arch_features): Likewise.
21631 (processor_target_table): Initializations for Core avx2.
21632 (cpu_names): New names "core-avx2".
21633 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
21634 PROCESSOR_CORE_HASWELL.
21635 (ix86_issue_rate): New case.
21636 (ia32_multipass_dfa_lookahead): Likewise.
21637 (ix86_sched_init_global): Likewise.
21639 * config/i386/i386.h (TARGET_HASWELL): New macro.
21640 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
21641 (processor_type): New PROCESSOR_HASWELL.
21643 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21645 PR tree-optimization/55955
21646 * tree-vect-loop.c (vectorizable_reduction): Give up early on
21647 *SHIFT_EXPR and *ROTATE_EXPR codes.
21649 PR tree-optimization/48766
21650 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
21651 -ftrapv disable -fwrapv.
21653 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21656 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
21657 etc. to 1 and not to __flash.
21658 Use LL suffix for __INT24_MAX__ with -mint8.
21659 Use ULL suffix for __UINT24_MAX__ with -mint8.
21661 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21663 * config/avr/avr-arch.h
21664 (struct base_arch_s): Use typedef avr_arch_t instead.
21665 (struct arch_info_s): Use typedef avr_arch_info_t instead.
21666 (struct mcu_type_s): Use typedef avr_mcu_t instead.
21667 * config/avr/avr.c: Same.
21668 * config/avr/avr-devices.c: Same.
21669 * config/avr/driver-avr.c: Same.
21670 * config/avr/gen-avr-mmcu-texi.c: Same.
21671 * config/avr/avr-mcus.def: Adjust comment.
21673 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
21675 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
21676 * config/aarch64/iterators.md (VALLDI): New.
21678 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
21679 Andi Kleen <ak@linux.intel.com>
21682 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
21683 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
21686 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21688 * config/avr/avr-stdint.h: Remove trailing blanks.
21689 * config/avr/avr-log.h: Same.
21690 * config/avr/avr-arch.h: Same.
21691 * config/avr/avr-devices.c: Same.
21692 * config/avr/avr-dimode.md: Same.
21693 * config/avr/predicates.md: Same.
21694 * config/avr/avr-c.c: Same. And fix typo.
21696 * config/avr/avr-protos.h: Same. And:
21697 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
21698 (init_cumulative_args): Rename to avr_init_cumulative_args.
21699 (expand_prologue): Rename to avr_expand_prologue.
21700 (expand_epilogue): Rename to avr_expand_epilogue.
21701 (adjust_insn_length): Rename to avr_adjust_insn_length.
21702 (notice_update_cc): Rename to avr_notice_update_cc.
21703 (final_prescan_insn): Rename to avr_final_prescan_insn.
21704 * config/avr/avr.c: Same.
21705 * config/avr/avr.h: Same.
21706 * config/avr/avr.md: Remove trailing blanks.
21707 (prologue): Use avr_expand_prologue.
21708 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
21710 2013-01-14 Richard Biener <rguenther@suse.de>
21712 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
21713 verify_location, collect_subblocks): New functions.
21714 (verify_gimple_in_cfg): Verify that locations only reference
21715 BLOCKs in the functions BLOCK tree.
21717 2013-01-14 Richard Biener <rguenther@suse.de>
21719 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
21721 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
21723 (insert_out_of_ssa_copy_on_edge): Likewise.
21724 (rewrite_close_phi_out_of_ssa): Likewise.
21725 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
21727 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
21728 propagated constants.
21729 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
21732 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21734 * config/avr/avr-modes.def: Add GPL copyright notice.
21736 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
21738 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
21739 MEMMODEL_MASK to determine memory model.
21740 (atomic_store<mode>): Ditto from operands[2].
21741 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
21743 2013-01-13 Jakub Jelinek <jakub@redhat.com>
21746 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
21747 (fold_gimple_assign): Don't call unshare_expr here.
21748 (fold_ctor_reference): Call unshare_expr.
21750 2013-01-13 Terry Guo <terry.guo@arm.com>
21752 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
21753 * doc/fragments.texi: Document MULTILIB_REUSE.
