1 2013-11-19 Jan Hubicka <jh@suse.cz>
3 * cgraph.c (cgraph_create_indirect_edge): Use get_polymorphic_call_info.
4 * cgraph.h (cgraph_indirect_call_info): Add outer_type, maybe_in_construction
5 and maybe_derived_type.
6 * ipa-utils.h (ipa_polymorphic_call_context): New structure.
7 (ipa_dummy_polymorphic_call_context): New global var.
8 (possible_polymorphic_call_targets): Add context paramter.
9 (dump_possible_polymorphic_call_targets): Likewise; update
11 (possible_polymorphic_call_target_p): Likewise.
12 (get_polymorphic_call_info): New function.
13 * ipa-devirt.c (ipa_dummy_polymorphic_call_context): New function.
14 (add_type_duplicate): Remove forgotten debug output.
15 (method_class_type): Add sanity check.
16 (maybe_record_node): Add FINALP parameter.
17 (record_binfo): Add OUTER_TYPE and OFFSET; walk the inner
18 by info by get_binfo_at_offset.
19 (possible_polymorphic_call_targets_1): Add OUTER_TYPE/OFFSET parameters;
20 pass them to record-binfo.
21 (polymorphic_call_target_d): Add context and FINAL.
22 (polymorphic_call_target_hasher::hash): Hash context.
23 (polymorphic_call_target_hasher::equal): Compare context.
24 (free_polymorphic_call_targets_hash):
25 (get_class_context): New function.
26 (contains_type_p): New function.
27 (get_polymorphic_call_info): New function.
28 (walk_bases): New function.
29 (possible_polymorphic_call_targets): Add context parameter; honnor it.
30 (dump_possible_polymorphic_call_targets): Dump context.
31 (possible_polymorphic_call_target_p): Add context.
32 (update_type_inheritance_graph): Update comment.s
33 (ipa_set_jf_known_type): Assert that compoentn type is known.
34 (ipa_note_param_call): Do not tamper with offsets.
35 (ipa_analyze_indirect_call_uses): When offset is being changed; clear
37 (update_indirect_edges_after_inlining): Likewise.
38 (ipa_write_indirect_edge_info): Stream new fields.
39 (ipa_read_indirect_edge_info): Stream in new fields.
41 2013-11-19 Jan Hubicka <jh@suse.cz>
43 * tree-pretty-print.c (dump_generic_node): Print class type of
46 2013-11-19 Joey Ye <joey.ye@arm.com>
48 * config/arm/arm.opt (-marm-pic-data-is-text-relative): New option.
49 * doc/invoke.texi (-marm-pic-data-is-text-relative): Documentation
51 * config/arm/arm.c (arm_option_override): By default disable
52 -marm-pic-data-is-text-relative.
53 (legitimize_pic_address): Use arm_pic_data_is_text_relative.
54 (arm_assemble_integer): Likewise.
55 * config/arm/arm.h (TARGET_DEFAULT_PIC_DATA_IS_TEXT_RELATIVE):
56 New macro to initialize -marm-pic-data-is-text-relative.
58 2013-11-19 Bin Cheng <bin.cheng@arm.com>
60 * tree-ssa-loop-ivopts.c (enum ainc_type): New.
61 (address_cost_data): New field.
62 (get_address_cost): Compute auto-increment rtx cost in ainc_costs.
63 Use ainc_costs for auto-increment rtx patterns.
66 2013-11-19 James Greenhalgh <james.greenhalgh@arm.com>
68 * config/aarch64/aarch64.md: Remove v8type from all insns.
70 2013-11-19 Richard Biener <rguenther@suse.de>
72 PR tree-optimization/57517
73 * tree-predcom.c (combinable_refs_p): Verify the combination
74 is always executed when the refs are.
76 2013-11-19 Jeff Law <law@redhat.com>
78 * tree-ssa-threadupdate.c: Include ssa-iterators.h
79 (copy_phi_arg_into_existing_phi): New function.
80 (any_remaining_duplicated_blocks): Likewise.
81 (ssa_fix_duplicate_block_edges): Handle multiple duplicated
82 blocks on a jump threading path.
84 * tree-ssa-threadupdate.c (thread_through_loop_header): Do not
85 thread through a joiner which has the latch edge.
87 2013-11-19 Jan Hubicka <jh@suse.cz>
89 * md.texi (setmem): Document new parameter.
90 * optabs.c (maybe_gen_insn): Support 9 operands.
91 * builtins.c (determine_block_size): Add probable_max_size;
93 (expand_builtin_memcpy. expand_builtin_memset_args): Pass around
95 * expr.c (emit_block_move_via_movmem, emit_block_move_hints,
96 emit_block_move, clear_storage_hints, set_storage_via_setmem):
98 * expr.h (emit_block_move_hints, clear_storage_hints,
99 set_storage_via_setmem): Update prototype.
100 * i386.md (setmem, movmem patterns): Add 9th operand.
101 * i386-protos.h (ix86_expand_set_or_movmem): Update prototype.
102 * i386.c (ix86_expand_set_or_movmem): Take probable_max_size_exp
103 argument; pass it to decide_alg.
105 2013-11-19 David Malcolm <dmalcolm@redhat.com>
107 * basic-block.h (n_basic_blocks_for_function): Rename macro to...
108 (n_basic_blocks_for_fn): ...this.
110 (n_basic_blocks): Eliminate macro as work towards making uses of
113 * cfgloop.c (init_loops_structure): Update for renaming of
114 "n_basic_blocks_for_function" to "n_basic_blocks_for_fn".
115 * graph.c (draw_cfg_nodes_no_loops): Likewise.
116 * ipa-utils.c (ipa_merge_profiles): Likewise.
117 * lto-streamer-in.c (make_new_block): Likewise.
118 * tree-cfg.c (init_empty_tree_cfg_for_function): Likewise.
119 (dump_function_to_file): Likewise.
121 * alias.c (init_alias_analysis): Replace usage of "n_basic_blocks"
122 macro with "n_basic_blocks_for_fn (cfun)".
123 * bb-reorder.c (partition_hot_cold_basic_blocks): Likewise.
124 (duplicate_computed_gotos): Likewise.
125 (reorder_basic_blocks): Likewise.
126 * bt-load.c (augment_live_range): Likewise.
127 * cfg.c (expunge_block): Likewise.
128 (compact_blocks): Likewise.
129 * cfganal.c (single_pred_before_succ_order): Likewise.
130 (compute_idf): Likewise.
131 (flow_dfs_compute_reverse_init): Likewise.
132 (pre_and_rev_post_order_compute): Likewise.
133 (pre_and_rev_post_order_compute_fn): Likewise.
134 (inverted_post_order_compute): Likewise.
135 (post_order_compute): Likewise.
136 (print_edge_list): Likewise.
137 (find_unreachable_blocks): Likewise.
138 (mark_dfs_back_edges): Likewise.
139 * cfgcleanup.c (try_optimize_cfg): Likewise.
140 (try_forward_edges): Likewise.
141 * cfghooks.c (dump_flow_info): Likewise.
142 * cfgloop.c (verify_loop_structure): Likewise.
143 (get_loop_body): Likewise.
144 (flow_loops_find): Likewise.
145 * cfgloopmanip.c (add_loop): Likewise.
146 (remove_path): Likewise.
147 (find_path): Likewise.
148 * cfgrtl.c (rtl_flow_call_edges_add): Likewise.
149 (rtl_verify_bb_layout): Likewise.
150 (entry_of_function): Likewise.
151 (rtl_create_basic_block): Likewise.
152 * coverage.c (coverage_compute_cfg_checksum): Likewise.
153 * cprop.c (one_cprop_pass): Likewise.
154 (is_too_expensive): Likewise.
155 * df-core.c (df_compute_cfg_image): Likewise.
156 (df_compact_blocks): Likewise.
157 (df_worklist_dataflow_doublequeue): Likewise.
158 * dominance.c (calculate_dominance_info): Likewise.
159 (calc_dfs_tree): Likewise.
160 (calc_dfs_tree_nonrec): Likewise.
161 (init_dom_info): Likewise.
162 * domwalk.c (cmp_bb_postorder): Likewise.
163 * function.c (thread_prologue_and_epilogue_insns): Likewise.
164 (generate_setjmp_warnings): Likewise.
165 * fwprop.c (build_single_def_use_links): Likewise.
166 * gcse.c (is_too_expensive): Likewise.
167 (one_code_hoisting_pass): Likewise.
168 (one_pre_gcse_pass): Likewise.
169 * graphite.c (graphite_initialize): Likewise.
170 * haifa-sched.c (haifa_sched_init): Likewise.
171 * ipa-inline-analysis.c (estimate_function_body_sizes): Likewise.
172 * ira.c (split_live_ranges_for_shrink_wrap): Likewise.
173 * ira-build.c (ira_build): Likewise.
174 * lcm.c (compute_nearerout): Likewise.
175 (compute_available): Likewise.
176 (compute_laterin): Likewise.
177 (compute_antinout_edge): Likewise.
178 * lra-lives.c (lra_create_live_ranges): Likewise.
179 * lra.c (has_nonexceptional_receiver): Likewise.
180 * mcf.c (create_fixup_graph): Likewise.
181 * profile.c (branch_prob): Likewise.
182 * reg-stack.c (convert_regs_2): Likewise.
183 * regrename.c (regrename_analyze): Likewise.
184 * reload1.c (has_nonexceptional_receiver): Likewise.
185 * reorg.c (dbr_schedule): Likewise.
186 * sched-deps.c (sched_deps_init): Likewise.
187 * sched-ebb.c (schedule_ebbs): Likewise.
188 * sched-rgn.c (extend_regions): Likewise.
189 (schedule_insns): Likewise.
190 (sched_rgn_init): Likewise.
191 (extend_rgns): Likewise.
192 (haifa_find_rgns): Likewise.
193 * sel-sched-ir.c (recompute_rev_top_order): Likewise.
194 (sel_recompute_toporder): Likewise.
195 * sel-sched.c (run_selective_scheduling): Likewise.
196 * store-motion.c (one_store_motion_pass): Likewise.
197 (remove_reachable_equiv_notes): Likewise.
198 * tracer.c (tracer): Likewise.
199 (tail_duplicate): Likewise.
200 * tree-cfg.c (gimple_flow_call_edges_add): Likewise.
201 (dump_cfg_stats): Likewise.
202 (gimple_dump_cfg): Likewise.
203 (create_bb): Likewise.
204 (build_gimple_cfg): Likewise.
205 * tree-cfgcleanup.c (merge_phi_nodes): Likewise.
206 * tree-inline.c (optimize_inline_calls): Likewise.
207 (fold_marked_statements): Likewise.
208 * tree-ssa-ifcombine.c (tree_ssa_ifcombine): Likewise.
209 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
210 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
211 * tree-ssa-loop-manip.c (compute_live_loop_exits): Likewise.
212 * tree-ssa-math-opts.c (execute_cse_reciprocals): Likewise.
213 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
214 * tree-ssa-pre.c (do_pre): Likewise.
215 (init_pre): Likewise.
216 (compute_avail): Likewise.
217 * tree-ssa-reassoc.c (init_reassoc): Likewise.
218 * tree-ssa-sccvn.c (init_scc_vn): Likewise.
219 * tree-ssa-tail-merge.c (alloc_cluster_vectors): Likewise.
220 (init_worklist): Likewise.
221 * tree-ssa-uncprop.c (associate_equivalences_with_edges): Likewise.
222 * var-tracking.c (variable_tracking_main_1): Likewise.
223 (vt_find_locations): Likewise.
224 (vt_stack_adjustments): Likewise.
225 * config/s390/s390.c (s390_optimize_nonescaping_tx): Likewise.
226 * config/spu/spu.c (spu_machine_dependent_reorg): Likewise.
228 2013-11-18 Jan Hubicka <jh@suse.cz>
230 * profile.c (compute_branch_probabilities): Do not sanity check run_max.
232 2013-11-18 Kenneth Zadeck <zadeck@naturalbridge.com>
234 * tree.c (int_fits_type_p): Change GET_MODE_BITSIZE to
236 * fold-const.c (fold_single_bit_test_into_sign_test)
237 (fold_binary_loc): Change GET_MODE_BITSIZE to
240 2013-11-18 Teresa Johnson <tejohnson@google.com>
242 * gcc/cfgrtl.c (cfg_layout_initialize): Assert if we
243 try to go into cfglayout after bb reordering.
244 * gcc/passes.def: Move compgotos before bb reordering
245 since it goes into cfglayout.
247 2013-11-18 Bernd Schmidt <bernds@codesourcery.com>
249 * cgraphunit.c (ipa_passes): Don't execute all_lto_gen_passes.
250 * lto-streamer-out.c (lto_output, produce_asm_for_decls): No longer
252 (pass_data_ipa_lto_gimple_out, pass_ipa_lto_gimple_out,
253 make_pass_ipa_lto_gimple_out, pass_data_ipa_lto_finish_out,
254 pass_ipa_lto_finish_out, make_pass_ipa_lto_finish_out): Remove.
255 * lto-streamer.h (lto_output, produce_asm_for_decls): Declare.
256 * pass-manager.h (GCC_PASS_LISTS, class pass_manager):
257 Remove all_lto_gen_passes.
258 * passes.c (pass_manager::dump_passes): Remove its use.
259 (pass_manager::register_pass): Likewise.
260 (ipa_read_summaries, ipa_read_optimization_summaries): Likewise.
261 (pass_manager::pass_manager): Don't initialize or use it.
262 (write_lto): New static function.
263 (ipa_write_summaries_1, ipa_write_optimization_summaries): Use it
264 instead of using all_lto_gen_passes.
265 * passes.def (all_lto_gen_passes, pass_ipa_lto_gimple_out,
266 pass_ipa_lto_finish_out): Delete.
267 * tree-pass.h (make_pass_ipa_lto_gimple_out,
268 make_pass_ipa_lto_finish_out): Don't declare.
270 2013-11-18 Jeff Law <law@redhat.com>
272 * tree-ssa-threadupdate.c (redirection_data): Record two
273 duplicated blocks instead of just one.
274 (local_info): Explain why we don't create a template for the
275 second duplicated block in a thread path.
276 (create_block_for_threading): Accept argument indicating array
277 index into redirection_data to store its result.
278 (lookup_redirection_data): Initialize both duplicate blocks.
279 (ssa_create_duplicates): If a jump threading path needs multiple
280 blocks duplicated, then duplicate them.
281 (ssa_fix_duplicate_block_edges): Corresponding changes.
282 (ssa_fixup_template_block, thread_single_edge): Likewise.
284 2013-11-18 Marek Polacek <polacek@redhat.com>
286 * doc/invoke.texi: Extend -fsanitize=undefined documentation.
288 2013-11-18 Andrew Pinski <apinski@cavium.com>
289 Steve Ellcey <sellcey@mips.com>
292 * config/mips/mips.md (*mov<GPR:mode>_on_<MOVECC:mode>): Remove
293 type restriction from equality_operator on conditonal move.
294 (*mov<SCALARF:mode>_on_<MOVECC:mode>): Ditto.
295 (*mov<GPR:mode>_on_<GPR2:mode>_ne): New.
297 2013-11-18 Jeff Law <law@redhat.com>
299 * tree-ssa-threadupdate.c: Fix file block comment.
300 Fix minor indention issue.
302 2013-11-18 Uros Bizjak <ubizjak@gmail.com>
304 * config/i386/i386.c (ix86_decompose_address): Use REG_P instead of
305 ix86_address_subreg_operand. Move subreg checks to
306 ix86_validate_address_register. Move address override check to
307 ix86_legitimate_address_p.
308 (ix86_validate_address_register): New function.
309 (ix86_legitimate_address_p): Call ix86_validate_address_register
310 to validate base and index registers. Add address override check
311 from ix86_decompose_address.
312 (ix86_decompose_address): Remove.
314 2013-11-18 Richard Biener <rguenther@suse.de>
316 PR tree-optimization/59125
317 PR tree-optimization/54570
318 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): When inlining
319 is not complete do not treat component-references with offset zero
320 but different fields as equal.
321 * tree-object-size.c: Include tree-phinodes.h and ssa-iterators.h.
322 (compute_object_sizes): Apply TLC. Propagate the constant
323 results into all uses and fold their stmts.
324 * passes.def (pass_all_optimizations): Move pass_object_sizes
325 after the first pass_forwprop and before pass_fre.
327 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
329 * tree.h (tree_to_uhwi): Return an unsigned HOST_WIDE_INT.
330 * tree.c (tree_to_uhwi): Return an unsigned HOST_WIDE_INT.
331 (tree_ctz): Remove cast to unsigned type.
332 * builtins.c (fold_builtin_memory_op): Likewise.
333 * dwarf2out.c (descr_info_loc): Likewise.
334 * godump.c (go_output_typedef): Likewise.
335 * omp-low.c (expand_omp_simd): Likewise.
336 * stor-layout.c (excess_unit_span): Likewise.
337 * tree-object-size.c (addr_object_size): Likewise.
338 * tree-sra.c (analyze_all_variable_accesses): Likewise.
339 * tree-ssa-forwprop.c (simplify_builtin_call): Likewise.
340 (simplify_rotate): Likewise.
341 * tree-ssa-strlen.c (adjust_last_stmt, handle_builtin_memcpy)
342 (handle_pointer_plus): Likewise.
343 * tree-switch-conversion.c (check_range): Likewise.
344 * tree-vect-patterns.c (vect_recog_rotate_pattern): Likewise.
345 * tsan.c (instrument_builtin_call): Likewise.
346 * cfgexpand.c (defer_stack_allocation): Add cast to HOST_WIDE_INT.
347 * trans-mem.c (tm_log_add): Likewise.
348 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Likewise.
349 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
350 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): Likewise.
351 * config/mips/mips.c (r10k_safe_mem_expr_p): Make offset unsigned.
353 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
355 * tree.h (host_integerp, tree_low_cst): Delete.
356 * tree.c (host_integerp, tree_low_cst): Delete.
358 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
360 * expr.h: Update comments to refer to tree_to_[su]hwi rather
362 * fold-const.c (fold_binary_loc): Likewise.
363 * expr.c (store_constructor): Use tree_to_uhwi rather than
365 * ipa-utils.h (possible_polymorphic_call_target_p): Likewise.
366 * stmt.c (emit_case_dispatch_table): Likewise.
367 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
369 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
371 * alias.c, asan.c, builtins.c, cfgexpand.c, cgraph.c,
372 config/aarch64/aarch64.c, config/alpha/predicates.md,
373 config/arm/arm.c, config/darwin.c, config/epiphany/epiphany.c,
374 config/i386/i386.c, config/iq2000/iq2000.c, config/m32c/m32c-pragma.c,
375 config/mep/mep-pragma.c, config/mips/mips.c,
376 config/picochip/picochip.c, config/rs6000/rs6000.c, cppbuiltin.c,
377 dbxout.c, dwarf2out.c, emit-rtl.c, except.c, expr.c, fold-const.c,
378 function.c, gimple-fold.c, godump.c, ipa-cp.c, ipa-prop.c, omp-low.c,
379 predict.c, sdbout.c, stor-layout.c, trans-mem.c, tree-object-size.c,
380 tree-sra.c, tree-ssa-ccp.c, tree-ssa-forwprop.c,
381 tree-ssa-loop-ivcanon.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-niter.c,
382 tree-ssa-loop-prefetch.c, tree-ssa-strlen.c, tree-stdarg.c,
383 tree-switch-conversion.c, tree-vect-generic.c, tree-vect-loop.c,
384 tree-vect-patterns.c, tree-vrp.c, tree.c, tsan.c, ubsan.c, varasm.c:
385 Replace tree_low_cst (..., 1) with tree_to_uhwi throughout.
387 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
389 * builtins.c, cilk-common.c, config/aarch64/aarch64.c,
390 config/alpha/alpha.c, config/arm/arm.c, config/c6x/predicates.md,
391 config/i386/i386.c, config/ia64/predicates.md, config/s390/s390.c,
392 coverage.c, dbxout.c, dwarf2out.c, except.c, explow.c, expr.c, expr.h,
393 fold-const.c, gimple-fold.c, godump.c, ipa-prop.c, omp-low.c,
394 predict.c, rtlanal.c, sdbout.c, stmt.c, stor-layout.c, targhooks.c,
395 tree-cfg.c, tree-data-ref.c, tree-inline.c, tree-ssa-forwprop.c,
396 tree-ssa-loop-prefetch.c, tree-ssa-phiopt.c, tree-ssa-sccvn.c,
397 tree-ssa-strlen.c, tree-stdarg.c, tree-vect-data-refs.c,
398 tree-vect-patterns.c, tree.c, tree.h, var-tracking.c, varasm.c:
399 Replace tree_low_cst (..., 0) with tree_to_shwi throughout.
401 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
403 * tree.h (tree_to_shwi, tree_to_uhwi): Declare, with inline expansions.
404 * tree.c (tree_to_shwi, tree_to_uhwi): New functions.
406 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
408 * expr.h: Update comments to refer to tree_fits_[su]hwi_p rather
411 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
413 * builtins.c, config/alpha/alpha.c, config/iq2000/iq2000.c,
414 config/mips/mips.c, dbxout.c, dwarf2out.c, expr.c, fold-const.c,
415 gimple-fold.c, godump.c, omp-low.c, predict.c, sdbout.c, stor-layout.c,
416 tree-dfa.c, tree-sra.c, tree-ssa-forwprop.c, tree-ssa-loop-prefetch.c,
417 tree-ssa-phiopt.c, tree-ssa-sccvn.c, tree-ssa-strlen.c,
418 tree-ssa-structalias.c, tree-vect-data-refs.c, tree-vect-patterns.c,
419 tree.c, varasm.c, alias.c, cfgexpand.c, config/aarch64/aarch64.c,
420 config/arm/arm.c, config/epiphany/epiphany.c, config/i386/i386.c,
421 config/m32c/m32c-pragma.c, config/mep/mep-pragma.c,
422 config/rs6000/rs6000.c, config/sparc/sparc.c, emit-rtl.c, function.c,
423 gimplify.c, ipa-prop.c, stmt.c, trans-mem.c, tree-cfg.c,
424 tree-object-size.c, tree-ssa-ccp.c, tree-ssa-loop-ivcanon.c,
425 tree-stdarg.c, tree-switch-conversion.c, tree-vect-generic.c,
426 tree-vrp.c, tsan.c, ubsan.c: Replace host_integerp (..., 1) with
427 tree_fits_uhwi_p throughout.
429 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
431 * builtins.c, config/alpha/alpha.c, config/c6x/predicates.md,
432 config/ia64/predicates.md, config/iq2000/iq2000.c, config/mips/mips.c,
433 config/s390/s390.c, dbxout.c, dwarf2out.c, except.c, explow.c, expr.c,
434 expr.h, fold-const.c, gimple-fold.c, gimple-ssa-strength-reduction.c,
435 gimple.c, godump.c, graphite-scop-detection.c, graphite-sese-to-poly.c,
436 omp-low.c, predict.c, rtlanal.c, sdbout.c, simplify-rtx.c,
437 stor-layout.c, tree-data-ref.c, tree-dfa.c, tree-pretty-print.c,
438 tree-sra.c, tree-ssa-alias.c, tree-ssa-forwprop.c,
439 tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c, tree-ssa-math-opts.c,
440 tree-ssa-phiopt.c, tree-ssa-reassoc.c, tree-ssa-sccvn.c,
441 tree-ssa-strlen.c, tree-ssa-structalias.c, tree-vect-data-refs.c,
442 tree-vect-patterns.c, tree-vectorizer.h, tree.c, var-tracking.c,
443 varasm.c: Replace host_integerp (..., 0) with tree_fits_shwi_p
446 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
448 * tree.h (tree_fits_shwi_p, tree_fits_uhwi_p): Declare.