21754 * gcc.c (multilib_reuse): New internal spec.
21755 (set_multilib_dir): Also search multilib from multilib_reuse.
21756 * genmultilib (tmpmultilib3): Refactor code.
21757 (tmpmultilib4): Ditto.
21758 (multilib_reuse): New multilib argument.
21760 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
21762 * Makefile.in: Update copyright.
21764 2013-01-12 Tom de Vries <tom@codesourcery.com>
21766 PR middle-end/55890
21767 * calls.c (expand_call): Check if arg_nr is valid.
21769 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
21771 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
21772 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
21773 documentation. Add missing '__' in front of
21774 __builtin_ia32_packssdw256.
21776 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21779 * config/s390/s390.c (s390_preferred_reload_class): Do not return
21780 NO_REGS for larl operands.
21781 (s390_reload_larl_operand): Use s390_load_address instead of
21784 2013-01-11 Richard Biener <rguenther@suse.de>
21786 * tree-cfg.c (verify_node_sharing_1): Split out from ...
21787 (verify_node_sharing): ... here.
21788 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
21790 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
21792 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
21793 Substitute TREECHECKING.
21794 * configure: Regenerate.
21795 * Makefile.in (TREECHECKING): New.
21797 2013-01-11 Richard Guenther <rguenther@suse.de>
21799 PR tree-optimization/44061
21800 * tree-vrp.c (extract_range_basic): Compute zero as
21801 value-range for __builtin_constant_p of function parameters.
21803 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
21805 Update copyright years.
21807 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
21809 PR rtl-optimization/55672
21810 * lra-eliminations.c (mark_not_eliminable): Permit addition with
21811 const to be eliminable.
21813 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
21815 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
21816 * configure: Regenerate.
21818 2013-01-10 Richard Biener <rguenther@suse.de>
21820 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
21822 2013-01-10 Richard Biener <rguenther@suse.de>
21825 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
21826 locations for virtual PHI arguments.
21827 (rewrite_update_phi_arguments): Likewise.
21829 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
21831 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
21834 2013-01-10 Jakub Jelinek <jakub@redhat.com>
21836 PR tree-optimization/55921
21837 * tree-complex.c (expand_complex_asm): New function.
21838 (expand_complex_operations_1): Call it for GIMPLE_ASM.
21840 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21843 * config/s390/s390.c (s390_symref_operand_p)
21844 (s390_loadrelative_operand_p): Merge the two functions.
21845 (s390_check_qrst_address, print_operand_address): Add parameters
21846 to s390_loadrelative_operand_p invokation.
21847 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
21848 (s390_reload_larl_operand, s390_secondary_reload): Use
21849 s390_loadrelative_operand_p instead of s390_symref_operand_p.
21850 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
21852 2013-01-09 Mike Stump <mikestump@comcast.net>
21854 * dse.c (record_store): Remove unnecessary assert.
21856 2013-01-09 Jan Hubicka <jh@suse.cz>
21858 PR tree-optimization/55569
21859 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
21860 * cfgloop.h (scale_loop_profile): Likewise.
21862 2013-01-09 Jan Hubicka <jh@suse.cz>
21865 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
21867 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
21869 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
21871 PR middle-end/55114
21872 * expr.h (maybe_emit_group_store): Declare.
21873 * expr.c (maybe_emit_group_store): New function.
21874 * builtins.c (expand_builtin_int_roundingfn): Call it.
21875 (expand_builtin_int_roundingfn_2): Likewise.
21877 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
21879 PR rtl-optimization/55829
21880 * lra-constraints.c (match_reload): Add code for absent output.
21881 (curr_insn_transform): Add code for reloads of matched inputs
21884 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
21886 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
21887 attribute of movddup insn to DF.
21888 (*vec_interleave_lowv2df): Ditto.
21889 (vec_dupv2df): Ditto.
21891 2013-01-09 Jan Hubicka <jh@suse.cz>
21893 PR tree-optimiation/55875
21894 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
21895 EVERY_ITERATION parameter.
21896 (number_of_iterations_exit): Check if exit is executed every iteration.
21897 (idx_infer_loop_bounds): Similarly here.