449 * tree.c (tree_fits_shwi_p, tree_fits_uhwi_p): Define.
451 2013-11-18 Kirill Yukhin <kirill.yukhin@intel.com>
453 * gcc/config/ia64/ia64.c (ia64_split_tmode_move): Mark load with
454 `dead' flag if it kills address, not its post-increment.
456 2013-11-18 Ilya Enkovich <ilya.enkovich@intel.com>
458 * builtin-types.def (BT_FN_PTR_CONST_PTR_VAR): New.
459 * chkp-builtins.def (BUILT_IN_CHKP_BIND_BOUNDS): New.
460 * cfgexpand.c (expand_call_stmt): Expand BUILT_IN_CHKP_BIND_BOUNDS.
461 * gimple.c (gimple_call_get_nobnd_arg_index): Remove.
462 * gimple.h (gf_mask): Add GF_CALL_WITH_BOUNDS.
463 (gimple_call_with_bounds_p): New.
464 (gimple_call_set_with_bounds): New.
465 (gimple_call_num_nobnd_args): Remove.
466 (gimple_call_nobnd_arg): Remove.
467 * tree.h (CALL_WITH_BOUNDS_P): New.
468 * rtl.h (CALL_EXPR_WITH_BOUNDS_P): New.
470 2013-11-18 Trevor Saunders <tsaunders@mozilla.com>
472 * cgraph.h (symtab_node_asm_name): Rename to symtab_node::asm_name.
473 (symtab_node_name): Rename to symtab_node::name.
474 (cgraph_node_asm_name): Remove.
475 (varpool_node_asm_name): Remove.
476 * cgraph.c cgraphclones.c cgraphunit.c ipa-cp.c ipa-devirt.c
477 ipa-inline-analysis.c ipa-inline-transform.c ipa-inline.c
478 ipa-profile.c ipa-prop.c ipa-pure-const.c ipa-ref.c ipa-reference.c
479 ipa-utils.c ipa.c symtab.c tree-inline.c tree-sra.c
480 tree-ssa-structalias.c value-prof.c varpool.c Adjust.
482 2013-11-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
484 * config/arm/aarch-cost-tables.h (cortexa53_extra_costs): New table.
485 * config/arm/arm.c (arm_cortex_a53_tune): New.
486 * config/arm/arm-cores.def (cortex-a53): Use cortex_a53 tuning struct.
488 2013-11-12 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
490 * config.gcc (i[34567]86-*-linux* | ...): Add bdver4.
491 (case ${target}): Add bdver4.
492 * config/i386/bdver3.md: Add bdver4.
493 * config/i386/driver-i386.c: (host_detect_local_cpu): Let
494 -march=native recognize bdver4 processors.
495 * config/i386/i386-c.c (ix86_target_macros_internal): Add
497 * config/i386/i386.c (struct processor_costs bdver4_cost): New.
498 (m_BDVER4): New definition.
499 (m_AMD_MULTIPLE): Includes m_BDVER4.
500 (processor_target_table): Add bdver4 entry.
501 (static const char *const cpu_names): Add bdver4 entry.
502 (software_prefetching_beneficial_p): Add bdver3.
503 (ix86_option_override_internal): Add bdver4 instruction sets.
504 (ix86_issue_rate): Add bdver4.
505 (ix86_adjust_cost): Add bdver4.
506 (ia32_multipass_dfa_lookahead): Add bdver4.
507 (enum processor_model): Add M_AMDFAM15H_BDVER4.
508 (struct _arch_names_table): Add M_AMDFAM15H_BDVER4.
509 (has_dispatch): Add bdver4.
510 * config/i386/i386.h (TARGET_BDVER4): New definition.
511 (enum target_cpu_default): Add TARGET_CPU_DEFAULT_bdver4.
512 (enum processor_type): Add PROCESSOR_BDVER4.
513 * config/i386/i386.md (define_attr "cpu"): Add bdver4.
514 * config/i386/i386.opt (flag_dispatch_scheduler): Add bdver4.
515 * gcc/doc/extend.texi: Add details about bdver4.
516 * gcc/doc/invoke.texi: Add details about bdver4. Add
517 fma4 and fsgsbase for bdver3. Add fma4 for bdver2.
519 2013-11-17 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
521 * config/rs6000/rs6000.c (rs6000_emit_move): Use low word of
522 sdmode_stack_slot also in little-endian mode.
524 2013-11-17 Jan Hubicka <jh@suse.cz>
526 * doc/md.texi (setmem, movstr): Update documentation.
527 * builtins.c (determine_block_size): New function.
528 (expand_builtin_memcpy): Use it and pass it to emit_block_move_hints.
529 (expand_builtin_memset_args): Use it and pass it to
530 set_storage_via_setmem.
531 * expr.c (emit_block_move_via_movmem): Add min_size/max_size
532 parameters; update call to expander.
533 (emit_block_move_hints): Add min_size/max_size parameters.
534 (clear_storage_hints): Likewise.
535 (set_storage_via_setmem): Likewise.
536 (clear_storage): Update.
537 * expr.h (emit_block_move_hints, clear_storage_hints,
538 set_storage_via_setmem): Update prototypes.
539 * i386.c (ix86_expand_set_or_movmem): Add bounds; export.
540 (ix86_expand_movmem, ix86_expand_setmem): Remove.
541 (ix86_expand_movmem, ix86_expand_setmem): Remove.
542 * i386.md (movmem, setmem): Pass parameters.
544 2013-11-17 Uros Bizjak <ubizjak@gmail.com>
547 * config/i386/i386.c (ix86_address_subreg_operand): Do not
548 reject non-integer subregs.
549 (ix86_decompose_address): Do not reject invalid CONST_INT RTXes.
550 Move check for invalid x32 constant addresses ...
551 (ix86_legitimate_address_p): ... here.
553 2011-11-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
555 * config/rs6000/rs6000.c (rs6000_frame_related): Add split_reg
556 parameter and use it in REG_FRAME_RELATED_EXPR note.
557 (emit_frame_save): Call rs6000_frame_related with extra NULL_RTX
559 (rs6000_emit_prologue): Likewise, but for little endian VSX
560 stores, pass the source register of the store instead.
562 2013-11-17 Andrew MacLeod <amacleod@redhat.com>
564 * gimple.h: Reorder prototypes to match .c declaration order, and
565 remove protyotypes for functions not in gimple.c.
566 (LABEL): Move to tree-into-ssa.c.
567 * gimple.c: Remove unused prototypes.
568 (get_base_address): Move to tree.c.
569 * tree.c (get_base_address): Relocate from gimple.c.
570 * builtins.h (validate_gimple_arglist): Add prototype.
571 * trans-mem.h (compute_transaction_bits, is_tm_ending): Add prototype.
572 * cfgexpand.h: New File.
573 (gimple_assign_rhs_to_tree, estimated_stack_frame_size): Add protoype.
574 * tree.h (build_addr): Move to tree-nested.h.
575 * tree-nested.h: New File.
576 (build_addr, lower_nested_functions, insert_field_into_struct): Add
578 * tree-inline.h (estimated_stack_frame_size): Remove prototype.
579 * ipa-inline-analysis.c: Include cfgexpand.h.
580 * cgraphunit.c: Include tree-nested.h.
581 * omp-low.c: Likewise.
582 * tree-parloops.c: Likewise.
583 * gimple-low.h: Likewise.
584 * tree-profile.h: Likewise.
585 * expr.c: Include cfgexpand.h.
586 * tree-affine.c: Likewise.
587 * tree-ssa.c: Likewise.
588 * tree-ssa-loop-im.c: Include trans-mem.h.
589 * tree-ssa-tail-merge.c: Likewise.
590 * value-prof.c: Include builtins.h and tree-nested.h.
591 * tree-into-ssa.c (LABEL): Define here.
593 2013-11-16 Joern Rennecke <joern.rennecke@embecosm.com>
595 * config/arc/arc.c (arc_predicate_delay_insns): New function.
596 (pass_data_arc_predicate_delay_insns): New pass_data instance.
597 (pass_arc_predicate_delay_insns): New subclass of rtl_opt_class.
598 (make_pass_arc_predicate_delay_insns): New function.
599 (arc_init): Register pass_arc_predicate_delay_insns if
600 flag_delayed_branch is active.
602 2013-11-16 Joern Rennecke <joern.rennecke@embecosm.com>
604 * config/arc/constraints.md (Rcq): Simplify register number test.
606 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
608 * gimple.h (enum gf_mask): Change the ordering of GF_OMP_* bits.
610 2013-11-15 Kaz Kojima <kkojima@gcc.gnu.org>
612 * config/sh/sh.c (barrier_align): Return 0 when barrier_or_label
615 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
617 * Makefile.in (C_COMMON_OBJS): Depend on c-cilkplus.o.
618 * gimple-pretty-print.c (dump_omp_for): Add case for
619 GF_OMP_FOR_KIND_CILKSIMD.
620 * gimple.h (enum gf_mask): Restructure entries to add
621 GF_OMP_FOR_KIND_CILKSIMD.
622 * gimplify.c (is_gimple_stmt): Add case for CILK_SIMD.
623 (gimplify_omp_for): Handle CILK_SIMD.
624 (gimplify_expr): Add ccase for CILK_SIMD.
625 * omp-low.c (extract_omp_for_data): Handle CILK_SIMD.
626 (build_outer_var_ref): Same.
627 (check_omp_nesting_restrictions): Same.
628 (lower_rec_input_clauses): Same.
629 (lower_lastprivate_clauses): Same.
630 (expand_omp_for): Same.
631 (execute_expand_omp): Check flag_enable_cilkplus.
632 (execute_lower_omp): Same.
633 (diagnose_sb_0): Handle CILK_SIMD.
634 (diagnose_omp_structured_block_errors): Check flag_enable_cilkplus.
635 (setjmp_or_longjmp_p): New.
636 (scan_omp_1_stmt): Error on setjmp/longjmp in a simd construct.
637 * tree-pretty-print.c (dump_generic_node): Add case for CILK_SIMD.
638 * tree.def: Add tree code for CILK_SIMD.
640 2013-11-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
642 * config/rs6000/altivec.md (UNSPEC_VPERM_X, UNSPEC_VPERM_UNS_X):
644 (altivec_vperm_<mode>): Revert earlier little endian change.
645 (*altivec_vperm_<mode>_internal): Remove.
646 (altivec_vperm_<mode>_uns): Revert earlier little endian change.
647 (*altivec_vperm_<mode>_uns_internal): Remove.
648 * config/rs6000/vector.md (vec_realign_load_<mode>): Revise commentary.
650 2013-11-15 Jeff Law <law@redhat.com>
652 * basic-block.h (has_abnormal_or_eh_outgoing_edge): Renamed from
653 has_abnormal_or_outgoing_edge. Check for EH edges as well.
654 * gimple-ssa-isolate-paths.c
655 (find_implicit_erroneous_behaviour): Corresponding changes.
656 Do not check stmt_ends_bb_p or GIMPLE_RETURN anymore.
657 (find_explicit_erroneous_behaviour): Likewise.
659 2013-11-15 Jeff Law <law@redhat.com>
661 * ifcvt.c (find_cond_trap): Properly handle case where
664 2013-11-15 Andreas Schwab <schwab@linux-m68k.org>
666 * configure: Regenerate.
668 2013-11-15 James Greenhalgh <james.greenhalgh@arm.com>
670 * config/aarch64/aarch64-simd.md: Remove simd_type from all patterns.
671 * config/aarch64/aarch64.md: Likewise, correct "type" attribute
672 where it is incorrect or missing.
674 2013-11-15 Richard Sandiford <rdsandiford@googlemail.com>
676 * dwarf2out.c (gen_enumeration_type_die): Remove unnecessary
678 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
679 Use TREE_INT_CST_LOW rather than tree_low_cst when reading the
681 * fold-const.c (fold_binary_loc): Replace a host_integerp/tree_low_cst
682 pair with a TREE_CODE test and TREE_INT_CST_LOW.
683 * tree-vect-generic.c (expand_vector_divmod): Likewise.
685 2013-11-15 Richard Biener <rguenther@suse.de>
687 PR tree-optimization/50262
688 * tree-ssa-alias.h (struct pt_solution): Split
689 vars_contains_global into vars_contains_nonlocal,
690 vars_contains_escaped and vars_contains_escaped_heap.
691 * tree-ssa-structalias.c (label_visit): Expand comment.
692 (handle_lhs_call): Adjust comment.
693 (set_uids_in_ptset): Set the new flags appropriately.
694 (pt_solution_set): Adjust.
695 (pt_solution_set_var): Likewise.
696 (pt_solution_ior_into): Likewise.
697 (pt_solution_includes_global): Likewise.
698 (pt_solutions_intersect_1): Optimize escaped handling.
699 (compute_points_to_sets): Remove heap variable globalization.
700 (ipa_escaped_pt): Adjust initializer.
701 (pass_data_ipa_pta): Do not run TODO_update_ssa.
702 * gimple-pretty-print.c (pp_points_to_solution): Print split flags.
703 * tree-ssa-alias.c (dump_points_to_solution): Likewise.
705 2013-11-15 Richard Biener <rguenther@suse.de>
707 * tree-loop-distribution.c (tree_loop_distribution): Make sure
708 to distribute all stores.
710 2013-11-15 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
712 * doc/invoke.texi (-mabi=elfv1, -mabi=elfv2): Document.
714 2013-11-15 Joseph Myers <joseph@codesourcery.com>
716 * acinclude.m4 (GCC_GLIBC_VERSION_GTE_IFELSE): New configure macro.
717 * configure.ac: Determine target_header_dir earlier.
718 (--with-glibc-version): New configure option.
719 Use GCC_GLIBC_VERSION_GTE_IFELSE in enable_gnu_unique_object,
720 gcc_cv_libc_provides_ssp and gcc_cv_target_ldbl128 tests.
721 * configure: Regenerate.
722 * doc/install.texi (--enable-gnu-unique-object): Don't refer to
723 native toolchains for default.
724 (--with-glibc-version): Document.
726 2013-11-15 Eric Botcazou <ebotcazou@adacore.com>
728 * fold-const.c (fold_binary_loc) <comparisons>: Reuse local variable.
730 2013-11-15 Uros Bizjak <ubizjak@gmail.com>
732 * lto-streamer-in.c (input function): Call cgraph_create_node if
733 cgraph_get_node failed.
735 2013-11-14 Olivier Hainque <hainque@adacore.com>
737 * cfgexpand.c (defer_stack_allocation): When optimization is enabled,
738 defer allocation of DECL_IGNORED_P variables at toplevel unless really
739 small. Factorize size threshold computation from the existing one.
740 (expand_used_vars): Refine comment.
742 2013-11-14 Cong Hou <congh@google.com>
744 * tree-vectorizer.h (struct dr_with_seg_len): Remove the base
745 address field as it can be obtained from dr. Rename the struct.
746 * tree-vect-data-refs.c (comp_dr_with_seg_len_pair): Consider
747 steps of data references during sort.
748 (vect_prune_runtime_alias_test_list): Adjust with the change to
749 struct dr_with_seg_len.
750 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
751 Adjust with the change to struct dr_with_seg_len.
753 2013-11-14 Jeff Law <law@redhat.com>
756 * basic-block.h (has_abnormal_outgoing_edge_p): Moved here from...
757 * tree-inline.c (has_abnormal_outgoing_edge_p): Remove.
758 * gimple-ssa-isolate-paths.c: Include tree-cfg.h.
759 (find_implicit_erroneous_behaviour): If a block has abnormal outgoing
760 edges, then ignore it. If the statement exhibiting erroneous
761 behaviour ends basic blocks, with the exception of GIMPLE_RETURNs,
762 then we can not optimize.
763 (find_explicit_erroneous_behaviour): Likewise.
765 2013-11-14 Andrew MacLeod <amacleod@redhat.com>
767 * gimplify-me.h: New file. Add prototypes.
768 * gimplify.h: Don't include gimple.h.
769 (struct gimplify_hasher, struct gimplify_ctx, is_gimple_sizepos):
770 Relocate from gimple.h.
771 * gimple.h (struct gimplify_hasher, struct gimplify_ctx,
772 is_gimple_sizepos): Move to gimplify.h.
773 (gimplify_hasher::hash, gimplify_hasher::equal): Move to gimplify.c.
774 (enum gsi_iterator_update): Move to gimple-iterator.h.
775 * gimple-iterator.h (enum gsi_iterator_update): Relocate from gimple.h.
776 * gimplify-me.c: New File.
777 (force_gimple_operand_1, force_gimple_operand,
778 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
779 gimple_regimplify_operands): Relocate from gimplify.c.
780 * gimplify.c (force_gimple_operand_1, force_gimple_operand,
781 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
782 gimple_regimplify_operands): Move to gimplify-me.c.
783 (gimplify_hasher::hash, gimplify_hasher::equal): Relocate
785 * Makefile.in (OBJS): Add gimplify-me.o
786 * asan.c: Include only gimplify.h, gimplify-me.h, and/or gimple.h as
788 * cfgloopmanip.c: Likewise.
789 * cgraphunit.c: Likewise.
790 * cilk-common.c: Likewise.
791 * fold-const.c: Likewise.
792 * function.c: Likewise.
793 * gimple-expr.c: Likewise.
794 * gimple-fold.c: Likewise.
795 * gimple-ssa-strength-reduction.c: Likewise.
796 * gimple.c: Likewise.
797 * graphite-clast-to-gimple.c: Likewise.
798 * graphite-sese-to-poly.c: Likewise.
799 * ipa-prop.c: Likewise.
800 * ipa-split.c: Likewise.
802 * langhooks.c: Likewise.
803 * omp-low.c: Likewise.
805 * stor-layout.c: Likewise.
806 * targhooks.c: Likewise.
807 * trans-mem.c: Likewise.
808 * tree-affine.c: Likewise.
809 * tree-cfg.c: Likewise.
810 * tree-cfgcleanup.c: Likewise.
811 * tree-complex.c: Likewise.
812 * tree-if-conv.c: Likewise.
813 * tree-inline.c: Likewise.
814 * tree-loop-distribution.c: Likewise.
815 * tree-nested.c: Likewise.
816 * tree-parloops.c: Likewise.
817 * tree-predcom.c: Likewise.
818 * tree-profile.c: Likewise.
819 * tree-scalar-evolution.c: Likewise.
820 * tree-sra.c: Likewise.
821 * tree-ssa-address.c: Likewise.
822 * tree-ssa-ccp.c: Likewise.
823 * tree-ssa-dce.c: Likewise.
824 * tree-ssa-forwprop.c: Likewise.
825 * tree-ssa-ifcombine.c: Likewise.
826 * tree-ssa-loop-im.c: Likewise.
827 * tree-ssa-loop-ivopts.c: Likewise.
828 * tree-ssa-loop-manip.c: Likewise.
829 * tree-ssa-loop-niter.c: Likewise.
830 * tree-ssa-loop-prefetch.c: Likewise.
831 * tree-ssa-loop-unswitch.c: Likewise.
832 * tree-ssa-math-opts.c: Likewise.
833 * tree-ssa-phiopt.c: Likewise.
834 * tree-ssa-phiprop.c: Likewise.
835 * tree-ssa-pre.c: Likewise.
836 * tree-ssa-propagate.c: Likewise.
837 * tree-ssa-reassoc.c: Likewise.
838 * tree-ssa-sccvn.c: Likewise.
839 * tree-ssa-strlen.c: Likewise.
840 * tree-ssa.c: Likewise.
841 * tree-switch-conversion.c: Likewise.
842 * tree-tailcall.c: Likewise.
843 * tree-vect-data-refs.c: Likewise.
844 * tree-vect-generic.c: Likewise.
845 * tree-vect-loop-manip.c: Likewise.
846 * tree-vect-loop.c: Likewise.
847 * tree-vect-patterns.c: Likewise.
848 * tree-vect-stmts.c: Likewise.
851 * value-prof.c: Likewise.
852 * config/aarch64/aarch64.c: Likewise.
853 * config/alpha/alpha.c: Likewise.
854 * config/darwin.c: Likewise.
855 * config/i386/i386.c: Likewise.
856 * config/ia64/ia64.c: Likewise.
857 * config/mep/mep.c: Likewise.
858 * config/mips/mips.c: Likewise.
859 * config/rs6000/rs6000.c: Likewise.
860 * config/s390/s390.c: Likewise.
861 * config/sh/sh.c: Likewise.
862 * config/sparc/sparc.c: Likewise.
863 * config/spu/spu.c: Likewise.
864 * config/stormy16/stormy16.c: Likewise.
865 * config/tilegx/tilegx.c: Likewise.
866 * config/tilepro/tilepro.c: Likewise.
867 * config/xtensa/xtensa.c: Likewise.
869 2013-11-14 Joern Rennecke <joern.rennecke@embecosm.com>
871 * config/arc/arc.md (doloop_begin_i): Remove extra alignment;
874 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
876 * config/rs6000/sysv4le.h (LINUX64_DEFAULT_ABI_ELFv2): Define.
878 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
879 Alan Modra <amodra@gmail.com>
881 * config/rs6000/rs6000.h (RS6000_SAVE_AREA): Handle ABI_ELFv2.
882 (RS6000_SAVE_TOC): Remove.
883 (RS6000_TOC_SAVE_SLOT): New macro.
884 * config/rs6000/rs6000.c (rs6000_parm_offset): New function.
885 (rs6000_parm_start): Use it.
886 (rs6000_function_arg_advance_1): Likewise.
887 (rs6000_emit_prologue): Use RS6000_TOC_SAVE_SLOT.
888 (rs6000_emit_epilogue): Likewise.
889 (rs6000_call_aix): Likewise.
890 (rs6000_output_function_prologue): Do not save/restore r11
891 around calling _mcount for ABI_ELFv2.
893 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
894 Alan Modra <amodra@gmail.com>
896 * config/rs6000/rs6000-protos.h (rs6000_reg_parm_stack_space):
898 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Remove.
899 (REG_PARM_STACK_SPACE): Call rs6000_reg_parm_stack_space.
900 * config/rs6000/rs6000.c (rs6000_parm_needs_stack): New function.
901 (rs6000_function_parms_need_stack): Likewise.
902 (rs6000_reg_parm_stack_space): Likewise.
903 (rs6000_function_arg): Do not replace BLKmode by Pmode when
904 returning a register argument.
906 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
907 Michael Gschwind <mkg@us.ibm.com>
909 * config/rs6000/rs6000.h (FP_ARG_MAX_RETURN): New macro.
910 (ALTIVEC_ARG_MAX_RETURN): Likewise.
911 (FUNCTION_VALUE_REGNO_P): Use them.
912 * config/rs6000/rs6000.c (TARGET_RETURN_IN_MSB): Define.
913 (rs6000_return_in_msb): New function.
914 (rs6000_return_in_memory): Handle ELFv2 homogeneous aggregates.
915 Handle aggregates of up to 16 bytes for ELFv2.
916 (rs6000_function_value): Handle ELFv2 homogeneous aggregates.
918 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
919 Michael Gschwind <mkg@us.ibm.com>
921 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
922 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
923 (rs6000_discover_homogeneous_aggregate): Likewise.
924 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
925 (rs6000_function_arg_advance_1): Likewise.
926 (rs6000_function_arg): Likewise.
927 (rs6000_arg_partial_bytes): Likewise.
928 (rs6000_psave_function_arg): Handle BLKmode arguments.
930 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
931 Michael Gschwind <mkg@us.ibm.com>
933 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
934 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
935 (rs6000_discover_homogeneous_aggregate): Likewise.
936 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
937 (rs6000_function_arg_advance_1): Likewise.
938 (rs6000_function_arg): Likewise.
939 (rs6000_arg_partial_bytes): Likewise.
940 (rs6000_psave_function_arg): Handle BLKmode arguments.
942 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
944 * config/rs6000/rs6000.c (machine_function): New member
946 (rs6000_emit_prologue): Set r2_setup_needed if necessary.
947 (rs6000_output_mi_thunk): Set r2_setup_needed.
948 (rs6000_output_function_prologue): Output global entry point
949 prologue and local entry point marker if needed for ABI_ELFv2.
950 Output -mprofile-kernel code here.
951 (output_function_profiler): Do not output -mprofile-kernel
952 code here; moved to rs6000_output_function_prologue.
953 (rs6000_file_start): Output ".abiversion 2" for ABI_ELFv2.
955 (rs6000_emit_move): Do not handle dot symbols for ABI_ELFv2.
956 (rs6000_output_function_entry): Likewise.
957 (rs6000_assemble_integer): Likewise.
958 (rs6000_elf_encode_section_info): Likewise.
959 (rs6000_elf_declare_function_name): Do not create dot symbols
960 or .opd section for ABI_ELFv2.
962 (rs6000_trampoline_size): Update for ABI_ELFv2 trampolines.
963 (rs6000_trampoline_init): Likewise.
964 (rs6000_elf_file_end): Call file_end_indicate_exec_stack for ABI_ELFv2.
966 (rs6000_call_aix): Handle ELFv2 indirect calls. Do not check
967 for function descriptors in ABI_ELFv2.
969 * config/rs6000/rs6000.md ("*call_indirect_aix<mode>"): Support
970 on ABI_AIX only, not ABI_ELFv2.
971 ("*call_value_indirect_aix<mode>"): Likewise.
972 ("*call_indirect_elfv2<mode>"): New pattern.
973 ("*call_value_indirect_elfv2<mode>"): Likewise.
975 * config/rs6000/predicates.md ("symbol_ref_operand"): Do not
976 check for function descriptors in ABI_ELFv2.
977 ("current_file_function_operand"): Likewise.
979 * config/rs6000/ppc-asm.h [__powerpc64__ && _CALL_ELF == 2]:
981 (FUNC_NAME): Define ELFv2 variant.
982 (JUMP_TARGET): Likewise.
983 (FUNC_START): Likewise.
984 (HIDDEN_FUNC): Likewise.
985 (FUNC_END): Likeiwse.
987 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
989 * config.gcc [powerpc*-*-* | rs6000-*-*]: Support --with-abi=elfv1
990 and --with-abi=elfv2.
991 * config/rs6000/option-defaults.h (OPTION_DEFAULT_SPECS): Add "abi".
992 * config/rs6000/rs6000.opt (mabi=elfv1): New option.
993 (mabi=elfv2): Likewise.
994 * config/rs6000/rs6000-opts.h (enum rs6000_abi): Add ABI_ELFv2.
995 * config/rs6000/linux64.h (DEFAULT_ABI): Do not hard-code to AIX_ABI
997 (ELFv2_ABI_CHECK): New macro.
998 (SUBSUBTARGET_OVERRIDE_OPTIONS): Use it to decide whether to set
999 rs6000_current_abi to ABI_AIX or ABI_ELFv2.
1000 (GLIBC_DYNAMIC_LINKER64): Support ELFv2 ld.so version.
1001 * config/rs6000/rs6000-c.c (rs6000_cpu_cpp_builtins): Predefine
1002 _CALL_ELF and __STRUCT_PARM_ALIGN__ if appropriate.
1004 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Handle ABI_ELFv2.
1005 (debug_stack_info): Likewise.
1006 (rs6000_file_start): Treat ABI_ELFv2 the same as ABI_AIX.
1007 (rs6000_legitimize_tls_address): Likewise.
1008 (rs6000_conditional_register_usage): Likewise.
1009 (rs6000_emit_move): Likewise.
1010 (init_cumulative_args): Likewise.
1011 (rs6000_function_arg_advance_1): Likewise.
1012 (rs6000_function_arg): Likewise.
1013 (rs6000_arg_partial_bytes): Likewise.
1014 (rs6000_output_function_entry): Likewise.
1015 (rs6000_assemble_integer): Likewise.
1016 (rs6000_savres_strategy): Likewise.
1017 (rs6000_stack_info): Likewise.
1018 (rs6000_function_ok_for_sibcall): Likewise.
1019 (rs6000_emit_load_toc_table): Likewise.
1020 (rs6000_savres_routine_name): Likewise.
1021 (ptr_regno_for_savres): Likewise.
1022 (rs6000_emit_prologue): Likewise.
1023 (rs6000_emit_epilogue): Likewise.
1024 (rs6000_output_function_epilogue): Likewise.
1025 (output_profile_hook): Likewise.
1026 (output_function_profiler): Likewise.
1027 (rs6000_trampoline_size): Likewise.
1028 (rs6000_trampoline_init): Likewise.
1029 (rs6000_elf_output_toc_section_asm_op): Likewise.
1030 (rs6000_elf_encode_section_info): Likewise.
1031 (rs6000_elf_reloc_rw_mask): Likewise.
1032 (rs6000_elf_declare_function_name): Likewise.
1033 (rs6000_function_arg_boundary): Treat ABI_ELFv2 the same as ABI_AIX,
1034 except that rs6000_compat_align_parm is always assumed false.
1035 (rs6000_gimplify_va_arg): Likewise.
1036 (rs6000_call_aix): Update comment.
1037 (rs6000_sibcall_aix): Likewise.
1038 * config/rs6000/rs6000.md ("tls_gd_aix<TLSmode:tls_abi_suffix>"):
1039 Treat ABI_ELFv2 the same as ABI_AIX.
1040 ("*tls_gd_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
1041 ("tls_ld_aix<TLSmode:tls_abi_suffix>"): Likewise.
1042 ("*tls_ld_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
1043 ("load_toc_aix_si"): Likewise.
1044 ("load_toc_aix_di"): Likewise.
1046 ("call_value"): Likewise.
1047 ("*call_local_aix<mode>"): Likewise.
1048 ("*call_value_local_aix<mode>"): Likewise.
1049 ("*call_nonlocal_aix<mode>"): Likewise.
1050 ("*call_value_nonlocal_aix<mode>"): Likewise.
1051 ("*call_indirect_aix<mode>"): Likewise.
1052 ("*call_value_indirect_aix<mode>"): Likewise.
1053 ("sibcall"): Likewise.
1054 ("sibcall_value"): Likewise.
1055 ("*sibcall_aix<mode>"): Likewise.
1056 ("*sibcall_value_aix<mode>"): Likewise.
1057 * config/rs6000/predicates.md ("symbol_ref_operand"): Likewise.
1058 ("current_file_function_operand"): Likewise.
1060 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1062 * config/rs6000/rs6000.c (rs6000_arg_partial_bytes): Simplify logic
1063 by making use of the fact that for vector / floating point arguments
1064 passed both in VRs/FPRs and in the fixed parameter area, the partial
1065 bytes mechanism is in fact not used.
1067 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1069 * config/rs6000/rs6000.c (rs6000_psave_function_arg): New function.
1070 (rs6000_finish_function_arg): Likewise.
1071 (rs6000_function_arg): Use rs6000_psave_function_arg and
1072 rs6000_finish_function_arg to handle both vector and floating
1073 point arguments that are also passed in GPRs / the stack.
1075 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1077 * config/rs6000/rs6000.c (USE_FP_FOR_ARG_P): Remove TYPE argument.
1078 (USE_ALTIVEC_FOR_ARG_P): Likewise.
1079 (rs6000_darwin64_record_arg_advance_recurse): Update uses.
1080 (rs6000_function_arg_advance_1):Likewise.
1081 (rs6000_darwin64_record_arg_recurse): Likewise.
1082 (rs6000_function_arg): Likewise.
1083 (rs6000_arg_partial_bytes): Likewise.
1085 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1087 * config/rs6000/rs6000.c (rs6000_option_override_internal): Replace
1088 "DEFAULT_ABI != ABI_AIX" test by testing for ABI_V4 or ABI_DARWIN.
1089 (rs6000_savres_strategy): Likewise.
1090 (rs6000_return_addr): Likewise.
1091 (rs6000_emit_load_toc_table): Replace "DEFAULT_ABI != ABI_AIX" by
1092 testing for ABI_V4 (since ABI_DARWIN is impossible here).
1093 (rs6000_emit_prologue): Likewise.
1094 (legitimate_lo_sum_address_p): Simplify DEFAULT_ABI test.
1095 (rs6000_elf_declare_function_name): Remove duplicated test.
1096 * config/rs6000/rs6000.md ("load_toc_v4_PIC_1"): Explicitly test
1097 for ABI_V4 (instead of "DEFAULT_ABI != ABI_AIX" test).
1098 ("load_toc_v4_PIC_1_normal"): Likewise.
1099 ("load_toc_v4_PIC_1_476"): Likewise.
1100 ("load_toc_v4_PIC_1b"): Likewise.
1101 ("load_toc_v4_PIC_1b_normal"): Likewise.
1102 ("load_toc_v4_PIC_1b_476"): Likewise.
1103 ("load_toc_v4_PIC_2"): Likewise.
1104 ("load_toc_v4_PIC_3b"): Likewise.
1105 ("load_toc_v4_PIC_3c"): Likewise.
1106 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Simplify DEFAULT_ABI test.
1107 (RS6000_SAVE_AREA): Likewise.
1108 (FP_ARG_MAX_REG): Likewise.
1109 (RETURN_ADDRESS_OFFSET): Likewise.
1110 * config/rs6000/sysv.h (TARGET_TOC): Test for ABI_V4 instead
1112 (SUBTARGET_OVERRIDE_OPTIONS): Likewise.
1113 (MINIMAL_TOC_SECTION_ASM_OP): Likewise.
1115 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1117 * config/rs6000/rs6000.c (rs6000_call_indirect_aix): Rename to ...
1118 (rs6000_call_aix): ... this. Handle both direct and indirect calls.
1119 Create call insn directly instead of via various gen_... routines.
1120 Mention special registers used by the call in CALL_INSN_FUNCTION_USAGE.
1121 (rs6000_sibcall_aix): New function.
1122 * config/rs6000/rs6000.md (TOC_SAVE_OFFSET_32BIT): Remove.
1123 (TOC_SAVE_OFFSET_64BIT): Likewise.
1124 (AIX_FUNC_DESC_TOC_32BIT): Likewise.
1125 (AIX_FUNC_DESC_TOC_64BIT): Likewise.
1126 (AIX_FUNC_DESC_SC_32BIT): Likewise.
1127 (AIX_FUNC_DESC_SC_64BIT): Likewise.
1128 ("call" expander): Call rs6000_call_aix.
1129 ("call_value" expander): Likewise.
1130 ("call_indirect_aix<ptrsize>"): Replace this pattern ...
1131 ("call_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
1132 ("*call_indirect_aix<mode>"): ... by this insn pattern.
1133 ("call_value_indirect_aix<ptrsize>"): Replace this pattern ...
1134 ("call_value_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
1135 ("*call_value_indirect_aix<mode>"): ... by this insn pattern.
1136 ("*call_nonlocal_aix32", "*call_nonlocal_aix64"): Replace by ...
1137 ("*call_nonlocal_aix<mode>"): ... this pattern.
1138 ("*call_value_nonlocal_aix32", "*call_value_nonlocal_aix64"): Replace
1139 ("*call_value_nonlocal_aix<mode>"): ... by this pattern.
1140 ("*call_local_aix<mode>"): New insn pattern.
1141 ("*call_value_local_aix<mode>"): Likewise.
1142 ("sibcall" expander): Call rs6000_sibcall_aix.
1143 ("sibcall_value" expander): Likewise. Move earlier in file.
1144 ("*sibcall_nonlocal_aix<mode>"): Replace by ...
1145 ("*sibcall_aix<mode>"): ... this pattern.
1146 ("*sibcall_value_nonlocal_aix<mode>"): Replace by ...
1147 ("*sibcall_value_aix<mode>"): ... this pattern.
1148 * config/rs6000/rs6000-protos.h (rs6000_call_indirect_aix): Remove.
1149 (rs6000_call_aix): Add prototype.
1150 (rs6000_sibcall_aix): Likewise.
1152 2013-11-14 Jakub Jelinek <jakub@redhat.com>
1155 * asan.c (asan_emit_stack_protection): Ensure -fsection-anchors
1156 isn't confused by the artificial decl.
1158 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1160 * config/rs6000/rs6000.c (rs6000_emit_prologue): Do not place a
1161 RTX_FRAME_RELATED_P marker on the UNSPEC_MOVESI_FROM_CR insn.
1162 Instead, add USEs of all modified call-saved CR fields to the
1163 insn storing the result to the stack slot, and provide an
1164 appropriate REG_FRAME_RELATED_EXPR for that insn.
1165 * config/rs6000/rs6000.md ("*crsave"): New insn pattern.
1166 * config/rs6000/predicates.md ("crsave_operation"): New predicate.
1168 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1169 Alan Modra <amodra@gmail.com>
1171 * function.c (assign_parms): Use all.reg_parm_stack_space instead
1172 of re-evaluating REG_PARM_STACK_SPACE target macro.
1173 (locate_and_pad_parm): New parameter REG_PARM_STACK_SPACE. Use it
1174 instead of evaluating target macro REG_PARM_STACK_SPACE every time.
1175 (assign_parm_find_entry_rtl): Update call.
1176 * calls.c (initialize_argument_information): Update call.
1177 (emit_library_call_value_1): Likewise.
1178 * expr.h (locate_and_pad_parm): Update prototype.
1180 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1182 * calls.c (store_unaligned_arguments_into_pseudos): Skip PARALLEL
1185 2013-11-14 DJ Delorie <dj@redhat.com>
1187 * config/rx/rx.c (rx_mode_dependent_address_p): Allow offsets up
1190 2013-11-14 Jeff Law <law@redhat.com>
1192 * tree-ssa-threadedge.c (thread_through_normal_block): Only push the
1193 EDGE_START_JUMP_THREAD marker if the jump threading path is empty.
1195 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
1197 * doc/invoke.texi: Update documentation for AArch64's -mcpu
1200 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
1202 * config/aarch64/aarch64-cores.def (example-1): Remove.
1203 (example-2): Likewise.
1204 * config/aarch64/aarch64-tune.md: Regenerate.
1205 * config/aarch64/aarch64.md: Do not include "large.md" or "small.md".
1206 (generic_sched): Remove "large", "small".
1207 * config/aarch64/large.md: Delete.
1208 * config/aarch64/small.md: Delete.
1210 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
1212 * config/aarch64/aarch64-cores.def (cortex-a57): Tune for cortexa15.
1213 * config/aarch64/aarch64-tune.md: Regenerate.
1214 * config/aarch64/aarch64.md: Include cortex-a15 pipeline model.
1215 (generic_sched): "no" if we are tuning for cortexa15.
1216 * config/arm/cortex-a15.md: Include cortex-a15-neon.md by
1219 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
1221 * config/aarch64/aarch64-arches.def (armv8-a): Tune for cortex-a53.
1222 * config/aarch64/aarch64.md: Do not include aarch64-generic.md.
1223 * config/aarch64/aarch64.c (aarch64_tune): Initialize to cortexa53.
1224 (all_cores): Use cortexa53 when tuning for "generic".
1225 (aarch64_override_options): Fix comment.
1226 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Set to cortexa53.
1227 * config/aarch64/aarch64-generic.md: Delete.
1229 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
1231 * config/aarch64/aarch64.c (all_architectures): Remove "generic".
1233 2013-11-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1235 * config/aarch64/aarch64.c: Include aarch-cost-tables.h.
1236 (generic_rtx_cost_table): Remove.
1237 (aarch64_rtx_costs): Use fields from cpu_cost_table.
1238 * config/aarch64/aarch64-protos.h (tune_params): Use cpu_cost_table for
1240 (cpu_rtx_cost_table): Remove.
1242 2013-11-14 Julian Brown <julian@codesourcery.com>
1243 Joey Ye <joey.ye@arm.com>
1245 * config/arm/arm.c (arm_cortex_m_branch_cost): New.
1246 (arm_v7m_tune): New.
1247 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune, arm_9e_tune,
1248 arm_v6t2_tune, arm_cortex_tune, arm_cortex_a15_tune,
1249 arm_cortex_a5_tune, arm_v6m_tune): Add comments for Sched adj cost.
1250 * config/arm/arm-cores.def (cortex-m4, cortex-m3): Use arm_v7m_tune.
1252 2013-11-14 Kirill Yukhin <kirill.yukhin@intel.com>
1255 * config/ia64/ia64.c (ia64_split_tmode_move): Relax `dead'
1258 2013-11-14 Jakub Jelinek <jakub@redhat.com>
1259 Uros Bizjak <ubizjak@gmail.com>
1262 * config/i386/i386.md (*anddi_2): Only allow CCZmode if
1263 operands[2] satisfies_constraint_Z that might have bit 31 set.
1265 2013-11-13 Jeff Law <law@redhat.com>
1267 PR tree-optimization/59102
1268 * gimple-ssa-isolate-paths.c
1269 (insert_trap_and_remove_trailing_statments): Ensure STMT is a
1270 gimple assignment before looking at gimple_assign_lhs.
1272 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
1274 * ira.c: Add comment about threads at the top of file.
1276 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
1278 * ira-color.c (coalesce_allocnos): Don't allocate and free
1281 2013-11-14 Tom de Vries <tom@codesourcery.com>
1283 * tree-ssa-tail-merge.c (gimple_equal_p): Add test for structural
1284 equality for GIMPLE_ASSIGN.
1286 2013-11-14 Tom de Vries <tom@codesourcery.com>
1288 * tree-ssa-tail-merge.c (gimple_operand_equal_value_p): Factor new
1290 (gimple_equal_p): ... here.
1292 2013-11-14 Tom de Vries <tom@codesourcery.com>
1294 * trans-mem.c (is_tm_ending): New function.
1295 * gimple.h (is_tm_ending): Declare.
1296 * tree-ssa-tail-merge.c (gimple_equal_p): Remove test on
1298 (find_duplicate): Use is_tm_ending instead of is_tm_ending_fndecl.
1300 2013-11-14 Tom de Vries <tom@codesourcery.com>
1302 * tree-ssa-tail-merge.c (gimple_equal_p): Remove equal variable.
1304 2013-11-13 Andrew MacLeod <amacleod@redhat.com>
1306 * gimple-walk.h: New File. Relocate prototypes from gimple.h.
1307 (struct walk_stmt_info): Relocate here from gimple.h.
1308 * gimple-iterator.h: New File. Relocate prototypes from gimple.h.
1309 (struct gimple_stmt_iterator_d): Relocate here from gimple.h.
1310 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
1311 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
1312 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
1313 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
1314 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Relocate here from gimple.h.
1315 * gimple.h (struct gimple_stmt_iterator_d): Move to gimple-iterator.h.
1316 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
1317 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
1318 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
1319 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
1320 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Move to gimple-iterator.h.
1321 (struct walk_stmt_info): Move to gimple-walk.h.
1322 (gimple_seq_set_location): Move to gimple.c
1323 * gimple-walk.c: New File.
1324 (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm, walk_gimple_op,
1325 walk_gimple_stmt, get_base_loadstore, walk_stmt_load_store_addr_ops,
1326 walk_stmt_load_store_ops): Relocate here from gimple.c.
1327 * gimple-iterator.c: Include gimple-iterator.h.
1328 * gimple.c (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm,
1329 walk_gimple_op, walk_gimple_stmt, get_base_loadstore,
1330 walk_stmt_load_store_addr_ops, walk_stmt_load_store_ops): Move to
1332 (gimple_seq_set_location): Relocate from gimple.h.
1333 * tree-phinodes.h (set_phi_nodes): Move to tree-phinodes.c.
1334 * tree-phinodes.c (set_phi_nodes): Relocate from tree-phinodes.h.
1335 * gengtype.c (open_base_files): Add gimple-iterator.h to include list.
1336 * Makefile.in (OBJS): Add gimple-walk.o
1337 * asan.c: Update Include list as required for gimple-iterator.h and
1339 * cfgexpand.c: Likewise.
1340 * cfgloop.c: Likewise.
1341 * cfgloopmanip.c: Likewise.
1342 * cgraph.c: Likewise.
1343 * cgraphbuild.c: Likewise.
1344 * cgraphunit.c: Likewise.
1345 * gimple-fold.c: Likewise.
1346 * gimple-low.c: Likewise.
1347 * gimple-pretty-print.c: Likewise.
1348 * gimple-ssa-isolate-paths.c: Likewise.
1349 * gimple-ssa-strength-reduction.c: Likewise.
1350 * gimple-streamer-in.c: Likewise.
1351 * gimple-streamer-out.c: Likewise.
1352 * gimplify.c: Likewise.
1353 * graphite-blocking.c: Likewise.
1354 * graphite-clast-to-gimple.c: Likewise.
1355 * graphite-dependences.c: Likewise.
1356 * graphite-interchange.c: Likewise.
1357 * graphite-optimize-isl.c: Likewise.
1358 * graphite-poly.c: Likewise.
1359 * graphite-scop-detection.c: Likewise.
1360 * graphite-sese-to-poly.c: Likewise.
1361 * graphite.c: Likewise.
1362 * ipa-inline-analysis.c: Likewise.
1363 * ipa-profile.c: Likewise.
1364 * ipa-prop.c: Likewise.
1365 * ipa-pure-const.c: Likewise.
1366 * ipa-split.c: Likewise.
1367 * lto-streamer-in.c: Likewise.
1368 * lto-streamer-out.c: Likewise.
1369 * omp-low.c: Likewise.
1370 * predict.c: Likewise.
1371 * profile.c: Likewise.
1373 * tracer.c: Likewise.
1374 * trans-mem.c: Likewise.
1375 * tree-call-cdce.c: Likewise.
1376 * tree-cfg.c: Likewise.
1377 * tree-cfgcleanup.c: Likewise.
1378 * tree-complex.c: Likewise.
1379 * tree-data-ref.c: Likewise.
1380 * tree-dfa.c: Likewise.
1381 * tree-eh.c: Likewise.
1382 * tree-emutls.c: Likewise.
1383 * tree-if-conv.c: Likewise.
1384 * tree-inline.c: Likewise.
1385 * tree-into-ssa.c: Likewise.
1386 * tree-loop-distribution.c: Likewise.
1387 * tree-nested.c: Likewise.
1388 * tree-nrv.c: Likewise.
1389 * tree-object-size.c: Likewise.
1390 * tree-outof-ssa.c: Likewise.
1391 * tree-parloops.c: Likewise.
1392 * tree-predcom.c: Likewise.
1393 * tree-profile.c: Likewise.
1394 * tree-scalar-evolution.c: Likewise.
1395 * tree-sra.c: Likewise.
1396 * tree-ssa-ccp.c: Likewise.
1397 * tree-ssa-coalesce.c: Likewise.
1398 * tree-ssa-copy.c: Likewise.
1399 * tree-ssa-copyrename.c: Likewise.