21898 (n_of_executions_at_most): Simplify
21899 to only test for cases where statement is dominated by the
21900 particular bound; handle correctly the "postdominance" test.
21901 (scev_probably_wraps_p): Use max loop iterations info
21902 as a global bound first.
21904 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
21905 Nick Clifton <nickc@redhat.com>
21907 * config/v850/v850.md (cbranchsf4): New pattern.
21908 (cstoresf4): New pattern.
21909 (cbranchdf4): New pattern.
21910 (cstoredf4): New pattern.
21911 (movsicc): Disallow floating point comparisons.
21912 (cmpsf_le_insn): Fix order of operators.
21913 (cmpsf_lt_insn): Likewise.
21914 (cmpsf_eq_insn): Likewise.
21915 (cmpdf_le_insn): Likewise.
21916 (cmpdf_lt_insn): Likewise.
21917 (cmpdf_eq_insn): Likewise.
21918 (cmpsf_ge_insn): Use LE comparison.
21919 (cmpdf_ge_insn): Likewise.
21920 (cmpsf_gt_insn): Use LT comparison.
21921 (cmpdf_gt_insn): Likewise.
21922 (cmpsf_ne_insn): Delete pattern.
21923 (cmpdf_ne_insn): Delete pattern.
21924 * config/v850/v850.c (v850_gen_float_compare): Use
21925 gen_cmpdf_eq_insn for NE comparison.
21926 (v850_float_z_comparison_operator)
21927 (v850_float_nz_comparison_operator): Move from here ...
21928 * config/v850/predicates.md: ... to here. Move GT and GE
21929 comparisons into v850_float_z_comparison_operator.
21930 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
21932 (v850_float_nz_comparison_operator): Likewise.
21934 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21936 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
21937 with calls to gen_insvsi/gen_insvdi.
21939 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21941 * config/i386/i386.c (initial_ix86_tune_features): Set up
21942 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
21944 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
21945 Jakub Jelinek <jakub@redhat.com>
21947 PR tree-optimization/48189
21948 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
21949 If nitercst is 0, don't predict the exit edge.
21951 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21953 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
21954 in asm_fprintf with reg_names.
21955 (aarch64_print_operand_address): Likewise.
21956 (aarch64_return_addr): Likewise.
21957 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
21959 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21961 * config/pa/pa.h (VAL_U6_BITS_P): Define.
21962 (INT_U6_BITS): Likewise.
21963 * config/pa/predicates.md (uint6_operand): New predicate.
21964 (shift5_operand, shift6_operand): Likewise.
21965 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
21967 (lshrdi3): Use shift6_operand.
21968 (shrpsi4, shrpdi4): New insn patterns.
21969 (extzv): Delete expander.
21970 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
21971 predicates in unamed zero extract patterns. Tighten common constraint.
21972 (extv): Delete expander.
21973 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
21974 predicates in unamed sign extract patterns. Tighten common constraint.
21975 (insv): Delete expander.
21976 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
21977 predicates in unamed insert patterns. Tighten common constraint.
21978 Change uint32_operand predicate to uint6_operand predicate in unamed
21979 DImode pattern to insert constant values of type 1...1xxxx.
21981 2013-01-04 Jan Hubicka <jh@suse.cz>
21983 PR tree-optimization/55823
21984 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
21987 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21988 Uros Bizjak <ubizjak@gmail.com>
21990 PR rtl-optimization/55845
21991 * df-problems.c (can_move_insns_across): Stop scanning at
21992 volatile_insn_p source instruction or give up if
21993 across_from .. across_to range contains any volatile_insn_p
21996 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21998 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
21999 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22001 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22002 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22004 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22007 * asan.c (asan_clear_shadow): New function.
22008 (asan_emit_stack_protection): Use it.
22010 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22012 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22013 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22014 with tab instead of space.
22016 2013-01-08 Nick Clifton <nickc@redhat.com>
22018 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22019 register bank 0 at the start of an interrupt handler.
22020 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22023 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22025 * config/aarch64/aarch64-simd.md
22026 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22027 (aarch64_simd_bsl): Likewise.
22028 (aarch64_vcond_internal<mode>): Likewise.
22029 (vcond<mode><mode>): Likewise.
22030 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22031 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22033 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22035 * config/aarch64/aarch64-builtins.c
22036 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22038 2013-01-08 Martin Jambor <mjambor@suse.cz>
22041 * tree-sra.c (analyze_access_subtree): Return true also after
22042 potentially creating a debug-only replacement.
22044 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22046 PR middle-end/55890
22047 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22049 PR tree-optimization/54120
22050 * tree-vrp.c (range_fits_type_p): Don't allow
22051 src_precision < precision from signed vr to unsigned_p
22052 if vr->min or vr->max is negative.
22053 (simplify_float_conversion_using_ranges): Test can_float_p
22054 against CODE_FOR_nothing.
22056 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22057 Richard Biener <rguenther@suse.de>
22059 PR middle-end/55851
22060 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
22061 types instead of just INTEGER_TYPE types.
22063 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
22065 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
22068 2013-01-07 Steve Ellcey <sellcey@mips.com>
22071 * config/mips/mips.opt: Change mad to mmad to match documentation.
22073 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22076 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
22077 .progmemx.data now.
22079 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22082 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
22083 (avr_addrspace_t): Add .section_name field.
22084 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
22086 (avr_addrspace): Same. Initialize .section_name. Remove last
22087 NULL entry. Put __memx into .progmemx.data.
22088 (progmem_section_prefix): Remove.
22089 (avr_asm_init_sections): No need to initialize progmem_section.
22090 (avr_asm_named_section): Use avr_addrspace[].section_name to get
22091 section name prefix.
22092 (avr_asm_select_section): Ditto. And use get_unnamed_section to
22093 retrieve the progmem section.
22094 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
22095 boundary to run over avr_addrspace[].
22096 (avr_register_target_pragmas): Ditto.
22098 2013-01-06 Jakub Jelinek <jakub@redhat.com>
22100 * varasm.c (output_constant_def_contents): For asan_protect_global
22101 protected strings, adjust DECL_ALIGN if needed, before testing for
22103 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
22104 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
22106 (output_object_block): For asan protected decls, emit asan padding
22107 after their contents.
22108 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
22109 (asan_finish_file): Test it here instead.
22111 2013-01-07 Nick Clifton <nickc@redhat.com>
22112 Matthias Klose <doko@debian.org>
22113 Doug Kwan <dougkwan@google.com>
22114 H.J. Lu <hongjiu.lu@intel.com>
22117 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
22119 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
22121 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
22123 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
22125 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
22127 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22130 * doc/install.texi (Cross-Compiler-Specific Options): Document
22133 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
22135 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
22136 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
22137 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
22138 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
22139 vqmovun_high_s64): Fix source operand number and update copyright.
22141 2013-01-07 Richard Biener <rguenther@suse.de>
22143 PR middle-end/55890
22144 * gimple.h (gimple_call_builtin_p): New overload.
22145 * gimple.c (validate_call): New function.
22146 (gimple_call_builtin_p): Likewise.
22147 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22148 Use gimple_call_builtin_p.
22149 (find_func_clobbers): Likewise.
22150 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
22151 (strlen_optimize_stmt): Likewise.
22153 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22155 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
22156 (vld1q_dup_*): Likewise.
22157 (vld1_*): Likewise.
22158 (vld1q_*): Likewise.
22159 (vld1_lane_*): Likewise.
22160 (vld1q_lane_*): Likewise.
22162 2013-01-07 Richard Biener <rguenther@suse.de>
22164 * lto-streamer.h (LTO_minor_version): Bump to 2.
22166 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22168 * config/aarch64/aarch64-protos.h
22169 (aarch64_const_double_zero_rtx_p): Rename to...
22170 (aarch64_float_const_zero_rtx_p): ...this.
22171 (aarch64_float_const_representable_p): New.
22172 (aarch64_output_simd_mov_immediate): Likewise.
22173 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
22174 move immediate case.
22175 * config/aarch64/aarch64.c
22176 (aarch64_const_double_zero_rtx_p): Rename to...