1400 * tree-ssa-dce.c: Likewise.
1401 * tree-ssa-dom.c: Likewise.
1402 * tree-ssa-dse.c: Likewise.
1403 * tree-ssa-forwprop.c: Likewise.
1404 * tree-ssa-ifcombine.c: Likewise.
1405 * tree-ssa-live.c: Likewise.
1406 * tree-ssa-loop-ch.c: Likewise.
1407 * tree-ssa-loop-im.c: Likewise.
1408 * tree-ssa-loop-ivcanon.c: Likewise.
1409 * tree-ssa-loop-ivopts.c: Likewise.
1410 * tree-ssa-loop-manip.c: Likewise.
1411 * tree-ssa-loop-niter.c: Likewise.
1412 * tree-ssa-loop-prefetch.c: Likewise.
1413 * tree-ssa-loop.c: Likewise.
1414 * tree-ssa-math-opts.c: Likewise.
1415 * tree-ssa-phiopt.c: Likewise.
1416 * tree-ssa-phiprop.c: Likewise.
1417 * tree-ssa-pre.c: Likewise.
1418 * tree-ssa-propagate.c: Likewise.
1419 * tree-ssa-reassoc.c: Likewise.
1420 * tree-ssa-sink.c: Likewise.
1421 * tree-ssa-strlen.c: Likewise.
1422 * tree-ssa-structalias.c: Likewise.
1423 * tree-ssa-tail-merge.c: Likewise.
1424 * tree-ssa-ter.c: Likewise.
1425 * tree-ssa-threadedge.c: Likewise.
1426 * tree-ssa-threadupdate.c: Likewise.
1427 * tree-ssa-uncprop.c: Likewise.
1428 * tree-ssa-uninit.c: Likewise.
1429 * tree-ssa.c: Likewise.
1430 * tree-stdarg.c: Likewise.
1431 * tree-switch-conversion.c: Likewise.
1432 * tree-tailcall.c: Likewise.
1433 * tree-vect-data-refs.c: Likewise.
1434 * tree-vect-generic.c: Likewise.
1435 * tree-vect-loop-manip.c: Likewise.
1436 * tree-vect-loop.c: Likewise.
1437 * tree-vect-patterns.c: Likewise.
1438 * tree-vect-slp.c: Likewise.
1439 * tree-vect-stmts.c: Likewise.
1440 * tree-vectorizer.c: Likewise.
1441 * tree-vrp.c: Likewise.
1444 * value-prof.c: Likewise.
1445 * vtable-verify.c: Likewise.
1447 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
1449 * gimple-ssa-isolate-paths.c (pass_isolate_erroneous_paths): Comment
1452 2013-11-13 Jeff Law <law@redhat.com>
1454 * PR middle-end/59119
1455 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behaviour): New
1456 function, extracted from gimple_ssa_isolate_erroneous_paths.
1457 (find_explicit_erroneous_behaviour): Similarly.
1458 (insert_trap_and_remove_trailing_statements): Remove statements
1461 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
1463 * cfgrtl.c (can_fallthru): Reorder code to move tablejump check up.
1464 Make that check explicit. BB_HEAD cannot be NULL, remove check for it.
1465 * haifa-sched.c (ready_remove_first_dispatch): Check INSN_P before
1466 looking at INSN_CODE.
1467 * reload1.c (delete_dead_insn) Do not expect JUMP_TABLE_DATA to be an
1468 active_insn_p object, respect basic block boundaries.
1469 * reorg.c (follow_jumps): Use invariant that JUMP_TABLE_DATA always
1470 follows immediately after the jump table data label.
1471 * config/nds32/nds32.c (nds32_output_casesi_pc_relative): Likewise.
1472 * config/sh/sh.c (barrier_align): Likewise. Rearrange code such
1473 that JUMP_TABLE_DATA is not expected to be an active_insn_p object.
1475 2013-11-13 Teresa Johnson <tejohnson@google.com>
1478 * predict.c (drop_profile): Error is currently too strict.
1479 (handle_missing_profiles): Pass call_count to drop_profile.
1481 2013-11-13 Teresa Johnson <tejohnson@google.com>
1484 * ipa-inline.c (edge_badness): Fix overflow.
1486 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
1488 PR rtl-optimization/59036
1489 * ira-color.c (struct allocno_color_data): Add new members
1490 first_thread_allocno, next_thread_allocno, thread_freq.
1491 (sorted_copies): New static var.
1492 (allocnos_conflict_by_live_ranges_p, copy_freq_compare_func): Move up.
1493 (allocno_thread_conflict_p, merge_threads)
1494 (form_threads_from_copies, form_threads_from_bucket)
1495 (form_threads_from_colorable_allocno, init_allocno_threads): New
1497 (bucket_allocno_compare_func): Add comparison by thread frequency
1499 (add_allocno_to_ordered_bucket): Rename to
1500 add_allocno_to_ordered_colorable_bucket. Remove parameter.
1501 (push_only_colorable): Call form_threads_from_bucket.
1502 (color_pass): Call init_allocno_threads. Use
1503 consideration_allocno_bitmap instead of coloring_allocno_bitmap
1504 for nuillify allocno color data.
1505 (ira_initiate_assign, ira_finish_assign): Allocate/free sorted_copies.
1506 (coalesce_allocnos): Use static sorted copies.
1508 2013-11-13 Jakub Jelinek <jakub@redhat.com>
1510 * passes.c (execute_todo): Don't call do_per_function if
1512 (execute_one_ipa_transform_pass, execute_one_pass): Don't call
1513 execute_function_dump if dump_file is NULL.
1515 2013-11-13 Martin Jambor <mjambor@suse.cz>
1517 * cgraph.c (cgraph_get_create_node): Do what
1518 cgraph_get_create_real_symbol_node used to do.
1519 (cgraph_get_create_real_symbol_node): Removed. Changed all users to
1520 call cgraph_get_create_node.
1521 * cgraph.h (cgraph_get_create_real_symbol_node): Removed.
1522 * lto-streamer-in.c (input_function): Call cgraph_get_node instead of
1523 cgraph_get_create_node. Assert we get a node.
1525 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
1527 * config/aarch64/aarch64-simd.md (vec_extract): New.
1529 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
1531 * config/aarch64/aarch64-simd.md (vec_set<mode>): Add w -> w option to
1534 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
1536 * cfgexpand.c (expand_used_vars): Allocate space for partitions based
1537 on PARM_DECLs or RESULT_DECLs only if they are ignored for debug info
1538 or if optimization is enabled.
1539 * tree-ssa-coalesce.c (coalesce_ssa_name): If optimization is disabled,
1540 require that all the names based on a PARM_DECL or a RESULT_DECL that
1541 isn't ignored for debug info be coalesced.
1543 2013-11-13 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1545 * config/c6x/c6x.c: Include "gimple-expr.h".
1547 2013-11-13 Richard Biener <rguenther@suse.de>
1549 * gimple-streamer-out.c (output_gimple_stmt): Also wrap
1550 decls in ADDR_EXPR operands inside a MEM_REF and optimize that.
1551 * gimple-streamer-in.c (input_gimple_stmt): Remove now dead code
1552 dealing with type mismatches inside component reference chains.
1554 2013-11-13 Marc Glisse <marc.glisse@inria.fr>
1556 PR tree-optimization/59077
1557 * ipa-pure-const.c (better_state): Update *state.
1559 2013-11-13 Christophe Lyon <christophe.lyon@linaro.org>
1561 * config/aarch64/aarch64.h (FRAME_GROWS_DOWNWARD): Define to 1.
1562 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1563 Update offset calculations.
1565 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
1568 * dwarf2out.c (add_byte_size_attribute): Also use int_size_in_bytes
1569 for fields. Do not add the attribute if the size is negative.
1571 2013-11-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1573 * config/arm/arm.c: Include aarch-cost-tables.h.
1574 (generic_extra_costs): Move from here...
1575 * config/arm/aarch-cost-tables.h: ... To here. New file.
1577 2013-11-13 Alexander Ivchenko <alexander.ivchenko@intel.com>
1578 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1579 Sergey Lega <sergey.s.lega@intel.com>
1580 Anna Tikhonova <anna.tikhonova@intel.com>
1581 Ilya Tocar <ilya.tocar@intel.com>
1582 Andrey Turetskiy <andrey.turetskiy@intel.com>
1583 Ilya Verbin <ilya.verbin@intel.com>
1584 Kirill Yukhin <kirill.yukhin@intel.com>
1585 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1587 * config/i386/i386.c (ix86_print_operand): Support z-masking.
1588 * config/i386/predicate.md (const_0_to_4_operand): New.
1589 (const_0_to_5_operand): Ditto.
1590 * config/i386/sse.md (UNSPEC_COMPRESS): New.
1591 (UNSPEC_COMPRESS_STORE): Ditto.
1592 (UNSPEC_EXPAND): Ditto.
1593 (UNSPEC_EMBEDDED_ROUNDING): Ditto.
1594 (define_mode_attr ssescalarsize): Ditto.
1595 (avx512f_load<mode>_mask): Ditto.
1596 (avx512f_store<mode>_mask): Ditto.
1597 (avx512f_storedqu<mode>_mask): Ditto.
1598 (avx512f_vmcmp<mode>3_mask): Ditto.
1599 (avx512f_fmadd_<mode>_mask): Ditto.
1600 (avx512f_fmadd_<mode>_mask3): Ditto.
1601 (avx512f_fmsub_<mode>_mask): Ditto.
1602 (avx512f_fmsub_<mode>_mask3): Ditto.
1603 (avx512f_fnmadd_<mode>_mask): Ditto.
1604 (avx512f_fnmadd_<mode>_mask3): Ditto.
1605 (avx512f_fnmsub_<mode>_mask): Ditto.
1606 (avx512f_fnmsub_<mode>_mask3): Ditto.
1607 (avx512f_fmaddsub_<mode>_mask): Ditto.
1608 (avx512f_fmaddsub_<mode>_mask3): Ditto.
1609 (avx512f_fmsubadd_<mode>_mask): Ditto.
1610 (avx512f_fmsubadd_<mode>_mask3): Ditto.
1611 (vec_unpacku_float_lo_v16si): Ditto.
1612 (avx512f_vextract<shuffletype>32x4_mask): Ditto.
1613 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
1614 (avx512f_vextract<shuffletype>64x4_mask): Ditto.
1615 (vec_extract_lo_<mode>_maskm): Ditto.
1616 (vec_extract_hi_<mode>_maskm): Ditto.
1617 (avx512f_vternlog<mode>_mask): Ditto.
1618 (avx512f_shufps512_mask): Ditto.
1619 (avx512f_fixupimm<mode>_mask): Ditto.
1620 (avx512f_shufpd512_mask): Ditto.
1621 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
1622 (avx512f_<code>v8div16qi2_mask/trunc): Ditto.
1623 (*avx512f_<code>v8div16qi2_store_mask): Ditto.
1624 (ashr<mode>3<mask_name>): Ditto.
1625 (avx512f_vinsert<shuffletype>32x4_mask): Ditto.
1626 (avx512f_vinsert<shuffletype>64x4_mask): Ditto.
1627 (avx512f_shuf_<shuffletype>64x2_mask): Ditto.
1628 (avx512f_shuf_<shuffletype>32x4_mask): Ditto.
1629 (avx512f_pshufdv3_mask): Ditto.
1630 (avx512f_perm<mode>_mask): Ditto.
1631 (avx512f_vpermi2var<mode>3_mask): Ditto.
1632 (avx512f_vpermt2var<mode>3_mask): Ditto.
1633 (avx512f_compress<mode>_mask): Ditto.
1634 (avx512f_compressstore<mode>_mask): Ditto.
1635 (avx512f_expand<mode>_mask): Ditto.
1636 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>): Extend
1638 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
1639 (<plusminus_insn><mode>3<mask_name>): Ditto.
1640 (*<plusminus_insn><mode>3<mask_name>): Ditto.
1641 (mul<mode>3<mask_name>): Ditto.
1642 (*mul<mode>3<mask_name>): Ditto.
1643 (<sse>_div<mode>3<mask_name>): Ditto.
1644 (<mask_codefor>rcp14<mode><mask_name>): Ditto.
1645 (<sse>_sqrt<mode>2<mask_name>): Ditto.
1646 (<mask_codefor>rsqrt14<mode><mask_name>): Ditto.
1647 (<code><mode>3<mask_name>/smaxmin): Ditto.
1648 (*<code><mode>3_finite<mask_name>/smaxmin): Ditto.
1649 (*<code><mode>3<mask_name>/smaxmin): Ditto.
1650 (float<sseintvecmodelower><mode>2<mask_name>): Ditto.
1651 (ufloatv16siv16sf2<mask_name>): Ditto.
1652 (<mask_codefor>avx512f_fix_notruncv16sfv16si<mask_name>): Ditto.
1653 (<mask_codefor>avx512f_ufix_notruncv16sfv16si<mask_name>): Ditto.
1654 (<fixsuffix>fix_truncv16sfv16si2<mask_name>): Ditto.
1655 (float<si2dfmodelower><mode>2<mask_name>): Ditto.
1656 (ufloatv8siv8df<mask_name>): Ditto.
1657 (<mask_codefor>avx512f_cvtpd2dq512<mask_name>): Ditto.
1658 (avx512f_ufix_notruncv8dfv8si<mask_name>): Ditto.
1659 (<fixsuffix>fix_truncv8dfv8si2<mask_name>): Ditto.
1660 (<mask_codefor>avx512f_cvtpd2ps512<mask_name>): Ditto.
1661 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix><mask_name>): Ditto.
1662 (<mask_codefor>avx512f_unpckhps512<mask_name>): Ditto.
1663 (<mask_codefor>avx512f_unpcklps512<mask_name>): Ditto.
1664 (<mask_codefor>avx512f_movshdup512<mask_name>): Ditto.
1665 (<mask_codefor>avx512f_movsldup512<mask_name>): Ditto.
1666 (<mask_codefor>avx512f_vextract<shuffletype>32x4_1<mask_name>): Ditto.
1667 (vec_extract_lo_<mode><mask_name>): Ditto.
1668 (vec_extract_hi_<mode><mask_name>): Ditto.
1669 (<mask_codefor>avx512f_unpckhpd512<mask_name>): Ditto.
1670 (avx512f_movddup512<mask_name>): Ditto.
1671 (avx512f_unpcklpd512<mask_name>): Ditto.
1672 (*avx512f_unpcklpd512<mask_name>): Ditto.
1673 (*avx512f_vmscalef<mode>): Ditto.
1674 (avx512f_scalef<mode><mask_name>): Ditto.
1675 (avx512f_getexp<mode><mask_name>): Ditto.
1676 (<mask_codefor>avx512f_align<mode><mask_name>): Ditto.
1677 (avx512f_rndscale<mode><mask_name>): Ditto.
1678 (avx512f_shufps512_1<mask_name>): Ditto.
1679 (avx512f_shufpd512_1<mask_name>): Ditto.
1680 (<plusminus_insn><mode>3<mask_name>): Ditto.
1681 (*<plusminus_insn><mode>3<mask_name>): Ditto.
1682 (vec_widen_umult_even_v16si<mask_name>): Ditto.
1683 (*vec_widen_umult_even_v16si<mask_name>): Ditto.
1684 (vec_widen_smult_even_v16si<mask_name>): Ditto.
1685 (*vec_widen_smult_even_v16si<mask_name>): Ditto.
1686 (mul<mode>3<mask_name>): Ditto.
1687 (*<sse4_1_avx2>_mul<mode>3<mask_name>): Ditto.
1688 (<shift_insn><mode>3<mask_name>): Ditto.
1689 (avx512f_<rotate>v<mode><mask_name>/rotate): Ditto.
1690 (avx512f_<rotate><mode><mask_name>): Ditto.
1691 (<code><mode>3<mask_name>/maxmin): Ditto.
1692 (*avx2_<code><mode>3<mask_name>/maxmin): Ditto.
1693 (<sse2_avx2>_andnot<mode>3<mask_name>): Ditto.
1694 (*andnot<mode>3<mask_name>): Ditto.
1695 (<mask_codefor><code><mode>3<mask_name>/any_logic): Ditto.
1696 (<mask_codefor>avx512f_interleave_highv16si<mask_name>): Ditto.
1697 (<mask_codefor>avx512f_interleave_lowv16si<mask_name>): Ditto.
1698 (<mask_codefor>avx512f_vinsert<shuffletype>32x4_1<mask_name>): Ditto.
1699 (vec_set_lo_<mode><mask_name>): Ditto.
1700 (vec_set_hi_<mode><mask_name>): Ditto.
1701 (avx512f_shuf_<shuffletype>64x2_1<mask_name>): Ditto.
1702 (avx512f_shuf_<shuffletype>32x4_1<mask_name>): Ditto.
1703 (avx512f_pshufd_1<mask_name>): Ditto.
1704 (<mask_codefor>abs<mode>2<mask_name>): Ditto.
1705 (<mask_codefor>avx512f_<code>v16qiv16si2<mask_name>): Ditto.
1706 (avx512f_<code>v16hiv16si2<mask_name>/any_extend): Ditto.
1707 (avx512f_<code>v8qiv8di2<mask_name>/any_extend): Ditto.
1708 (avx512f_<code>v8hiv8di2<mask_name>/any_extend): Ditto.
1709 (avx512f_<code>v8siv8di2<mask_name>/any_extend): Ditto.
1710 (avx512er_exp2<mode><mask_name>): Ditto.
1711 (<mask_codefor>avx512er_rcp28<mode><mask_name>): Ditto.
1712 (<mask_codefor>avx512er_rsqrt28<mode><mask_name>): Ditto.
1713 (<avx2_avx512f>_permvar<mode><mask_name>): Ditto.
1714 (<avx2_avx512f>_perm<mode>_1<mask_name>): Ditto.
1715 (<mask_codefor>avx512f_vec_dup<mode><mask_name>): Ditto.
1716 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V16FI): Ditto.
1717 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V8FI): Ditto.
1718 (<mask_codefor>avx512f_vec_dup_gpr<mode><mask_name>): Ditto.
1719 (<mask_codefor>avx512f_vec_dup_mem<mode><mask_name>): Ditto.
1720 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF2): Ditto.
1721 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF1): Ditto.
1722 (*<sse2_avx_avx512f>_vpermilp<mode><mask_name>): Ditto.
1723 (<sse2_avx_avx512f>_vpermilvar<mode>3<mask_name>): Ditto.
1724 (<avx2_avx512f>_ashrv<mode><mask_name>): Ditto.
1725 (<avx2_avx512f>_<shift_insn>v<mode><mask_name>): Ditto.
1726 (<mask_codefor>avx512f_vcvtph2ps512<mask_name>): Ditto.
1727 (<mask_codefor>avx512f_vcvtps2ph512<mask_name>): Ditto.
1728 (avx512f_getmant<mode><mask_name>): Ditto.
1729 (clz<mode>2<mask_name>): Ditto.
1730 (<mask_codefor>conflict<mode><mask_name>): Ditto.
1731 (*srcp14<mode>): Remove visibility.
1732 (*rsqrt14<mode>): Ditto.
1733 (*fma_fmsub_<mode>): Ditto.
1734 (*fma_fnmadd_<mode>): Ditto.
1735 (*avx512f_rndscale<mode>): Ditto.
1736 * config/i386/subst.md: New file.
1738 2013-11-13 Joseph Myers <joseph@codesourcery.com>
1740 * doc/extend.texi (Statement Exprs, Typeof): Discuss __auto_type.
1741 * ginclude/stdatomic.h (kill_dependency, atomic_store_explicit)
1742 (atomic_load_explicit, atomic_exchange_explicit)
1743 (atomic_compare_exchange_strong_explicit)
1744 (atomic_compare_exchange_weak_explicit): Use __auto_type to
1745 declare variable initialized with PTR argument.
1747 2013-11-12 Jeff Law <law@redhat.com>
1749 * tree-ssa-threadedge.c (thread_around_empty_blocks): New argument
1750 backedge_seen_p. Set, use and pass it to children appropriately.
1751 (thread_through_normal_block): Similarly.
1752 (thread_across_edge): Similarly.
1754 * gimple-ssa-isolate-paths.c (check_loadstore): Mark discovered
1755 memory references as volatile.
1756 (insert_trap_and_remove_trailing_statements): Fix comment.
1758 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
1761 * ira-costs.c (record_operand_costs): Check operands number for
1764 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1767 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
1768 constraints that would allow DImode into the traditional Altivec
1769 registers, but cause undesirable code generation when loading 0 as
1771 (movdi_internal64): Likewise.
1772 (cmp<mode>_fpr): Do not use %x for CR register output.
1773 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
1774 -mallow-upper-sf debug switches are used.
1776 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
1778 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
1779 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
1780 Relocate prototypes from gimple.h.
1781 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
1782 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
1784 * gimple.h: Move some prototypes to gimplify.h.
1785 (gimple_predicate, enum fallback, enum gimplify_status): Move to
1787 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
1788 Relocate from gimpify.c.
1789 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1790 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
1791 is_gimple_reg_rhs) Relocate from gimplify.c.
1792 * gimplify.c (mark_addressable): Move to gimple-expr.c.
1793 (gimple_seq_add_stmt_without_update): Move to gimple.c.
1794 (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1795 create_tmp_var_raw, create_tmp_var, create_tmp_reg,
1796 is_gimple_reg_rhs): Move to gimple-expr.c.
1797 (should_carry_location_p): Move to gimple.c.
1798 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
1800 (annotate_one_with_location, annotate_all_with_location_after,
1801 annotate_all_with_location): Move to gimple.c.
1802 (compare_case_labels, sort_case_labels,
1803 preprocess_case_label_vec_for_gimple): Move to gimple.c.
1804 (rhs_predicate_for): Make static.
1805 (gimplify_assign): Relocate from gimple.c.
1806 * gimple.c (gimplify_assign): Move to gimplify.c.
1807 (gimple_seq_add_stmt_without_update, should_carry_location_p,
1808 annotate_one_with_location, annotate_all_with_location_after,
1809 annotate_all_with_location, compare_case_labels, sort_case_labels,
1810 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
1811 * tree.h (unshare_expr, unshare_expr_without_location,
1812 mark_addressable): Move prototypes to gimplify.h.
1813 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
1815 * asan.c: Include gimplify.h rather than gimple.h.
1816 * cfgloopmanip.c: Likewise.
1817 * cgraphunit.c: Likewise.
1818 * cilk-common.c: Likewise.
1819 * dwarf2out.c: Dont include gimple.h.
1820 * fold-const.c: Include gimplify.h rather than gimple.h.
1821 * function.c: Likewise.
1822 * gimple-fold.c: Likewise.
1823 * gimple-ssa-strength-reduction.c: Likewise.
1824 * graphite-clast-to-gimple.c: Likewise.
1825 * graphite-sese-to-poly.c: Likewise.
1826 * ipa-prop.c: Likewise.
1827 * ipa-split.c: Likewise.
1829 * langhooks.c: Dont include gimple.h.
1830 * loop-init.c: Include gimplify.h rather than gimple.h.
1831 * omp-low.c: Likewise.
1833 * stor-layout.c: Likewise.
1834 * targhooks.c: Likewise.
1835 * trans-mem.c: Likewise.
1836 * tree-affine.c: Likewise.
1837 * tree-cfg.c: Likewise.
1838 * tree-cfgcleanup.c: Likewise.
1839 * tree-complex.c: Likewise.
1840 * tree-if-conv.c: Likewise.