22177 (aarch64_float_const_zero_rtx_p): ...this.
22178 (aarch64_print_operand): Allow printing of new constants.
22179 (aarch64_valid_floating_const): New.
22180 (aarch64_legitimate_constant_p): Check for valid floating-point
22182 (aarch64_simd_valid_immediate): Likewise.
22183 (aarch64_vect_float_const_representable_p): New.
22184 (aarch64_float_const_representable_p): Likewise.
22185 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
22186 (aarch64_output_simd_mov_immediate): New.
22187 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
22188 (*movdf_aarch64): Likewise.
22189 * config/aarch64/constraints.md (Ufc): New.
22190 (Y): call aarch64_float_const_zero_rtx.
22191 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
22193 2013-01-07 Richard Biener <rguenther@suse.de>
22195 PR tree-optimization/55888
22196 PR tree-optimization/55862
22197 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
22198 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
22199 not if it is contained therein.
22201 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22203 * config/avr/t-avr: Typo.
22205 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22208 * config/avr/t-avr: Don't automatically rebuild
22209 $(srcdir)/config/avr/t-multilib
22210 $(srcdir)/config/avr/avr-tables.opt
22211 $(srcdir)/doc/avr-mmcu.texi
22212 (avr-mcus): New phony target to build them on request.
22213 (s-avr-mlib, s-avr-mmcu-texi): Remove.
22214 * avr/avr-mcus.def: Adjust comments.
22216 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
22218 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
22220 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22222 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
22224 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22226 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
22228 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
22230 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
22231 to generate profiling.
22232 * config/rs6000/aix64.h (LIB_SPEC): Same.
22234 2013-01-04 Andrew Pinski <apinski@cavium.com>
22236 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
22238 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
22240 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
22242 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
22244 (ix86_expand_move): Ditto.
22245 (ix86_zero_extend_to_Pmode): Ditto.
22246 (ix86_expand_call): Ditto.
22247 (ix86_expand_special_args_builtin): Ditto.
22248 (ix86_expand_builtin): Ditto.
22250 2013-01-04 Richard Biener <rguenther@suse.de>
22252 PR tree-optimization/55862
22253 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
22254 translating them through PHI nodes.
22256 2013-01-04 Martin Jambor <mjambor@suse.cz>
22258 PR tree-optimization/55755
22259 * tree-sra.c (sra_modify_assign): Do not check that an access has no
22260 children when trying to avoid producing a VIEW_CONVERT_EXPR.
22262 2013-01-04 Marek Polacek <polacek@redhat.com>
22264 PR middle-end/55859
22265 * opts.c (default_options_optimization): Clarify error message.
22267 2013-01-04 Richard Biener <rguenther@suse.de>
22269 PR middle-end/55863
22270 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
22273 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22276 * config/pa/pa.md (movsi): Revert previous change.
22277 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
22280 2013-01-03 Richard Henderson <rth@redhat.com>
22282 * config/i386/i386.c (ix86_expand_move): Always assign to op1
22283 after eliminating TLS symbols.
22285 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
22288 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
22289 * graphite-poly.c (debug_gmp_value): Likewise.
22291 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
22294 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
22295 selected code model, define __code_mode_small__, __code_model_medium__,
22296 __code_model_large__, __code_model_32__ or __code_model_kernel__.
22297 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
22298 xchg temporary register with %k. Declare temporary register as
22300 [__x86_64__]: For medium and large code models, preserve %rbx register.
22302 2013-01-03 Richard Biener <rguenther@suse.de>
22304 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
22305 (dump_subscript): Adjust.
22306 (finalize_ddr_dependent): Do not dump redundant info.
22307 (analyze_siv_subscript): Adjust.
22308 (subscript_dependence_tester): Likewise.
22309 (compute_affine_dependence): Likewise.
22311 2013-01-03 Richard Biener <rguenther@suse.de>
22314 2013-01-03 Richard Biener <rguenther@suse.de>
22316 PR tree-optimization/55857
22317 * tree-vect-stmts.c (vectorizable_load): Do not setup
22318 re-alignment for invariant loads.