1841 * tree-inline.c: Likewise.
1842 * tree-iterator.c: Likewise.
1843 * tree-loop-distribution.c: Likewise.
1844 * tree-nested.c: Likewise.
1845 * tree-parloops.c: Likewise.
1846 * tree-predcom.c: Likewise.
1847 * tree-profile.c: Likewise.
1848 * tree-scalar-evolution.c: Likewise.
1849 * tree-sra.c: Likewise.
1850 * tree-ssa-address.c: Likewise.
1851 * tree-ssa-ccp.c: Likewise.
1852 * tree-ssa-dce.c: Likewise.
1853 * tree-ssa-forwprop.c: Likewise.
1854 * tree-ssa-ifcombine.c: Likewise.
1855 * tree-ssa-loop-im.c: Likewise.
1856 * tree-ssa-loop-ivopts.c: Likewise.
1857 * tree-ssa-loop-manip.c: Likewise.
1858 * tree-ssa-loop-niter.c: Likewise.
1859 * tree-ssa-loop-prefetch.c: Likewise.
1860 * tree-ssa-loop-unswitch.c: Likewise.
1861 * tree-ssa-math-opts.c: Likewise.
1862 * tree-ssa-phiopt.c: Likewise.
1863 * tree-ssa-phiprop.c: Likewise.
1864 * tree-ssa-pre.c: Likewise.
1865 * tree-ssa-propagate.c: Likewise.
1866 * tree-ssa-reassoc.c: Likewise.
1867 * tree-ssa-sccvn.c: Likewise.
1868 * tree-ssa-strlen.c: Likewise.
1869 * tree-ssa.c: Likewise.
1870 * tree-switch-conversio: Likewise.n.c
1871 * tree-tailcall.c: Likewise.
1872 * tree-vect-data-refs.c: Likewise.
1873 * tree-vect-generic.c: Likewise.
1874 * tree-vect-loop-manip.c: Likewise.
1875 * tree-vect-loop.c: Likewise.
1876 * tree-vect-patterns.c: Likewise.
1877 * tree-vect-stmts.c: Likewise.
1879 * value-prof.c: Likewise.
1880 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
1881 * config/alpha/alpha.c: Likewise.
1882 * config/darwin.c: Likewise.
1883 * config/i386/i386.c: Likewise.
1884 * config/ia64/ia64.c: Likewise.
1885 * config/mep/mep.c: Likewise.
1886 * config/mips/mips.c: Likewise.
1887 * config/rs6000/rs6000.c: Likewise.
1888 * config/s390/s390.c: Likewise.
1889 * config/sh/sh.c: Likewise.
1890 * config/sparc/sparc.c: Likewise.
1891 * config/spu/spu.c: Likewise.
1892 * config/stormy16/stormy16.c: Likewise.
1893 * config/tilegx/tilegx.c: Likewise.
1894 * config/tilepro/tilepro.c: Likewise.
1895 * config/xtensa/xtensa.c: Likewise.
1897 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
1899 * tree.c (grow_tree_vec_stat): New function ...
1900 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
1901 and macro front-end.
1903 2013-11-12 Marek Polacek <polacek@redhat.com>
1905 * final.c (update_alignments): Initialize label to NULL_RTX.
1907 2013-11-12 Jeff Law <law@redhat.com>
1909 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
1910 (insert_trap_and_remove_trailing_statements): New argument OP which
1911 is the NULL pointer. Emit the trap after the load/store through
1912 the NULL pointer. Simplify the RHS of a store through a NULL pointer
1913 when trivial to do so.
1914 (isolate_path): Corresponding changes.
1915 (gimple_ssa_isolate_erroneous_path): Likewise.
1917 2013-11-12 Teresa Johnson <tejohnson@google.com>
1918 Jan Hubicka <jh@suse.cz>
1920 * predict.c (drop_profile): New function.
1921 (handle_missing_profiles): Ditto.
1922 (counts_to_freqs): Don't overwrite estimated frequencies
1923 when function has no profile counts.
1924 * predict.h (handle_missing_profiles): Declare.
1925 * tree-inline.c (freqs_to_counts): New function.
1926 (copy_cfg_body): Invoke freqs_to_counts as needed.
1927 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
1929 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1932 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
1934 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
1936 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1939 * config/i386/i386.c (ix86_option_override_internal): Check
1940 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
1941 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
1942 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
1943 MASK_AVX256_SPLIT_UNALIGNED_STORE.
1945 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
1946 Clear m_COREI7_AVX and update comments.
1947 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
1949 2013-11-12 Martin Jambor <mjambor@suse.cz>
1951 PR rtl-optimization/10474
1952 * ira.c (interesting_dest_for_shprep): New function.
1953 (split_live_ranges_for_shrink_wrap): Likewise.
1954 (find_moveable_pseudos): Move calculation of dominance info,
1955 df_analysios and the final anlyses to...
1956 (ira): ...here, call split_live_ranges_for_shrink_wrap.
1958 2013-11-12 Bin Cheng <bin.cheng@arm.com>
1960 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
1961 Handle type conversion.
1963 2013-11-11 Martin Liska <marxin.liska@gmail.com>
1964 Jan Hubicka <jh@suse.cz>
1966 * cgraph.c (dump_cgraph_node): Profile dump added.
1967 * cgraph.h (struct cgraph_node): New time profile variable added.
1968 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
1969 * gcov-io.h (gcov_type): New profiler type introduced.
1970 * ipa-profile.c (lto_output_node): Streaming for time profile added.
1971 (input_node): Time profiler is read from LTO stream.
1972 * predict.c (maybe_hot_count_p): Hot prediction changed.
1973 * profile.c (instrument_values): New case for time profiler added.
1974 (compute_value_histograms): Read of time profile.
1975 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
1976 * tree-profile.c (init_ic_make_global_vars): Time profiler
1978 (gimple_init_edge_profiler): TP function instrumentation.
1979 (gimple_gen_time_profiler): New.
1980 * value-prof.c (gimple_add_histogram_value): Support for time profiler
1982 (dump_histogram_value): TP type added to dumps.
1983 (visit_hist): More sensitive check that takes TP into account.
1984 (gimple_find_values_to_profile): TP instrumentation.
1985 * value-prof.h (hist_type): New histogram type added.
1986 (struct histogram_value_t): Pointer to struct function added.
1987 * libgcc/Makefile.in: New GCOV merge function for TP added.
1988 * libgcov.c: function_counter variable introduced.
1989 (_gcov_merge_time_profile): New.
1990 (_gcov_time_profiler): New.
1992 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
1993 Jeff Law <law@redhat.com>
1995 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
1996 ao_ref_init_from_ptr_and_size for builtins.
1998 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
1999 H.J. Lu <hongjiu.lu@intel.com>
2002 * config/i386/x86-tune.def
2003 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
2004 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
2005 * config/i386/i386.h
2006 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
2007 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
2008 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
2009 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
2010 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
2011 misaligned_prologue_used. Check that
2012 desired_aling <= epilogue_size_needed.
2014 2013-11-11 Cong Hou <congh@google.com>
2016 PR tree-optimization/59050
2017 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
2019 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
2022 * expmed.c (emit_store_flag): Fail for const-const comparison.
2024 2013-11-11 Tristan Gingold <gingold@adacore.com>
2025 Eric Botcazou <ebotcazou@adacore.com>
2027 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
2028 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
2029 * tree.def (CONSTRUCTOR): Likewise.
2030 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
2031 * gimplify.c (gimplify_init_constructor): Do not clear the object when
2032 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
2034 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
2036 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
2039 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2041 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
2044 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
2046 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
2048 2013-11-08 Jeff Law <law@redhat.com>
2050 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
2051 threading paths first, then perform PHI node checks if applicable.
2053 2013-11-10 Karlson2k <k2k@narod.ru>
2054 Kai Tietz <ktietz@redhat.com>
2057 * configure.ac: Adding for exported symbols check
2058 and for rdynamic-check executable-extension.
2059 * configure: Regenerated.
2061 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
2063 * mode-switching.c (optimize_mode_switching): Mark block as
2064 nontransparent, if last_mode at block exit is different from no_mode.
2066 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2068 * function.c (NAME__MAIN): Move to...
2069 * cfgexpand.c (NAME__MAIN): ...here.
2071 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
2073 * target.def (can_use_doloop_p): New hook.
2074 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
2075 * doc/tm.texi: Regenerate.
2076 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
2077 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
2078 * hooks.c (hook_bool_dint_dint_uint_true): New function.
2079 * targhooks.h (can_use_doloop_if_innermost): Declare.
2080 * targhooks.c (can_use_doloop_if_innermost): New function.
2081 * target.h: Include double-int.h.
2082 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
2083 Remove iteration count, maximum iteration count, loop depth and
2084 enter-at-top inputs from doloop_begin and doloop_end.
2085 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
2087 * config/arc/arc.c (arc_can_use_doloop_p): New function.
2088 (TARGET_CAN_USE_DOLOOP_P): Define.
2089 * config/arm/thumb2.md (doloop_end): Update for new interface.
2090 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
2091 * config/bfin/bfin.md (doloop_end): Update for new interface.
2092 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
2093 (TARGET_CAN_USE_DOLOOP_P): Define.
2094 * config/c6x/c6x.md (doloop_end): Update for new interface.
2095 * config/ia64/ia64.md (doloop_end): Update for new interface.
2096 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
2097 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
2099 * config/mep/mep.c (mep_emit_doloop): Likewise.
2100 (TARGET_CAN_USE_DOLOOP_P): Define.
2101 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
2102 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
2103 * config/s390/s390.md (doloop_end): Update for new interface.
2104 * config/sh/sh.md (doloop_end): Likewise.
2105 * config/spu/spu.md (doloop_end): Likewise.
2106 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
2107 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
2108 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
2109 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
2110 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
2111 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
2113 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
2115 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
2118 * doc/extend.texi: Move Cilk Plus Builtins node before Other
2121 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
2122 Joseph Myers <joseph@codesourcery.com>
2124 * ginclude/stdatomic.h: New file.
2125 * Makefile.in (USER_H): Add stdatomic.h.
2127 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2129 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
2132 2013-11-08 Jeff Law <law@redhat.com>
2134 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
2135 * tree-ssa-threadupdate.c (delete_thread_path): New function.
2136 (ssa_redirect_edges, thread_block_1): Use it.
2137 (thread_through_loop_header, mark_threaded_blocks): Likewise.
2138 (thread_through_all_blocks, register_jump_thread): Likewise.
2139 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
2141 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
2143 * config/arm/aarch-common.c
2144 (search_term): New typedef.
2145 (shift_rtx_costs): New array.
2146 (arm_rtx_shift_left_p): New.
2147 (arm_find_sub_rtx_with_search_term): Likewise.
2148 (arm_find_sub_rtx_with_code): Likewise.
2149 (arm_early_load_addr_dep): Add sanity checking.
2150 (arm_no_early_alu_shift_dep): Likewise.
2151 (arm_no_early_alu_shift_value_dep): Likewise.
2152 (arm_no_early_mul_dep): Likewise.
2153 (arm_no_early_store_addr_dep): Likewise.
2155 2013-11-08 Richard Biener <rguenther@suse.de>
2157 PR tree-optimization/59047
2158 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
2160 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
2162 * common.opt (fcheck-pointer-bounds): Move to ...
2163 * c-family/c.opt: ... here.
2164 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
2165 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
2166 * langhooks.h (lang_hooks): Remove chkp_supported field.
2167 * toplev.c (process_options): Remove chkp_supported check.
2169 2013-11-08 Richard Biener <rguenther@suse.de>
2171 PR tree-optimization/59038
2172 PR tree-optimization/58955
2173 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
2174 previous change. Handle known dependences correctly.
2176 2013-11-08 Tom de Vries <tom@codesourcery.com>
2178 * config/rs6000/t-xilinx: Remove duplicate contents.
2180 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
2181 Joseph Myers <joseph@codesourcery.com>
2183 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
2184 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
2185 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
2186 (struct tree_base): Add atomic_flag field.
2187 * tree.h (TYPE_ATOMIC): New accessor macro.
2188 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
2189 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
2190 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
2191 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
2192 * tree.c (set_type_quals): Set TYPE_ATOMIC.
2193 (find_atomic_core_type): New function.
2194 (build_qualified_type): Adjust alignment for qualified types.
2195 (build_atomic_base): New function
2196 (build_common_tree_nodes): Build atomicQI_type_node,
2197 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
2199 * print-tree.c (print_node): Print atomic qualifier.
2200 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
2201 * target.def (atomic_assign_expand_fenv): New hook.
2202 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
2203 * doc/tm.texi: Regenerate.
2204 * targhooks.c (default_atomic_assign_expand_fenv): New function.
2205 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
2206 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
2207 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
2209 * config/i386/i386.c (enum ix86_builtins): Add
2210 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
2211 IX86_BUILTIN_FNCLEX.
2212 (bdesc_special_args): Add __builtin_ia32_fnstenv,
2213 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
2214 (ix86_expand_builtin): Handle the new built-in functions.
2215 (ix86_atomic_assign_expand_fenv): New function.
2216 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
2217 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
2218 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
2219 (fnstenv, fldenv, fnstsw, fnclex): New insns.
2221 2013-11-07 Steve Ellcey <sellcey@mips.com>
2223 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
2224 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
2225 (MULTILIB_DIRNAMES): Add fp64 directory.
2226 (MULTILIB_EXCEPTIONS): Add new exclusions.
2228 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
2230 * gimplify.c (gimple_regimplify_operands): Do not set
2232 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
2233 (rewrite_close_phi_out_of_ssa): Same.
2234 (rewrite_phi_out_of_ssa): Same.
2235 (rewrite_degenerate_phi): Same.
2236 (handle_scalar_deps_crossing_scop_limits): Same.
2237 * tree-if-conv.c (predicate_scalar_phi): Same.
2238 * tree-parloops.c (create_loads_for_reductions): Same.
2239 (create_final_loads_for_reduction): Same.
2240 (create_loads_and_stores_for_name): Same.
2241 (transform_to_exit_first_loop): Same.
2242 (create_parallel_loop): Same.
2243 * tree-ssa-loop-im.c
2244 (move_computations_dom_walker::before_dom_children): Same.
2245 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
2246 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
2247 * tree-ssa-propagate.c (substitute_and_fold): Same.
2248 * tree-vect-loop.c (vect_finalize_reduction): Same.
2249 * tree-vect-stmts.c (vectorizable_call): Same.
2251 2013-11-07 Mike Stump <mikestump@comcast.net>
2253 * config/pdp11/pdp11.c: Include dbxout.h.
2254 * config/picochip/picochip.c: Likewise.
2256 2013-11-07 Cong Hou <congh@google.com>
2258 PR tree-optimization/56764
2259 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
2260 Combine alias checks if it is possible to amortize the runtime
2261 overhead. Return the number of alias checks after merging.
2262 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
2263 Use the function vect_create_cond_for_alias_checks () to check
2264 the number of alias checks.
2266 2013-11-07 Jeff Law <law@redhat.com>
2268 * varpool.c (ctor_for_folding): Fix typo in comment.
2270 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
2272 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
2273 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
2275 2013-11-07 Jeff Law <law@redhat.com>
2277 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
2279 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
2280 No longer check if we have __builtin_trap, assume it's available.
2282 2013-11-07 Diego Novillo <dnovillo@google.com>
2284 * attribs.c (lookup_scoped_attribute_spec): Make static.
2285 (get_attribute_namespace): Likewise.
2286 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
2287 (validate_arglist): Move earlier in the file. Make static.
2288 (expand_stack_restore): Move from stmt.c
2289 (expand_stack_save): Move from stmt.c
2290 (rewrite_call_expr_array): Move earlier in the file.
2291 (rewrite_call_expr_valist): Likewise.
2292 * cfgexpand.c: Include hard-reg-set.h before tree.h
2295 (expand_asm_loc): Move from stmt.c.
2296 (n_occurrences): Move from stmt.c.
2297 (check_operand_nalternatives): Move from stmt.c.
2298 (tree_conflicts_with_clobbers_p): Move from stmt.c.
2299 (expand_asm_operands): Move from stmt.c
2300 (expand_asm_stmt): Move from stmt.c
2301 (expand_computed_goto): Move from stmt.c
2302 (expand_goto): Move from stmt.c
2303 (expand_null_return_1): Move from stmt.c
2304 (expand_null_return): Move from stmt.c
2305 (expand_value_return): Move from stmt.c
2306 (expand_return): Move from stmt.c
2307 (expand_main_function): Move from function.c
2308 (stack_protect_prologue): Move from function.c
2309 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
2310 (build_function_decl_skip_args): Move from tree.c.
2311 * explow.c (tree_expr_size): Move from tree.c.
2312 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
2313 (fields_length): Move from tree.c.
2314 * fold-const.c (size_low_cst): Move from tree.c.
2315 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
2316 (tree_expr_nonzero_p): Make static. Move earlier in the file.
2317 (fold_build3_initializer_loc): Remove.
2318 (tree_invalid_nonnegative_warnv_p): Make static.
2319 * function.c (expand_main_function): Move to cfgexpand.c.
2320 (stack_protect_prologue): Move to cfgexpand.c.
2321 (set_insn_locations): Move earlier in the file.
2322 * gimple-fold.c: Include langhooks.h.
2323 (truth_type_for): Move from tree.c.
2324 * print-tree.c (print_vec_tree): Remove.
2325 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
2326 (expand_goto): Move to cfgexpand.c.
2327 (n_occurrences): Move to cfgexpand.c.
2328 (expand_asm_loc): Move to cfgexpand.c
2329 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
2330 (expand_asm_operands): Move to cfgexpand.c.
2331 (expand_asm_stmt): Move to cfgexpand.c.
2332 (check_operand_nalternatives): Move to cfgexpand.c
2333 (expand_null_return): Move to cfgexpand.c.
2334 (expand_value_return): Move to cfgexpand.c.
2335 (expand_null_return_1): Move to cfgexpand.c.
2336 (expand_return): Move to cfgexpand.c.
2337 (expand_stack_save): Move to builtins.c.
2338 (expand_stack_restore): Move to builtins.c
2339 * symtab.c: Include output.h.
2340 (decl_assembler_name_hash): Move from tree.c.
2341 (decl_assembler_name_equal): Move from tree.c.
2342 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
2343 * tree-eh.c (in_array_bounds_p): Move from tree.c.
2344 (range_in_array_bounds_p): Move from tree.c.
2345 * tree-object-size.c (fini_object_sizes): Make static.
2346 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
2347 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
2348 * tree.c (decl_assembler_name_equal): Move to symtab.c.
2349 (tree_expr_size): Move to explow.c.
2350 (decl_assembler_name_hash): Move to symtab.c.
2351 (real_twop): Remove.
2352 (tree_expr_size): Move to explow.c.
2353 (stabilize_reference_1): Move earlier in the file. Make static.
2354 (omp_remove_redundant_declare_simd_attrs): Remove.
2355 (simple_cst_list_equal): Move earlier in the file. Make static.
2356 (size_low_cst): Move to fold-const.c.
2357 (build_type_no_quals): Remove.
2358 (build_function_type_skip_args): Move to cgraphclones.c.
2359 (build_function_decl_skip_args): Move to cgraphclones.c.
2360 (in_array_bounds_p): Move to tree-eh.c.
2361 (range_in_array_bounds_p): Move to tree-eh.c.
2362 (truth_type_for): Move to gimple-fold.c.
2363 (list_equal_p): Remove.
2364 * tree.h (decl_assembler_name_equal): Remove.
2365 (decl_assembler_name_hash): Remove.
2366 (truth_type_for): Remove.
2367 (build_type_no_quals): Remove.
2368 (build_function_decl_skip_args): Remove.
2369 (in_array_bounds_p): Remove.
2370 (range_in_array_bounds_p): Remove.
2371 (size_low_cst): Remove.
2372 (omp_remove_redundant_declare_simd_attrs): Remove.
2373 (tree_expr_size): Remove.
2374 (fields_length): Remove.
2375 (stabilize_reference_1): Remove.
2376 (expand_goto): Remove.
2377 (expand_stack_save): Remove.
2378 (expand_stack_restore): Remove.
2379 (expand_return): Remove.
2380 (fold_build3_initializer_loc): Remove.
2381 (tree_expr_nonzero_p): Remove.
2382 (tree_invalid_nonnegative_warnv_p): Remove.
2383 (tree_expr_nonzero_warnv_p): Remove.
2384 (fold_builtin_snprintf_chk): Remove.
2385 (validate_arglist): Remove.
2386 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
2387 (simple_cst_list_equal): Remove.
2388 (real_twop): Remove.
2389 (expand_main_function): Remove.
2390 (stack_protect_prologue): Remove.
2391 (print_vec_tree): Remove.
2392 (lookup_scoped_attribute_spec): Remove.
2393 (get_attribute_namespace): Remove.
2394 (expand_computed_goto): Remove.
2395 (expand_asm_stmt): Remove.
2396 (list_equal_p): Remove.
2397 (ssa_name_nonnegative_p): Remove.
2398 (fini_object_sizes): Remove.
2399 (addr_expr_of_non_mem_decl_p): Remove.
2400 (is_tm_safe_or_pure): Move to trans-mem.c.
2401 (more_const_call_expr_args_p): Remove.
2402 (save_vtable_map_decl): Remove.
2404 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
2406 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
2407 linker plugins, too.
2409 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
2410 link to libgomp and its dependencies.
2411 * config/ia64/hpux.h (LIB_SPEC): Likewise.
2412 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
2413 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
2414 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
2416 2013-11-07 Jakub Jelinek <jakub@redhat.com>
2418 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
2419 (determine_value_range): Add loop argument. Use get_range_info to
2421 (bound_difference): Adjust caller.
2423 2013-11-07 Richard Biener <rguenther@suse.de>
2424 Jakub Jelinek <jakub@redhat.com>
2426 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
2427 latches from header PHI arguments from the latch edge.
2429 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
2432 * varasm.c (output_constant): Handle NULLPTR_TYPE.
2434 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
2436 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
2437 misaligned_prologue_used when it has been set.
2439 2013-11-07 Yury Gribov <y.gribov@samsung.com>
2440 Jakub Jelinek <jakub@redhat.com>
2443 * asan.c (get_mem_refs_of_builtin_call): Allow
2444 integer literals as addresses in instrumented builtins.
2446 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2448 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
2449 Explain why plus_constant is not used.
2451 2013-11-07 Richard Biener <rguenther@suse.de>
2453 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
2454 (canonicalize_value): ... this. Also handle stripping of
2456 (get_value, set_lattice_value, get_value_for_expr): Adjust.
2457 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
2458 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
2460 2013-11-07 Richard Biener <rguenther@suse.de>
2462 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
2464 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
2467 * config/i386/i386.md (push peepholer/splitter): Use Pmode
2468 with stack_pointer_rtx.
2470 2013-11-07 Bin Cheng <bin.cheng@arm.com>
2472 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
2473 using operand_equal_p.
2475 2013-11-07 Bin Cheng <bin.cheng@arm.com>
2477 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
2478 * tree-affine.c (get_inner_reference_aff): Return base.
2479 * tree-affine.h (get_inner_reference_aff): Change prototype.
2481 2013-11-06 Tobias Burnus <burnus@net-b.de>
2483 * doc/invoke.texi (Wdate-time): Fix typo.
2485 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
2487 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
2488 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
2489 fp_arith_reg_operand instead of arith_reg_operand.
2491 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
2493 * config/sh/sh.md (adddi3): Remove empty constraints.