22320 2013-01-02 Richard Biener <rguenther@suse.de>
22322 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
22323 invariant load do not generate a vector load from the scalar location.
22325 2013-01-03 Richard Biener <rguenther@suse.de>
22327 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
22328 for not vectorizing.
22329 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
22330 not build INDIRECT_REFs, call get_name once only.
22331 (vect_create_data_ref_ptr): Likewise. Dump base object kind
22332 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
22334 2013-01-03 Richard Biener <rguenther@suse.de>
22336 PR tree-optimization/55857
22337 * tree-vect-stmts.c (vectorizable_load): Do not setup
22338 re-alignment for invariant loads.
22340 2013-01-03 Richard Biener <rguenther@suse.de>
22343 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
22344 prefer a built-in decl.
22346 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22348 * gcc.c (process_command): Update copyright notice dates.
22349 * gcov.c (print_version): Likewise.
22350 * gcov-dump.c (print_version): Likewise.
22352 PR rtl-optimization/55838
22353 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
22354 iv0.step, iv1.step and step.
22356 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22357 Marc Glisse <marc.glisse@inria.fr>
22359 PR tree-optimization/55832
22360 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
22361 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
22362 integer_{one,zero}_node.
22364 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22367 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
22368 * var-tracking.c (reverse_op): Don't add reverse ops to
22369 VALUEs that have already
22370 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
22372 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
22374 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
22376 2013-01-02 Teresa Johnson <tejohnson@google.com>
22378 * dumpfile.c (dump_loc): Print filename with location.
22379 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
22380 new location_t parameter to emit complete unroll message with
22381 new dump framework.
22382 (canonicalize_loop_induction_variables): Compute loops location
22383 and pass to try_unroll_loop_completely.
22384 * loop-unroll.c (report_unroll_peel): New function.
22385 (peel_loops_completely): Use new dump format with location
22386 for main dumpfile message, and invoke report_unroll_peel on success.
22387 (decide_unrolling_and_peeling): Ditto.
22388 (decide_peel_once_rolling): Remove old dumpfile message subsumed
22389 by report_unroll_peel.
22390 (decide_peel_completely): Ditto.
22391 (decide_unroll_constant_iterations): Ditto.
22392 (decide_unroll_runtime_iterations): Ditto.
22393 (decide_peel_simple): Ditto.
22394 (decide_unroll_stupid): Ditto.
22395 * cfgloop.c (get_loop_location): New function.
22396 * cfgloop.h (get_loop_location): Declare.
22398 2013-01-02 Sriraman Tallam <tmsriram@google.com>
22400 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
22403 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22405 PR middle-end/55198
22406 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
22407 BLKmode objects when EXPAND_MEMORY is specified.
22409 2013-01-02 Sriraman Tallam <tmsriram@google.com>
22411 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
22413 (fold_builtin_cpu): Do not share cpu model decls across statements.
22415 2013-01-02 Jason Merrill <jason@redhat.com>
22418 * tree.c (build_array_type_1): Revert earlier change.
22420 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
22422 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
22424 * config/aarch64/aarch64-tune.md: Re-generate.
22426 2013-01-02 Richard Biener <rguenther@suse.de>
22428 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
22429 invariant load do not generate a vector load from the scalar location.
22431 2013-01-02 Richard Biener <rguenther@suse.de>
22434 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
22435 * configure: Regenerate.
22437 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
22439 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
22440 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
22441 (expand_builtin_int_roundingfn_2): Keep the original target around
22442 for the fallback case.
22444 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
22446 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
22447 to be clear for sign changes.
22449 2013-01-01 Jan Hubicka <jh@suse.cz>
22451 * ipa-inline-analysis.c: Fix formatting.
22453 2013-01-01 Jakub Jelinek <jakub@redhat.com>
22455 PR tree-optimization/55831
22456 * tree-vect-loop.c (get_initial_def_for_induction): Use
22457 gsi_after_labels instead of gsi_start_bb.
22459 Copyright (C) 2013 Free Software Foundation, Inc.
22461 Copying and distribution of this file, with or without modification,
22462 are permitted in any medium without royalty provided the copyright
22463 notice and this notice are preserved.