2494 Remove can_create_pseudo_p and arith_reg_operand check.
2495 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
2496 Split before reload.
2498 2013-11-06 Jeff Law <law@redhat.com>
2499 Tom Tromey <tromey@redhat.com>
2501 * gdbinit.in: Disable strict type checking.
2503 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
2505 * tree-pass.h (make_pass_live_range_shrinkage): New external.
2506 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
2507 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
2508 (rest_of_handle_live_range_shrinkage): Ditto
2509 (class pass_live_range_shrinkage): Ditto.
2510 (pass_data_live_range_shrinkage): Ditto.
2511 (make_pass_live_range_shrinkage): Ditto.
2512 * sched-int.h (initialize_live_range_shrinkage): New prototype.
2513 (finish_live_range_shrinkage): Ditto.
2514 * sched-deps.c (create_insn_reg_set): Make void return value.
2515 * passes.def: Add pass_live_range_shrinkage.
2516 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
2517 * haifa-sched.c (live_range_shrinkage_p): New.
2518 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
2520 (rank_for_schedule): Add code for pressure relief through live
2522 (schedule_insn): Print more debug info.
2523 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
2524 through live range shrinkage.
2525 * doc/invoke.texi (-flive-range-shrinkage): New.
2526 * common.opt (flive-range-shrinkage): New.
2528 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
2531 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
2532 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
2533 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
2534 RTXes that return in AVX256 register.
2536 2013-11-06 Richard Biener <rguenther@suse.de>
2538 PR tree-optimization/58653
2539 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
2540 (prepare_initializers_chain): Adjust.
2542 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
2544 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
2545 (gimple_in_transaction): Use bb_in_transaction.
2546 * basic-block.h (bb_in_transaction): Relocate here and rename.
2547 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
2549 2013-11-06 Richard Biener <rguenther@suse.de>
2551 * tree.c (drop_tree_overflow): New function.
2552 * tree.h (drop_tree_overflow): Declare.
2553 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
2554 * tree-vrp.c (range_int_cst_singleton_p): Use
2555 is_overflow_infinity instead of testing TREE_OVERFLOW.
2556 (extract_range_from_assert): Likewise.
2557 (zero_nonzero_bits_from_vr): Likewise.
2558 (extract_range_basic): Likewise.
2559 (register_new_assert_for): Use drop_tree_overflow.
2560 (vrp_visit_phi_node): Likewise.
2562 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
2564 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
2565 checking for leaf functions without dynamic stack allocation.
2566 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
2567 (ia64_expand_prologue): Likewise.
2568 * config/mips/mips.c (mips_expand_prologue): Likewise.
2569 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
2570 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
2571 (sparc_flat_expand_prologue): Likewise.
2573 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
2575 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
2576 (__ST3_LANE_FUNC): Likewise.
2577 (__ST4_LANE_FUNC): Likewise.
2579 2013-11-06 Nick Clifton <nickc@redhat.com>
2581 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
2582 name returned by msp430_mcu_name.
2583 (LIB_SPEC): If a -T option has not been specified then set a
2584 default, mcu-specific, linker script.
2585 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
2586 * config/msp430/msp430.c (msp430x_names): Likewise.
2587 Alpha sort the names for ease of comparison.
2588 (msp430_mcu_name): New function: Returns a string suitable for
2589 use as a C preprocessor symbol based upon the name of the MCU
2591 (msp430_option_override): Accept msp430x and msp430xv2 as generic
2593 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
2595 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
2596 unless it is the last part of a spec substring.
2597 * doc/invoke.texi (Spec Files): Document space insertion
2600 2013-11-06 Christian Bruel <christian.bruel@st.com>
2602 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
2603 Factorize probabilities, Use adjust_address instead of
2604 adjust_automodify_address when possible. Enable for optimize.
2605 (sh_expand_strlen): New function.
2606 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
2607 * config/sh/sh.md (strlensi): New pattern.
2608 (UNSPEC_BUILTIN_STRLEN): Define.
2610 2013-11-06 Jakub Jelinek <jakub@redhat.com>
2613 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
2614 (expand_assignment): If *bitpos is negative, set *offset
2615 and adjust *bitpos, so that it is not negative.
2617 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
2619 * config/i386/bdver3.md : Added two additional decoder units
2620 to support issue rate of 4 and remodeled vector unit.
2621 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
2622 architectures is set to 4.
2623 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
2624 lookahead is set to 4 for BD architectures.
2626 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2628 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2629 Remove restriction against use of VSX instructions when generating
2630 code for little endian mode.
2632 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2634 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
2635 for both big and little endian.
2636 (mulv8hi3): Swap input operands for merge high and merge low
2637 instructions for little endian.
2639 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2641 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
2642 define_insn to define_expand that uses even patterns for big
2643 endian and odd patterns for little endian.
2644 (vec_widen_smult_even_v16qi): Likewise.
2645 (vec_widen_umult_even_v8hi): Likewise.
2646 (vec_widen_smult_even_v8hi): Likewise.
2647 (vec_widen_umult_odd_v16qi): Likewise.
2648 (vec_widen_smult_odd_v16qi): Likewise.
2649 (vec_widen_umult_odd_v8hi): Likewise.
2650 (vec_widen_smult_odd_v8hi): Likewise.
2651 (altivec_vmuleub): New define_insn.
2652 (altivec_vmuloub): Likewise.
2653 (altivec_vmulesb): Likewise.
2654 (altivec_vmulosb): Likewise.
2655 (altivec_vmuleuh): Likewise.
2656 (altivec_vmulouh): Likewise.
2657 (altivec_vmulesh): Likewise.
2658 (altivec_vmulosh): Likewise.
2660 2013-11-05 Mike Stump <mikestump@comcast.net>
2662 * Makefile.in (mostlyclean): Remove c-family objects.
2664 2013-11-05 Ian Lance Taylor <iant@google.com>
2666 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
2667 If possible, add .cfi directives to record change to bx.
2668 * config/i386/i386.c (ix86_emit_cfi): New function.
2669 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
2671 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
2673 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
2674 immediately after a label for a tablejump pattern.
2676 * config/arm/arm.c (is_jump_table): Remove.
2677 (create_fix_barrier): Use tablejump_p instead.
2678 (arm_reorg): Likewise.
2679 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
2680 (thumb2_output_casesi): Likewise.
2681 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
2682 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
2683 casesi_shift_media, casesi_load_media): Likewise.
2684 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
2685 * config/microblaze/microblaze.md: Likewise.
2687 2013-11-05 Tobias Burnus <burnus@net-b.de>
2689 * doc/invoke.texi (-Wdate-time): Document.
2691 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
2693 * double-int.c (lshift_double, rshift_double): Remove
2694 SHIFT_COUNT_TRUNCATED handling.
2696 2013-11-05 Jeff Law <law@redhat.com>
2698 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
2699 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
2700 * gimple-ssa-isolate-paths.c: New file.
2701 * gimple.c (check_loadstore): New function.
2702 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
2703 Verify OP is in the argument list and the argument corresponding
2704 to OP is a pointer type. Use operand_equal_p rather than
2705 pointer equality when testing if OP is on the nonnull list.
2706 Use check_loadstore rather than count_ptr_derefs. Handle
2707 GIMPLE_RETURN statements.
2708 * tree-vrp.c (infer_nonnull_range): Remove.
2709 * gimple.h (infer_nonnull_range): Declare.
2710 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
2711 * passes.def: Add pass_isolate_erroneous_paths.
2712 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
2713 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
2714 * tree-ssa.c (struct count_ptr_d): Remove.
2715 (count_ptr_derefs, count_uses_and_derefs): Remove.
2716 * tree-ssa.h (count_uses_and_derefs): Remove.
2718 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2720 PR rtl-optimization/58997
2721 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
2722 get_iv_value to be in iv->mode rather than iv->extend_mode.
2723 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
2724 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
2725 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
2728 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2730 * gimple.h: Move some prototypes to gimple-expr.h and add to include
2732 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
2733 Move to gimple-expr.h.
2734 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
2735 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
2736 is_gimple_id, virtual_operand_p, is_gimple_addressable,
2737 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
2739 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
2740 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
2741 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
2742 is_gimple_address, is_gimple_invariant_address,
2743 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2744 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
2745 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
2746 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
2747 gimple_decl_printable_name, useless_type_conversion_p,
2748 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
2750 * gimple-expr.c: New File.
2751 (useless_type_conversion_p, gimple_set_body, gimple_body,
2752 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
2753 gimple_can_coalesce_p, extract_ops_from_tree_1,
2754 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
2755 is_gimple_address, is_gimple_invariant_address,
2756 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2757 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
2758 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
2759 is_gimple_mem_ref_addr): Relocate here.
2760 * Makefile.in (OBJS): Add gimple-expr.o.
2762 2013-11-05 David Malcolm <dmalcolm@redhat.com>
2764 * gengtype-parse.c (struct_field_seq): Support empty structs.
2766 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2768 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
2770 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2772 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
2774 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
2776 (LONG_DOUBLE_TYPE_SIZE): New define.
2777 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
2779 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
2782 * doc/extend.texi [visibility ("visibility_type")]: Add example
2783 about visibility attribute on namespace declaration.
2785 2013-11-05 Richard Biener <rguenther@suse.de>
2788 * passes.def (all_passes): Start with pass_fixup_cfg again.
2790 2013-11-05 Richard Biener <rguenther@suse.de>
2792 PR tree-optimization/58955
2793 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
2796 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2798 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
2800 (vec_pack_ufix_trunc_v2df): Likewise.
2802 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
2805 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
2806 pattern, instead of word_mode.
2808 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
2810 (set_storage_via_setmem): Likewise.
2812 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2814 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
2815 gimple.h and the rest of the condition in eliminate_build.
2816 (eliminate_build): Call new routine.
2817 * gimple.h (phi_ssa_name_p): Delete.
2819 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
2821 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
2822 case of no prefix and reserving zero slots, because when that's the
2823 case we'll never get here.
2824 * vec.h (va_heap::reserve): Don't try and handle
2825 vec_prefix::calculate_allocation returning zero because that should
2828 2013-11-05 Richard Biener <rguenther@suse.de>
2831 * tree-dfa.c (get_ref_base_and_extent): Merge common code
2832 in MEM_REF and TARGET_MEM_REF handling. Make sure to
2833 process trailing array detection before diving into the
2834 view-converted object (and possibly apply some extra offset).
2836 2013-11-05 Joseph Myers <joseph@codesourcery.com>
2838 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
2840 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
2842 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2844 PR tree-optimization/58958
2845 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
2846 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
2848 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2850 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
2851 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
2853 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2855 PR tree-optimization/58984
2856 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
2857 set *SIZE_P if non-NULL on success.
2858 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
2860 (ipcp_transform_function): Likewise. Punt if size of access
2861 is different from TYPE_SIZE on v->value's type.
2863 2013-11-05 Tobias Burnus <burnus@net-b.de>
2865 * doc/invoke.texi (-fopenmp-simd): Document new option.
2866 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
2867 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
2868 * tree.c (attribute_value_equal): Ditto.
2870 2013-11-04 Wei Mi <wmi@google.com>
2872 * sched-rgn.c (add_branch_dependences): Keep insns in
2873 a SCHED_GROUP at the end of BB to remain their location.
2875 2013-11-04 Wei Mi <wmi@google.com>
2877 * config/i386/i386.c (memory_address_length): Extract a part
2878 of code to rip_relative_addr_p.
2879 (rip_relative_addr_p): New Function.
2880 (ix86_macro_fusion_p): Ditto.
2881 (ix86_macro_fusion_pair_p): Ditto.
2882 * config/i386/i386.h: Add new tune features about macro-fusion.
2883 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2884 * doc/tm.texi: Generated.
2885 * doc/tm.texi.in: Ditto.
2886 * haifa-sched.c (try_group_insn): New Function.
2887 (group_insns_for_macro_fusion): Ditto.
2888 (sched_init): Call group_insns_for_macro_fusion.
2889 * target.def: Add two hooks: macro_fusion_p and
2890 macro_fusion_pair_p.
2892 2013-11-04 Kostya Serebryany <kcc@google.com>
2894 Update to match the changed asan API.
2895 * asan.c (asan_function_start): New function.
2896 (asan_emit_stack_protection): Update the string stored in the
2897 stack red zone to match new API. Store the PC of the current
2898 function in the red zone.
2899 (asan_global_struct): Update the __asan_global definition to match
2901 (asan_add_global): Ditto.
2902 * asan.h (asan_function_start): New prototype.
2903 * final.c (final_start_function): Call asan_function_start.
2904 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
2907 2013-11-04 Wei Mi <wmi@google.com>
2909 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
2910 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
2911 * config/i386/i386.c (ix86_option_override_internal): Ditto.
2912 (ix86_issue_rate): Ditto.
2913 (ix86_adjust_cost): Ditto.
2914 (ia32_multipass_dfa_lookahead): Ditto.
2915 (ix86_sched_init_global): Ditto.
2916 (get_builtin_code_for_version): Ditto.
2917 * config/i386/i386.h (enum target_cpu_default): Ditto.
2918 (enum processor_type): Ditto.
2919 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2921 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2923 PR rtl-optimization/58967
2924 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
2925 !lra_in_progress for mode sizes bigger word.
2927 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2929 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
2930 arguments to merge instruction for little endian.
2931 (vec_widen_umult_lo_v16qi): Likewise.
2932 (vec_widen_smult_hi_v16qi): Likewise.
2933 (vec_widen_smult_lo_v16qi): Likewise.
2934 (vec_widen_umult_hi_v8hi): Likewise.
2935 (vec_widen_umult_lo_v8hi): Likewise.
2936 (vec_widen_smult_hi_v8hi): Likewise.
2937 (vec_widen_smult_lo_v8hi): Likewise.
2939 2013-11-04 Ian Lance Taylor <iant@google.com>
2941 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
2942 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
2943 builtins that take pointers.
2944 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
2945 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
2946 (run_gcc): Pass -fnon-call-exceptions.
2948 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2950 * optabs.c (expand_vec_perm): Revert one incorrect line from
2953 PR tree-optimization/58978
2954 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
2955 use_stmt by single_imm_use directly. Only call single_imm_use
2958 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2960 PR rtl-optimization/58968
2961 * lra-spills.c (return_regno_p): New function.
2962 (lra_final_code_change): Use it.
2964 2013-11-04 Joseph Myers <joseph@codesourcery.com>
2966 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
2967 * target.def (float_exceptions_rounding_supported_p): New hook.
2968 * targhooks.c (default_float_exceptions_rounding_supported_p): New
2970 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
2971 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
2973 * doc/tm.texi: Regenerate.
2974 * config.gcc (powerpc*-*-linux*): Set extra_objs.
2975 * config/rs6000/rs6000-linux.c: New file.
2976 * config/rs6000/rs6000-protos.h
2977 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
2978 * config/rs6000/linux.h
2979 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
2980 * config/rs6000/linux64.h
2981 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
2982 * config/rs6000/t-linux (rs6000-linux.o): New rule.
2983 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
2985 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2987 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
2988 Replace the define_insn_and_split with a define_insn and two
2989 define_splits, with the split after reload re-permuting the source
2990 register to its original value.
2991 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
2992 (*vsx_le_perm_store_v8hi): Likewise.
2993 (*vsx_le_perm_store_v16qi): Likewise.
2995 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2997 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
3000 2013-11-04 Jakub Jelinek <jakub@redhat.com>
3002 PR tree-optimization/58946
3003 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
3004 bbs with bbinfo[idx].op != NULL before all blocks with
3005 bbinfo[idx].op == NULL.
3007 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
3009 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
3010 (avr_dump_double_int_hex): Likewise.
3011 (avr_log_vadump): Remove %D and %X handling.
3012 * config/avr/avr.c (avr_double_int_push_digit): Delete.
3013 (avr_map_op_t): Change map from double_int to unsigned int.
3014 (avr_map_op): Update accordingly.
3015 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
3016 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
3017 unsigned ints rather than double_ints.
3019 2013-11-03 Marek Polacek <polacek@redhat.com>
3021 Implement -fsanitize=vla-bound.
3022 * opts.c (common_handle_option): Handle vla-bound.
3023 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
3024 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
3025 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
3027 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3029 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
3032 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
3034 * config/i386/constraints.md (Ts, Tv): New address constrains.
3035 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
3036 constraint for address_no_seg_operand.
3037 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
3038 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
3039 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
3040 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
3041 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
3042 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
3043 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
3044 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
3045 constraint for vsib_address_operand.
3047 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
3049 * gcse.c (pre_delete): Remove references to regmove from comments.
3050 * recog.c: (validate_replace_rtx_1): Likewise.
3051 * config/rl78/rl78.c: Likewise.
3052 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
3053 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
3054 * common/config/mmix/mmix-common.c: Likewise.
3056 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
3058 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
3059 * gimplify.c (gimplify_compound_lval): Likewise.
3060 * graphite-clast-to-gimple.c (gloog): Likewise.
3061 * graphite-dependences.c (loop_is_parallel_p): Likewise.
3062 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
3063 (limit_scop); Likewise.
3064 (build_scops): Likewise.
3065 (dot_scop): Likewise.
3066 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3067 (build_scop_drs): Likewise.
3068 (insert_stmts): Likewise.
3069 (insert_out_of_ssa_copy): Likewise.
3070 (remove_phi): Likewise.
3071 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
3072 * hw-doloop.c (discover_loop): Likewise.
3073 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
3074 * tree-dfa.c (dump_enumerated_decls): Likewise.
3075 * tree-if-conv.c (if_convertable_loop_p): Likewise.
3076 * tree-inline.c (tree_function_versioning): Likewise.
3077 * tree-loop-distribution.c (build_rdg): Likewise.
3078 (rdg_flag_vertex_and_dependent): Likewise.
3079 (distribute_loop): Likewise.
3080 * tree-parloops.c (loop_parallel_p): Likewise.
3081 (eliminate_local_variables): Likewise.
3082 (separate_decls_in_region): Likewise.
3083 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
3084 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
3085 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3086 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
3087 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
3088 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
3089 (vectorizable_condition): Likewise.
3091 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
3093 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
3094 * configure: Regenerate.
3095 * config/i386/i386.md (*movdi_internal): Change
3096 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
3097 (*movdf_internal): Ditto.
3098 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
3099 * config/i386/sse.md (vec_concatv2di): Output interunit movq
3100 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
3102 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
3104 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
3107 2013-10-31 David Malcolm <dmalcolm@redhat.com>
3109 Automated part of renaming of symtab_node_base to symtab_node.
3111 Patch autogenerated by rename_symtab.py from
3112 https://github.com/davidmalcolm/gcc-refactoring-scripts
3113 revision 58bb219cc090b2f4516a9297d868c245495ee622
3114 with ChangeLog entry fixed up by hand.
3116 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
3118 (cgraph_node_for_asm): Likewise.
3119 * cgraph.h (symtab_node_base): Likewise.
3120 (cgraph_node): Likewise.
3121 (varpool_node): Likewise.
3122 (is_a_helper <cgraph_node>::test): Likewise.
3123 (is_a_helper <varpool_node>::test): Likewise.
3124 (symtab_nodes): Likewise.
3125 (symtab_register_node): Likewise.
3126 (symtab_unregister_node): Likewise.
3127 (symtab_remove_node): Likewise.
3128 (symtab_get_node): Likewise.
3129 (symtab_node_for_asm): Likewise.
3130 (symtab_node_asm_name): Likewise.
3131 (symtab_node_name): Likewise.
3132 (symtab_insert_node_to_hashtable): Likewise.
3133 (symtab_add_to_same_comdat_group): Likewise.
3134 (symtab_dissolve_same_comdat_group_list): Likewise.
3135 (dump_symtab_node): Likewise.
3136 (debug_symtab_node): Likewise.
3137 (dump_symtab_base): Likewise.
3138 (verify_symtab_node): Likewise.
3139 (verify_symtab_base): Likewise.
3140 (symtab_used_from_object_file_p): Likewise.
3141 (symtab_alias_ultimate_target): Likewise.
3142 (symtab_resolve_alias): Likewise.
3143 (fixup_same_cpp_alias_visibility): Likewise.
3144 (symtab_for_node_and_aliases): Likewise.
3145 (symtab_nonoverwritable_alias): Likewise.
3146 (availability symtab_node_availability): Likewise.
3147 (symtab_semantically_equivalent_p): Likewise.
3148 (fixup_same_cpp_alias_visibility): Likewise.
3149 (symtab_prevail_in_asm_name_hash): Likewise.
3151 (varpool): Likewise.
3152 (varpool_first_variable): Likewise.
3153 (varpool_next_variable): Likewise.
3154 (varpool_first_static_initializer): Likewise.
3155 (varpool_next_static_initializer): Likewise.
3156 (varpool_first_defined_variable): Likewise.
3157 (varpool_next_defined_variable): Likewise.
3158 (cgraph_first_defined_function): Likewise.
3159 (cgraph_next_defined_function): Likewise.
3160 (cgraph_first_function): Likewise.
3161 (cgraph_next_function): Likewise.
3162 (cgraph_first_function_with_gimple_body): Likewise.
3163 (cgraph_next_function_with_gimple_body): Likewise.
3164 (symtab_alias_target): Likewise.
3165 (symtab_real_symbol_p): Likewise.
3166 (symtab_can_be_discarded): Likewise.
3167 * cgraphbuild.c (mark_address): Likewise.
3168 (mark_load): Likewise.
3169 (mark_store): Likewise.
3170 * cgraphunit.c (decide_is_symbol_needed): Likewise.
3172 (enqueue_node): Likewise.
3173 (referred_to_p): Likewise.
3174 (cgraph_process_same_body_aliases): Likewise.
3175 (analyze_functions): Likewise.
3176 (handle_alias_pairs): Likewise.
3177 (output_weakrefs): Likewise.
3178 (compile): Likewise.
3179 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3180 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3181 * ipa-prop.c (remove_described_reference): Likewise.
3182 (try_decrement_rdesc_refcount): Likewise.
3183 (ipa_edge_duplication_hook): Likewise.
3184 * ipa-ref.c (ipa_record_reference): Likewise.
3185 (ipa_maybe_record_reference): Likewise.
3186 (ipa_clone_ref): Likewise.
3187 (ipa_clone_references): Likewise.
3188 (ipa_clone_referring): Likewise.
3189 (ipa_find_reference): Likewise.
3190 (ipa_remove_stmt_references): Likewise.
3191 (ipa_clear_stmts_in_references): Likewise.
3192 * ipa-ref.h (symtab_node_base): Likewise.
3193 (ipa_ref): Likewise.
3194 (ipa_record_reference): Likewise.
3195 (ipa_maybe_record_reference): Likewise.
3196 (ipa_clone_references): Likewise.
3197 (ipa_clone_referring): Likewise.
3198 (ipa_clone_ref): Likewise.
3199 (ipa_find_reference): Likewise.
3200 (ipa_remove_stmt_references): Likewise.
3201 (ipa_clear_stmts_in_references): Likewise.
3202 * ipa-reference.c (ipa_reference_write_optimization_summary):
3204 * ipa.c (enqueue_node): Likewise.
3205 (process_references): Likewise.
3206 (walk_polymorphic_call_targets): Likewise.
3207 (symtab_remove_unreachable_nodes): Likewise.
3208 (address_taken_from_non_vtable_p): Likewise.
3209 (comdat_can_be_unshared_p_1): Likewise.
3210 (comdat_can_be_unshared_p): Likewise.
3211 (can_replace_by_local_alias): Likewise.
3212 (function_and_variable_visibility): Likewise.
3213 * is-a.h: Likewise (within example in comment).
3214 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
3215 (lto_symtab_encoder_encode): Likewise.
3216 (lto_symtab_encoder_delete_node): Likewise.
3217 (lto_symtab_encoder_in_partition_p): Likewise.
3218 (lto_set_symtab_encoder_in_partition): Likewise.
3219 (output_refs): Likewise.
3220 (compute_ltrans_boundary): Likewise.
3221 (output_symtab): Likewise.
3222 (input_node): Likewise.
3223 (input_ref): Likewise.
3224 (input_edge): Likewise.
3225 (input_cgraph_1): Likewise.
3226 (input_refs): Likewise.
3227 (output_cgraph_opt_summary): Likewise.
3228 (input_node_opt_summary): Likewise.
3229 (input_cgraph_opt_section): Likewise.
3230 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
3232 * lto-streamer-out.c (lto_output): Likewise.
3233 (output_symbol_p): Likewise.
3234 (produce_symtab): Likewise.
3235 * lto-streamer.h (lto_encoder_entry): Likewise.
3236 (lto_free_function_in_decl_state_for_node): Likewise.
3237 (lto_symtab_encoder_encode): Likewise.
3238 (lto_symtab_encoder_delete_node): Likewise.
3239 (lto_symtab_encoder_in_partition_p): Likewise.
3240 (lto_set_symtab_encoder_in_partition): Likewise.
3241 (lto_symtab_encoder_lookup): Likewise.
3242 (lsei_node): Likewise.
3243 (lto_symtab_encoder_deref): Likewise.
3244 * symtab.c (symtab_hash): Likewise.
3245 (assembler_name_hash): Likewise.
3246 (symtab_nodes): Likewise.
3247 (hash_node): Likewise.
3248 (eq_node): Likewise.
3249 (hash_node_by_assembler_name): Likewise.
3250 (eq_assembler_name): Likewise.
3251 (insert_to_assembler_name_hash): Likewise.
3252 (unlink_from_assembler_name_hash): Likewise.
3253 (symtab_prevail_in_asm_name_hash): Likewise.
3254 (symtab_register_node): Likewise.
3255 (symtab_insert_node_to_hashtable): Likewise.
3256 (symtab_unregister_node): Likewise.
3257 (symtab_get_node): Likewise.
3258 (symtab_remove_node): Likewise.
3259 (symtab_initialize_asm_name_hash): Likewise.
3260 (symtab_node_for_asm): Likewise.
3261 (symtab_add_to_same_comdat_group): Likewise.
3262 (symtab_dissolve_same_comdat_group_list): Likewise.
3263 (symtab_node_asm_name): Likewise.
3264 (symtab_node_name): Likewise.
3265 (dump_symtab_base): Likewise.
3266 (dump_symtab_node): Likewise.
3267 (dump_symtab): Likewise.
3268 (debug_symtab_node): Likewise.
3269 (verify_symtab_base): Likewise.
3270 (verify_symtab_node): Likewise.
3271 (verify_symtab): Likewise.
3272 (symtab_used_from_object_file_p): Likewise.
3273 (symtab_node_availability): Likewise.
3274 (symtab_alias_ultimate_target): Likewise.
3275 (fixup_same_cpp_alias_visibility): Likewise.
3276 (symtab_resolve_alias): Likewise.
3277 (symtab_for_node_and_aliases): Likewise.
3278 (symtab_for_node_and_aliases): Likewise.
3279 (symtab_nonoverwritable_alias_1): Likewise.
3280 (symtab_nonoverwritable_alias): Likewise.
3281 (symtab_semantically_equivalent_p): Likewise.
3282 * value-prof.c (init_node_map): Likewise.
3283 * varasm.c (find_decl): Likewise.
3284 * varpool.c (varpool_node_for_asm): Likewise.
3285 (varpool_remove_unreferenced_decls): Likewise.
3287 2013-10-31 David Malcolm <dmalcolm@redhat.com>
3289 Manual part of renaming of symtab_node_base to symtab_node.
3291 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
3292 clashes with the preferred name for the base class.
3293 (const_symtab_node): Remove redundant typedef.
3295 2013-10-31 Jakub Jelinek <jakub@redhat.com>
3297 * optabs.c (expand_vec_perm): Avoid vector mode punning
3298 SUBREGs in SET_DEST.
3299 * expmed.c (store_bit_field_1): Likewise.
3300 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
3301 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
3302 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
3303 * config/i386/i386.c (ix86_expand_vector_move_misalign,
3304 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
3305 ix86_expand_sse_unpack, ix86_expand_args_builtin,
3306 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
3307 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
3308 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
3309 expand_vec_perm_vpshufb2_vpermq,
3310 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
3311 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
3312 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
3313 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
3316 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
3318 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
3319 Use gcc_unreachable for unhandled modes. Do not check results of
3320 expand_simple_binop. If not expanded to target, move the result.
3322 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
3323 Shiva Chen <shiva0217@gmail.com>
3325 * config.gcc (nds32*-*-*): Add nds32 target.
3326 * config/nds32/nds32.c: New file.
3327 * config/nds32/nds32.h: New file.
3328 * config/nds32/nds32.md: New file.
3329 * config/nds32/constants.md: New file.
3330 * config/nds32/constraints.md: New file.
3331 * config/nds32/iterators.md: New file.
3332 * config/nds32/nds32-doubleword.md: New file.
3333 * config/nds32/nds32-intrinsic.md: New file.
3334 * config/nds32/nds32_intrinsic.h: New file.
3335 * config/nds32/nds32-modes.def: New file.
3336 * config/nds32/nds32-multiple.md: New file.
3337 * config/nds32/nds32.opt: New file.
3338 * config/nds32/nds32-opts.h: New file.
3339 * config/nds32/nds32-protos.h: New file.
3340 * config/nds32/nds32-peephole2.md: New file.
3341 * config/nds32/pipelines.md: New file.
3342 * config/nds32/predicates.md: New file.
3343 * config/nds32/t-mlibs: New file.
3344 * common/config/nds32: New directory and files.
3346 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
3347 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
3348 * doc/install.texi (Cross-Compiler-Specific Options): Document
3349 --with-nds32-lib for nds32 target.
3350 * doc/extend.texi (Function Attributes, Target Builtins): Document
3351 nds32 specific attributes.
3353 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
3355 * lra-constraints (process_alt_operands): Use the result
3356 elimination register for operand when matching constraints.
3358 2013-10-31 Jakub Jelinek <jakub@redhat.com>
3360 * tree-vrp.c (maybe_set_nonzero_bits): New function.
3361 (remove_range_assertions): Call it.
3363 * tree.c (tree_ctz): New function.
3364 * tree.h (tree_ctz): New prototype.
3365 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
3366 first argument from tree to const_tree.
3367 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
3368 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
3369 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
3371 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
3372 * tree-vect-loop.c (vect_analyze_loop_operations,
3373 vect_transform_loop): Don't force scalar loop for bound just because
3374 number of iterations is unknown, only do it if it is not known to be
3375 a multiple of vectorization_factor.
3376 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
3378 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
3379 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
3380 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
3381 (set_nonzero_bits, get_nonzero_bits): New prototypes.
3382 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
3383 a default def isn't partially constant.
3384 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
3385 is known to be partially zero.
3386 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
3387 to see if a default def isn't partially constant.
3388 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
3389 creation of a range, if VR_RANGE, try to improve nonzero_bits from
3391 (set_nonzero_bits, get_nonzero_bits): New functions.
3393 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
3394 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
3395 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
3396 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
3397 uses but in the condition and ASSERT_EXPR and the other successor of
3398 the predecessor bb is __builtin_unreachable (), set_range_info of the
3399 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
3401 2013-10-31 Martin Jambor <mjambor@suse.cz>
3403 PR rtl-optimization/58934
3405 2013-10-30 Martin Jambor <mjambor@suse.cz>
3406 PR rtl-optimization/10474
3407 * ira.c (find_moveable_pseudos): Do not calculate dominance info
3409 (interesting_dest_for_shprep): New function.
3410 (split_live_ranges_for_shrink_wrap): Likewise.
3411 (ira): Calculate dominance info and df analysis. Call
3412 split_live_ranges_for_shrink_wrap.
3414 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3415 Yury Gribov <y.gribov@samsung.com>
3418 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
3419 shadow pointer to avoid clobbering the main one.
3421 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3423 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
3425 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3428 * ira-color.c (update_costs_from_copies): Add new parameter. Use
3429 it for calling update_costs_from_allocno.
3430 (assign_hard_reg): Call restore_costs_from_copies only for
3431 !retry_p. Pass new argument to update_costs_from_copies.
3432 (color_pass): Pass new argument to update_costs_from_copies.
3433 (ira_mark_allocation_change): Ditto.
3435 2013-10-30 Sharad Singhai <singhai@google.com>
3438 * opts.c (common_handle_option): Remove deprecated option
3439 -ftree-vectorizer-verbose.
3440 * doc/invoke.texi (Debugging Options): Ditto.
3441 * opts-global.c (handle_common_deferred_options): Ditto.
3442 (dump_remap_tree_vectorizer_verbose): Delete.
3443 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
3445 2013-10-30 DJ Delorie <dj@redhat.com>
3447 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
3448 that take no arguments.
3450 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
3453 * reload1.c (update_eliminables_and_spill): New function, broken
3455 (reload): Use it. Check for frame size change after frame size
3456 alignment, and call update_eliminables_and_spill first if continue-ing.
3458 2013-10-30 Cong Hou <congh@google.com>
3461 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
3462 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
3463 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
3465 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
3467 PR rtl-optimization/58369
3468 * reload1.c (compute_reload_subreg_offset): New function.
3469 (choose_reload_regs): Use it to pass endian-correct
3470 offset to subreg_regno_offset.
3472 2013-10-30 Tobias Burnus <burnus@net-b.de>
3475 * tree-cfg.c (replace_loop_annotate): Replace warning by
3478 2013-10-30 Jason Merrill <jason@redhat.com>
3480 * configure.ac (loose_warn): Add -Wno-format if
3481 --disable-build-format-warnings.
3483 2013-10-30 David Malcolm <dmalcolm@redhat.com>
3485 * cgraphunit.c (analyze_functions): Split symtab_node declarations
3486 onto multiple lines to make things easier for rename_symtab.py.
3488 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
3489 (symtab_semantically_equivalent_p): Likewise.
3491 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3494 * lra.c (check_rtl): Remove address check before LRA work.
3496 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
3498 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
3499 POINTER_PLUS_EXPR in the defining statement.
3501 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3503 * regmove.c: Remove.
3504 * tree-pass.h (make_pass_regmove): Remove.
3505 * timevar.def (TV_REGMOVE): Remove.
3506 * passes.def (pass_regmove): Remove.
3507 * opts.c (default_options_table): Remove entry for regmove.
3508 * doc/passes.texi: Remove regmove pass description.
3509 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
3511 (-fdump-rtl-regmove): Ditto.
3512 * common.opt (foptimize-register-move, fregmove): Ignore.
3513 * Makefile.in (OBJS): Remove regmove.o.
3514 * regmove.c: Remove.
3515 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
3517 (struct ira_allocno) New member allocno_prefs.
3518 (ALLOCNO_PREFS): New macro.
3519 (ira_prefs, ira_prefs_num): New external vars.
3520 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
3521 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
3522 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
3524 (ira_add_allocno_copy_to_list): Remove prototype.
3525 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
3526 (ira_pref_iterator): New type.
3527 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
3528 (FOR_EACH_PREF): New macro.
3529 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
3530 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
3532 (ira_setup_alts): New function.
3533 (decrease_live_ranges_number): New function.
3534 (ira): Call the above function.
3535 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
3536 (ira_create_allocno): Initialize allocno prefs.
3537 (pref_pool, pref_vec): New static vars.
3538 (initiate_prefs, find_allocno_pref, ira_create_pref): New
3540 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
3541 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
3542 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
3543 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
3544 (ira_add_allocno_copy_to_list): Make static. Rename to
3545 add_allocno_copy_to_list.
3546 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
3547 swap_allocno_copy_ends_if_necessary.
3548 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
3549 ira_remove_allocno_prefs.
3550 (ira_flattening): Ditto.
3551 (ira_build): Call initiate_prefs, print_prefs.
3552 (ira_destroy): Call finish_prefs.
3553 * ira-color.c (struct update_cost_record): New.
3554 (struct allocno_color_data): Add new member update_cost_records.
3555 (update_cost_record_pool): New static var.
3556 (init_update_cost_records, get_update_cost_record): New functions.
3557 (free_update_cost_record_list, finish_update_cost_records): Ditto.
3558 (struct update_cost_queue_elem): Add member from.
3559 (initiate_cost_update): Call init_update_cost_records.
3560 (finish_cost_update): Call finish_update_cost_records.
3561 (queue_update_cost, get_next_update_cost): Add new param from.
3562 (Update_allocno_cost, update_costs_from_allocno): New functions.
3563 (update_costs_from_prefs): Ditto.
3564 (update_copy_costs): Rename to update_costs_from_copies.
3565 (restore_costs_from_copies): New function.
3566 (update_conflict_hard_regno_costs): Don't go back.
3567 (assign_hard_reg): Call restore_costs_from_copies. Add printing
3569 (pop_allocnos): Add priniting more debug info.
3570 (color_allocnos): Remove prefs for conflicting hard regs.
3571 Call update_costs_from_prefs.
3572 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
3573 (get_dup_num): Rename, modify, and move to ira.c
3574 (process_regs_for_copy): Add prefs.
3575 (add_insn_allocno_copies): Put src as first arg of
3576 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
3577 * ira-costs.c (record_reg_classes): Modify and move code into
3578 record_operands_costs.
3579 (find_costs_and_classes): Create prefs for the hard reg of small
3581 (process_bb_node_for_hard_reg_moves): Add prefs.
3583 2013-10-30 Richard Biener <rguenther@suse.de>
3586 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
3587 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
3589 (pre_and_rev_post_order_compute): ... which now wraps it.
3590 * graph.c (draw_cfg_nodes_no_loops): Use
3591 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
3594 2013-10-30 Christian Bruel <christian.bruel@st.com>
3596 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
3597 (sh_expand_cmpstr): Handle known align and schedule improvements.
3598 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
3599 * config/sh/sh.md (cmpstrnsi): New pattern.
3601 2013-10-30 Martin Jambor <mjambor@suse.cz>
3603 PR rtl-optimization/10474
3604 * ira.c (find_moveable_pseudos): Do not calculate dominance info
3606 (interesting_dest_for_shprep): New function.
3607 (split_live_ranges_for_shrink_wrap): Likewise.
3608 (ira): Calculate dominance info and df analysis. Call
3609 split_live_ranges_for_shrink_wrap.
3611 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3614 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
3616 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3618 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
3619 * tree.h (POINTER_BOUNDS_P): New.
3620 (BOUNDED_TYPE_P): New.
3622 (pointer_bounds_type_node): New.
3623 * tree.c (build_common_tree_nodes): Initialize
3624 pointer_bounds_type_node.
3625 * gimple.h (gimple_call_get_nobnd_arg_index): New.
3626 (gimple_call_num_nobnd_args): New.
3627 (gimple_call_nobnd_arg): New.
3628 (gimple_return_retbnd): New.
3629 (gimple_return_set_retbnd): New
3630 * gimple.c (gimple_build_return): Increase number of ops
3631 for return statement.
3632 (gimple_call_get_nobnd_arg_index): New.
3633 * gimple-pretty-print.c (dump_gimple_return): Print second op.
3635 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3637 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
3638 with "chkp ctor" and "bnd_legacy" attributes.
3639 * gimplify.c (gimplify_init_constructor): Avoid infinite
3640 loop during gimplification of bounds initializer.
3642 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3644 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
3645 (handle_bnd_legacy): New.
3646 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
3647 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
3650 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
3652 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
3653 (BT_FN_PTR_CONST_PTR): New.
3654 (BT_FN_CONST_PTR_CONST_PTR): New.
3655 (BT_FN_PTR_CONST_PTR_SIZE): New.
3656 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
3657 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
3658 (BT_FN_VOID_CONST_PTR_SIZE): New.
3659 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
3660 * chkp-builtins.def: New.
3661 * builtins.def: include chkp-builtins.def.
3662 (DEF_CHKP_BUILTIN): New.
3663 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
3664 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
3665 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
3666 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
3667 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
3668 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
3669 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
3670 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
3671 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
3672 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
3673 * common.opt (fcheck-pointer-bounds): New.
3674 * toplev.c (process_options): Check Pointer Bounds Checker is
3676 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
3678 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3680 * target.def (builtin_chkp_function): New.
3681 (chkp_bound_type): New.
3682 (chkp_bound_mode): New.
3683 (fn_abi_va_list_bounds_size): New.
3684 (load_bounds_for_arg): New.
3685 (store_bounds_for_arg): New.
3686 * targhooks.h (default_load_bounds_for_arg): New.
3687 (default_store_bounds_for_arg): New.
3688 (default_fn_abi_va_list_bounds_size): New.
3689 (default_chkp_bound_type): New.
3690 (default_chkp_bound_mode): New.
3691 (default_builtin_chkp_function): New.
3692 * targhooks.c (default_load_bounds_for_arg): New.
3693 (default_store_bounds_for_arg): New.
3694 (default_fn_abi_va_list_bounds_size): New.
3695 (default_chkp_bound_type): New.
3696 (default_chkp_bound_mode); New.
3697 (default_builtin_chkp_function): New.
3698 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
3699 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
3700 (TARGET_STORE_BOUNDS_FOR_ARG): New.
3701 (TARGET_BUILTIN_CHKP_FUNCTION): New.
3702 (TARGET_CHKP_BOUND_TYPE): New.
3703 (TARGET_CHKP_BOUND_MODE): New.
3704 * doc/tm.texi: Regenerated.
3705 * langhooks.h (lang_hooks): Add chkp_supported field.
3706 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
3707 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
3709 2013-10-29 Andrew Pinski <apinski@cavium.com>
3711 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
3712 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
3713 fails, combining the branches anyways.
3714 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
3715 increases the number of combinings.
3716 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
3718 2013-10-29 Mike Stump <mikestump@comcast.net>
3720 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
3721 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
3722 (make_vector_mode): Increase namebuf to 16.
3723 (emit_insn_modes_h): When processing BImode, don't
3724 also match partial int modes.
3725 (emit_class_narrowest_mode): Likewise.
3727 * config/bfin/bfin-modes.def: Add precision to PDI.
3728 * config/m32c/m32c-modes.def: Add precision to PSI.
3729 * config/msp430/msp430-modes.def: Add precision to PSI.
3730 * config/rs6000/rs6000-modes.def: Add precision to PTI.
3731 * config/sh/sh-modes.def: Add precision to PSI and PDI.
3733 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
3736 * config/sh/sh.md (*addc): Rename existing variations to ...
3737 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
3738 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
3739 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
3740 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
3742 2013-10-29 Teresa Johnson <tejohnson@google.com>
3745 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
3746 insanities when updating probabilities.
3748 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3750 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
3751 prettyprinter for cgraph_node to reflect the conversion of the
3752 symtable types to a C++ class hierarchy: it now *is* a
3753 symtab_node_base, rather than having one (named "symbol").
3755 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
3757 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
3758 __cilkrts_pop_frame. If matched, then return true for built-in
3760 (expand_builtin): Added BUILT_IN_CILK_DETACH and
3761 BUILT_IN_CILK_POP_FRAME case.
3762 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
3763 (lhs_cilk_detect_spawn): Likewise.
3764 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
3765 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
3766 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
3767 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
3768 (LANG_HOOKS_CILKPLUS): Likewise.
3769 * tree.h (CILK_SPAWN_FN): Likewise.
3770 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
3771 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
3772 (OBJS): Added cilk-common.o.
3773 (BUILTINS_DEF): Added cilk-builtins.def.
3774 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
3775 (lhd_cilk_detect_spawn): Likewise.
3776 * langhooks.h (lang_hooks_for_cilkplus): New struct.
3777 (struct lang_hooks): Added new field called "cilkplus."
3778 * cilk-common.c: New file.
3780 * cilk-builtins.def: Likewise.
3781 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
3782 "__cilk" macro and set it to 200.
3783 * function.h (struct function::cilk_frame_decl): New field.
3784 (struct function::is_cilk_function): Likewise.
3785 (struct function::calls_cilk_spawn): Likewise.
3786 * gimplify.c (gimplify_call_expr): Added a check if the function call
3787 being gimplified is a spawn detach point. If so, then add pop_frame
3788 and detach function calls.
3789 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
3790 for gimplifying _Cilk_spawn and _Cilk_sync statements.
3791 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
3792 function. If so, added a _Cilk_sync and gimplified it.
3793 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
3794 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
3795 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
3797 (can_inline_edge_p): Prevent inling of spawnee function.
3798 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
3799 for functions that use Cilk keywords.
3800 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
3801 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
3802 CILK_SYNC_STMT cases.
3803 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
3805 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
3806 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
3807 * passes.texi (Cilk Keywords): New section that describes the compiler
3808 code changes for handling Cilk Keywords.
3810 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3812 Patch autogenerated by refactor_symtab.py from
3813 https://github.com/davidmalcolm/gcc-refactoring-scripts
3814 revision 58bb219cc090b2f4516a9297d868c245495ee622
3816 * asan.c (asan_finish_file): Update for conversion of symtab types to
3817 a true class hierarchy.
3818 * cfgexpand.c (estimated_stack_frame_size): Likewise.
3819 * cgraph.c (cgraph_get_body): Likewise.
3820 (cgraph_get_create_real_symbol_node): Likewise.
3821 (verify_cgraph_node): Likewise.
3822 (verify_edge_corresponds_to_fndecl): Likewise.
3823 (verify_edge_count_and_frequency): Likewise.
3824 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
3825 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3826 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
3827 (cgraph_node_cannot_return): Likewise.
3828 (cgraph_set_pure_flag_1): Likewise.
3829 (cgraph_set_const_flag_1): Likewise.
3830 (cgraph_set_nothrow_flag_1): Likewise.
3831 (cgraph_make_node_local_1): Likewise.
3832 (cgraph_for_node_and_aliases): Likewise.
3833 (cgraph_for_node_thunks_and_aliases): Likewise.
3834 (cgraph_node_can_be_local_p): Likewise.
3835 (cgraph_node_cannot_be_local_p_1): Likewise.
3836 (cgraph_function_body_availability): Likewise.
3837 (dump_cgraph_node): Likewise.
3838 (cgraph_rtl_info): Likewise.
3839 (cgraph_mark_address_taken_node): Likewise.
3840 (cgraph_remove_node): Likewise.
3841 (cgraph_release_function_body): Likewise.
3842 (cgraph_update_edges_for_call_stmt_node): Likewise.
3843 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
3844 (cgraph_make_edge_direct): Likewise.
3845 (cgraph_resolve_speculation): Likewise.
3846 (cgraph_speculative_call_info): Likewise.
3847 (cgraph_turn_edge_to_speculative): Likewise.
3848 (cgraph_create_edge_1): Likewise.
3849 (cgraph_set_call_stmt): Likewise.
3850 (cgraph_node_for_asm): Likewise.
3851 (cgraph_add_thunk): Likewise.
3852 (cgraph_same_body_alias): Likewise.
3853 (cgraph_create_function_alias): Likewise.
3854 (cgraph_create_node): Likewise.
3855 (cgraph_create_empty_node): Likewise.
3856 (record_function_versions): Likewise.
3857 (used_from_object_file_p): Likewise.
3858 * cgraph.h (symtab_can_be_discarded): Likewise.
3859 (symtab_real_symbol_p): Likewise.
3860 (cgraph_mark_force_output_node): Likewise.
3861 (cgraph_edge_recursive_p): Likewise.
3862 (symtab_alias_target): Likewise.
3863 (varpool_all_refs_explicit_p): Likewise.
3864 (varpool_can_remove_if_no_refs): Likewise.
3865 (cgraph_only_called_directly_or_aliased_p): Likewise.
3866 (cgraph_next_function_with_gimple_body): Likewise.
3867 (cgraph_first_function_with_gimple_body): Likewise.
3868 (cgraph_function_with_gimple_body_p): Likewise.
3869 (cgraph_next_function): Likewise.
3870 (cgraph_first_function): Likewise.
3871 (cgraph_next_defined_function): Likewise.
3872 (cgraph_first_defined_function): Likewise.
3873 (varpool_next_defined_variable): Likewise.
3874 (varpool_first_defined_variable): Likewise.
3875 (varpool_next_static_initializer): Likewise.
3876 (varpool_first_static_initializer): Likewise.
3877 (varpool_next_variable): Likewise.
3878 (varpool_first_variable): Likewise.
3879 (varpool_node_name): Likewise.
3880 (varpool): Likewise.
3882 (is_a_helper <varpool_node>::test): Likewise.
3883 (is_a_helper <cgraph_node>::test): Likewise.
3884 (varpool_variable_node): Likewise.
3885 (cgraph_function_or_thunk_node): Likewise.
3886 (varpool_alias_target): Likewise.
3887 (cgraph_alias_target): Likewise.
3888 (cgraph_node_name): Likewise.
3889 (varpool_node_asm_name): Likewise.
3890 (cgraph_node_asm_name): Likewise.
3891 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
3892 (cgraph_rebuild_references): Likewise.
3893 (rebuild_cgraph_edges): Likewise.
3894 (record_eh_tables): Likewise.
3895 (build_cgraph_edges): Likewise.
3896 (mark_store): Likewise.
3897 (mark_load): Likewise.
3898 (mark_address): Likewise.
3899 (record_type_list): Likewise.
3900 (record_reference): Likewise.
3901 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
3902 (cgraph_materialize_clone): Likewise.
3903 (cgraph_function_versioning): Likewise.
3904 (cgraph_copy_node_for_versioning): Likewise.
3905 (update_call_expr): Likewise.
3906 (cgraph_find_replacement_node): Likewise.
3907 (cgraph_create_virtual_clone): Likewise.
3908 (cgraph_clone_node): Likewise.
3909 * cgraphunit.c (compile): Likewise.
3910 (output_weakrefs): Likewise.
3911 (output_in_order): Likewise.
3912 (expand_function): Likewise.
3913 (assemble_thunks_and_aliases): Likewise.
3914 (expand_thunk): Likewise.
3915 (mark_functions_to_output): Likewise.
3916 (handle_alias_pairs): Likewise.
3917 (analyze_functions): Likewise.
3918 (walk_polymorphic_call_targets): Likewise.
3919 (varpool_finalize_decl): Likewise.
3920 (process_function_and_variable_attributes): Likewise.
3921 (cgraph_process_same_body_aliases): Likewise.
3922 (analyze_function): Likewise.
3923 (cgraph_add_new_function): Likewise.
3924 (cgraph_finalize_function): Likewise.
3925 (referred_to_p): Likewise.
3926 (cgraph_reset_node): Likewise.
3927 (cgraph_process_new_functions): Likewise.
3928 (enqueue_node): Likewise.
3929 (decide_is_symbol_needed): Likewise.
3930 * coverage.c (coverage_compute_profile_id): Likewise.
3931 * dbxout.c (dbxout_expand_expr): Likewise.
3932 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
3933 (reference_to_unused): Likewise.
3934 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3935 * gimplify.c (unvisit_body): Likewise.
3936 (unshare_body): Likewise.
3937 * ipa-cp.c (ipcp_generate_summary): Likewise.
3938 (ipcp_decision_stage): Likewise.
3939 (identify_dead_nodes): Likewise.
3940 (decide_whether_version_node): Likewise.
3941 (decide_about_value): Likewise.
3942 (perhaps_add_new_callers): Likewise.
3943 (create_specialized_node): Likewise.
3944 (update_profiling_info): Likewise.
3945 (ipcp_propagate_stage): Likewise.
3946 (estimate_local_effects): Likewise.
3947 (good_cloning_opportunity_p): Likewise.
3948 (devirtualization_time_bonus): Likewise.
3949 (propagate_constants_accross_call): Likewise.
3950 (initialize_node_lattices): Likewise.
3951 (ipcp_cloning_candidate_p): Likewise.
3952 (determine_versionability): Likewise.
3953 (print_all_lattices): Likewise.
3954 (print_lattice): Likewise.
3955 (ipcp_discover_new_direct_edges): Likewise.
3956 * ipa-devirt.c (ipa_devirt): Likewise.
3957 (likely_target_p): Likewise.
3958 (update_type_inheritance_graph): Likewise.
3959 (possible_polymorphic_call_target_p): Likewise.
3960 (dump_possible_polymorphic_call_targets): Likewise.
3961 (devirt_variable_node_removal_hook): Likewise.
3962 (record_binfo): Likewise.
3963 (maybe_record_node): Likewise.
3964 (build_type_inheritance_graph): Likewise.
3965 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3966 (inline_generate_summary): Likewise.
3967 (inline_analyze_function): Likewise.
3968 (do_estimate_growth): Likewise.
3969 (simple_edge_hints): Likewise.
3970 (estimate_node_size_and_time): Likewise.
3971 (estimate_edge_devirt_benefit): Likewise.
3972 (compute_inline_parameters): Likewise.
3973 (estimate_function_body_sizes): Likewise.
3974 (compute_bb_predicates): Likewise.
3975 (initialize_inline_failed): Likewise.
3976 (dump_inline_summary): Likewise.
3977 (dump_inline_edge_summary): Likewise.
3978 * ipa-inline-transform.c (inline_transform): Likewise.
3979 (preserve_function_body_p): Likewise.
3980 (save_inline_function_body): Likewise.
3981 (inline_call): Likewise.
3982 (clone_inlined_nodes): Likewise.
3983 (can_remove_node_now_p): Likewise.
3984 (can_remove_node_now_p_1): Likewise.
3985 * ipa-inline.c (early_inliner): Likewise.
3986 (early_inline_small_functions): Likewise.
3987 (inline_always_inline_functions): Likewise.
3988 (ipa_inline): Likewise.
3989 (flatten_function): Likewise.
3990 (inline_small_functions): Likewise.
3991 (speculation_useful_p): Likewise.
3992 (recursive_inlining): Likewise.
3993 (update_caller_keys): Likewise.
3994 (reset_edge_caches): Likewise.
3995 (update_edge_key): Likewise.
3996 (edge_badness): Likewise.
3997 (relative_time_benefit): Likewise.
3998 (want_inline_self_recursive_call_p): Likewise.
3999 (want_inline_small_function_p): Likewise.
4000 (want_early_inline_function_p): Likewise.
4001 (num_calls): Likewise.
4002 (can_early_inline_edge_p): Likewise.
4003 (can_inline_edge_p): Likewise.
4004 (report_inline_failed_reason): Likewise.
4005 * ipa-profile.c (ipa_profile): Likewise.
4006 (ipa_propagate_frequency): Likewise.
4007 (ipa_propagate_frequency_1): Likewise.
4008 (ipa_profile_generate_summary): Likewise.
4009 * ipa-prop.c (ipcp_transform_function): Likewise.
4010 (read_replacements_section): Likewise.
4011 (ipa_prop_read_section): Likewise.
4012 (ipa_modify_call_arguments): Likewise.
4013 (ipa_print_node_params): Likewise.
4014 (propagate_controlled_uses): Likewise.
4015 (update_indirect_edges_after_inlining): Likewise.
4016 (remove_described_reference): Likewise.
4017 (ipa_make_edge_direct_to_target): Likewise.
4018 (ipa_analyze_node): Likewise.
4019 (ipa_analyze_params_uses): Likewise.
4020 (ipa_compute_jump_functions): Likewise.
4021 (ipa_get_callee_param_type): Likewise.
4022 (ipa_print_node_jump_functions): Likewise.
4023 (ipa_initialize_node_params): Likewise.
4024 (ipa_populate_param_decls): Likewise.
4025 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
4026 (write_agg_replacement_chain): Likewise.
4027 (ipa_write_node_info): Likewise.
4028 (ipa_edge_duplication_hook): Likewise.
4029 (try_decrement_rdesc_refcount): Likewise.
4030 * ipa-pure-const.c (propagate_nothrow): Likewise.
4031 (propagate_pure_const): Likewise.
4032 (pure_const_read_summary): Likewise.
4033 (pure_const_write_summary): Likewise.
4034 (analyze_function): Likewise.
4035 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
4036 (ipa_ref_referring_ref_list): Likewise.
4037 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
4038 (ipa_remove_stmt_references): Likewise.
4039 (ipa_find_reference): Likewise.
4040 (ipa_dump_referring): Likewise.
4041 (ipa_dump_references): Likewise.
4042 (ipa_record_reference): Likewise.
4043 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
4044 (ipa_reference_write_optimization_summary): Likewise.
4045 (write_node_summary_p): Likewise.
4046 (propagate): Likewise.
4047 (read_write_all_from_decl): Likewise.
4048 (generate_summary): Likewise.
4049 (analyze_function): Likewise.
4050 (propagate_bits): Likewise.
4051 (ipa_reference_get_not_written_global): Likewise.
4052 (ipa_reference_get_not_read_global): Likewise.
4053 * ipa-split.c (execute_split_functions): Likewise.
4054 (split_function): Likewise.
4055 * ipa-utils.c (ipa_merge_profiles): Likewise.
4056 (dump_cgraph_node_set): Likewise.
4057 (ipa_reverse_postorder): Likewise.
4058 (ipa_edge_within_scc): Likewise.
4059 (ipa_get_nodes_in_cycle): Likewise.
4060 (ipa_free_postorder_info): Likewise.
4061 (ipa_reduced_postorder): Likewise.
4062 (searchc): Likewise.
4063 (recursive_call_p): Likewise.
4064 * ipa.c (ipa_cdtor_merge): Likewise.
4065 (record_cdtor_fn): Likewise.
4066 (function_and_variable_visibility): Likewise.
4067 (varpool_externally_visible_p): Likewise.
4068 (cgraph_externally_visible_p): Likewise.
4069 (comdat_can_be_unshared_p): Likewise.
4070 (comdat_can_be_unshared_p_1): Likewise.
4071 (address_taken_from_non_vtable_p): Likewise.
4072 (ipa_discover_readonly_nonaddressable_vars): Likewise.
4073 (symtab_remove_unreachable_nodes): Likewise.
4074 (walk_polymorphic_call_targets): Likewise.
4075 (process_references): Likewise.
4076 (enqueue_node): Likewise.
4077 (has_addr_references_p): Likewise.
4078 (cgraph_non_local_node_p_1): Likewise.
4079 * is-a.h (varpool_analyze_node): Likewise.
4080 * lto-cgraph.c (input_symtab): Likewise.
4081 (merge_profile_summaries): Likewise.
4082 (input_cgraph_1): Likewise.
4083 (input_edge): Likewise.
4084 (input_varpool_node): Likewise.
4085 (input_node): Likewise.
4086 (input_overwrite_node): Likewise.
4087 (compute_ltrans_boundary): Likewise.
4088 (output_refs): Likewise.
4089 (lto_output_varpool_node): Likewise.
4090 (lto_output_node): Likewise.
4091 (reachable_from_other_partition_p): Likewise.
4092 (referenced_from_other_partition_p): Likewise.
4093 (lto_output_edge): Likewise.
4094 (output_node_opt_summary): Likewise.
4095 (add_node_to): Likewise.
4096 (reachable_from_this_partition_p): Likewise.
4097 (lto_set_symtab_encoder_in_partition): Likewise.
4098 (lto_symtab_encoder_in_partition_p): Likewise.
4099 (lto_set_symtab_encoder_encode_initializer): Likewise.
4100 (lto_symtab_encoder_encode_initializer_p): Likewise.
4101 (lto_set_symtab_encoder_encode_body): Likewise.
4102 (lto_symtab_encoder_encode_body_p): Likewise.
4103 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
4105 * lto-streamer-in.c (lto_read_body): Likewise.
4106 (fixup_call_stmt_edges): Likewise.
4107 (fixup_call_stmt_edges_1): Likewise.
4108 * lto-streamer-out.c (produce_symtab): Likewise.
4109 (output_symbol_p): Likewise.
4110 (write_symbol): Likewise.
4111 (lto_output): Likewise.
4112 (copy_function): Likewise.
4113 (output_function): Likewise.
4114 * passes.c (function_called_by_processed_nodes_p): Likewise.
4115 (ipa_write_optimization_summaries): Likewise.
4116 (ipa_write_summaries): Likewise.
4117 (do_per_function_toporder): Likewise.
4118 (do_per_function): Likewise.
4119 (dump_passes): Likewise.
4120 * symtab.c (symtab_semantically_equivalent_p): Likewise.
4121 (symtab_nonoverwritable_alias): Likewise.
4122 (symtab_nonoverwritable_alias_1): Likewise.
4123 (symtab_for_node_and_aliases): Likewise.
4124 (symtab_resolve_alias): Likewise.
4125 (fixup_same_cpp_alias_visibility): Likewise.
4126 (symtab_alias_ultimate_target): Likewise.
4127 (symtab_used_from_object_file_p): Likewise.
4128 (verify_symtab_base): Likewise.
4129 (dump_symtab_base): Likewise.
4130 (symtab_node_name): Likewise.
4131 (symtab_node_asm_name): Likewise.
4132 (symtab_dissolve_same_comdat_group_list): Likewise.
4133 (symtab_add_to_same_comdat_group): Likewise.
4134 (symtab_unregister_node): Likewise.
4135 (symtab_insert_node_to_hashtable): Likewise.
4136 (symtab_register_node): Likewise.
4137 (unlink_from_assembler_name_hash): Likewise.
4138 (insert_to_assembler_name_hash): Likewise.
4139 (eq_assembler_name): Likewise.
4140 (hash_node_by_assembler_name): Likewise.
4141 (eq_node): Likewise.
4142 (hash_node): Likewise.
4143 * toplev.c (wrapup_global_declaration_2): Likewise.
4144 * trans-mem.c (ipa_tm_execute): Likewise.
4145 (ipa_tm_transform_clone): Likewise.
4146 (ipa_tm_transform_transaction): Likewise.
4147 (ipa_tm_transform_calls_redirect): Likewise.
4148 (ipa_tm_insert_gettmclone_call): Likewise.
4149 (ipa_tm_insert_irr_call): Likewise.
4150 (ipa_tm_create_version): Likewise.
4151 (ipa_tm_create_version_alias): Likewise.
4152 (ipa_tm_mark_forced_by_abi_node): Likewise.
4153 (ipa_tm_mark_force_output_node): Likewise.
4154 (ipa_tm_diagnose_tm_safe): Likewise.
4155 (ipa_tm_mayenterirr_function): Likewise.
4156 (ipa_tm_scan_irr_function): Likewise.
4157 (ipa_tm_note_irrevocable): Likewise.
4158 (ipa_tm_scan_calls_clone): Likewise.
4159 (get_cg_data): Likewise.
4160 * tree-eh.c (tree_could_trap_p): Likewise.
4161 * tree-emutls.c (ipa_lower_emutls): Likewise.
4162 (create_emultls_var): Likewise.
4163 (lower_emutls_function_body): Likewise.
4164 (gen_emutls_addr): Likewise.
4165 (emutls_decl): Likewise.
4166 (new_emutls_decl): Likewise.
4167 * tree-inline.c (tree_function_versioning): Likewise.
4168 (optimize_inline_calls): Likewise.
4169 (expand_call_inline): Likewise.
4170 (estimate_num_insns): Likewise.
4171 (copy_bb): Likewise.
4172 (delete_unreachable_blocks_update_callgraph): Likewise.
4173 * tree-nested.c (gimplify_all_functions): Likewise.
4174 (create_nesting_tree): Likewise.
4175 (check_for_nested_with_variably_modified): Likewise.
4176 * tree-pretty-print.c (dump_function_header): Likewise.
4177 * tree-profile.c (tree_profiling): Likewise.
4178 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
4179 (modify_function): Likewise.
4180 (convert_callers): Likewise.
4181 (convert_callers_for_node): Likewise.
4182 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
4183 (associate_varinfo_to_alias): Likewise.
4184 (create_variable_info_for): Likewise.
4185 (get_constraint_for_ssa_var): Likewise.
4186 * tree-vectorizer.c (increase_alignment): Likewise.
4187 * tree.c (find_decls_types_in_var): Likewise.
4188 (find_decls_types_in_node): Likewise.
4189 (free_lang_data_in_decl): Likewise.
4190 * value-prof.c (gimple_ic_transform): Likewise.
4191 (gimple_ic): Likewise.
4192 (check_ic_target): Likewise.
4193 (init_node_map): Likewise.
4194 * varasm.c (decl_binds_to_current_def_p): Likewise.
4195 (default_binds_local_p_1): Likewise.
4196 (dump_tm_clone_pairs): Likewise.
4197 (assemble_alias): Likewise.
4198 (find_decl): Likewise.
4199 (mark_decl_referenced): Likewise.
4200 * varpool.c (varpool_for_node_and_aliases): Likewise.
4201 (varpool_extra_name_alias): Likewise.
4202 (varpool_create_variable_alias): Likewise.
4203 (add_new_static_var): Likewise.
4204 (varpool_finalize_named_section_flags): Likewise.
4205 (varpool_remove_unreferenced_decls): Likewise.
4206 (enqueue_node): Likewise.
4207 (varpool_assemble_decl): Likewise.
4208 (assemble_aliases): Likewise.
4209 (varpool_analyze_node): Likewise.
4210 (cgraph_variable_initializer_availability): Likewise.
4211 (varpool_add_new_variable): Likewise.
4212 (ctor_for_folding): Likewise.
4213 (dump_varpool_node): Likewise.
4214 (varpool_remove_initializer): Likewise.
4215 (varpool_remove_node): Likewise.
4216 (varpool_node_for_decl): Likewise.
4217 (varpool_create_empty_node): Likewise.
4218 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
4219 (ix86_get_function_versions_dispatcher): Likewise.
4221 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4223 * cgraph.h (symtab_node_base): Convert to a class;
4224 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
4225 chain_next/prev from symtab_node_def.
4226 (cgraph_node): Inherit from symtab_node; add GTY option
4227 tag ("SYMTAB_FUNCTION").
4228 (varpool_node): Inherit from symtab_node; add GTY option
4229 tag ("SYMTAB_VARIABLE").
4230 (symtab_node_def): Remove.
4231 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
4232 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
4233 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
4234 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
4236 * ipa-ref.h (symtab_node_def): Drop.
4237 (symtab_node): Change underlying type from symtab_node_def to
4239 (const_symtab_node): Likwise.
4241 * is-a.h: Update examples in comment.
4243 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
4244 (assembler_name_hash): Likewise.
4246 2013-10-29 Martin Liska <marxin.liska@gmail.com>
4248 * doc/tree-ssa.texi (gimple_phi_result): Document.
4249 (gimple_phi_num_args, gimple_phi_arg): Likewise.
4250 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
4251 (PHI_RESULT, PHI_NUM_ARGS): Remove.
4252 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
4254 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
4256 * expr.h: Revert change and include tree-core.h.
4257 * rtl.h: Revert change and don't include tree-core.h.
4259 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
4261 * config/darwin.c: Include gimple.h.
4262 * config/i386/winnt.c: Likewise.
4264 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
4266 PR tree-optimization/19831
4267 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
4269 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
4271 * tree-outof-ssa.h: Remove include files.
4272 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
4274 * tree-ssa-coalesce.c: Likewise.
4275 * cfgexpand.c: Likewise.
4276 * tree-ssa-ter.c: Likewise.
4277 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
4278 * lto-streamer.h: Likewise.
4279 * cgraphbuild.c: Add gimple.h to include list.
4280 * data-streamer-in.c: Likewise.
4281 * ipa-cp.c: Likewise.
4282 * tree-streamer.c: Likewise.
4283 * lto-compress.c: Likewise.
4284 * ipa-reference.c: Likewise.
4285 * data-streamer-out.c: Likewise.
4286 * lto-cgraph.c: Likewise.
4287 * cgraphclones.c: Likewise.
4288 * ipa-utils.c: Likewise.
4289 * data-streamer.c: Likewise.
4290 * ipa-split.c: Likewise.
4291 * lto-section-in.c: Likewise.
4292 * tree-streamer-out.c: Likewise.
4293 * ipa-prop.c: Likewise.
4294 * tree-streamer-in.c: Likewise.
4295 * symtab.c: Likewise.
4296 * opts-global.c: Likewise.
4297 * lto-opts.c: Likewise.
4298 * lto-section-out.c: Likewise.
4299 * lto-streamer.c: Likewise.
4300 * rtl.h: Add tree-core.h to include list.
4301 * expr.h: Remove tree-core.h from include list.
4302 * gimple.h: Likewise.
4303 * ipa-utils.h: Likewise.
4304 * streamer-hooks.h: Likewise.
4305 * streamer-hooks.c: Include input.h.
4307 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4309 * config/arm/arm.c (cortexa7_extra_costs): New table.
4310 (arm_cortex_a7_tune): New.
4311 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
4313 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
4315 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
4317 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4319 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
4320 to use autogenerated markers for a class-hierarchy, every class
4321 must have a GTY marker.
4322 * gengtype.h (struct type): Add linked list of subclasses to
4323 the "s" member of the union.
4324 (add_subclass): New decl.
4325 * gengtype-state.c (read_state_struct_type): Set up subclass
4327 * gengtype.c (get_ultimate_base_class): New.
4328 (add_subclass): New.
4329 (new_structure): Set up subclass linked list.
4330 (set_gc_used_type): Propagate usage information to subclasses.
4331 (output_mangled_typename): Use get_ultimate_base_class.
4332 (walk_subclasses): Use the subclass linked list, avoiding an
4333 O(N^2) when writing out all types.
4334 (walk_type): Issue an error if the base class is missing a tag,
4335 rather than generating bogus C code. Add a gcc_unreachable
4336 default case, in case people omit tags from concrete subclasses,
4337 or get the values wrong.
4338 (write_func_for_structure): Issue an error for subclasses for
4339 which the base doesn't have a "desc", since otherwise the
4340 autogenerated routines for the base would silently fail to visit
4341 any subclass fields.
4342 (write_root): Use get_ultimate_base_class, tweaking constness of
4343 tp to match that function's signature.
4345 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4347 * doc/gty.texi (GTY Options): Add note about inheritance to
4348 description of desc and tag.
4349 (Inheritance and GTY): New.
4351 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4353 * gengtype-parse.c (opts_have): Drop "static" so that
4354 we can use this from gengtype.c.
4355 * gengtype.c (set_gc_used_type): Mark any base class as used;
4356 update field traversal to visit inherited fields.
4357 (output_mangled_typename): Convert references to classes within
4358 an inheritance hierarchy to reference the ultimate base class,
4359 since only it will have gt_ functions.
4360 (get_string_option): New.
4361 (walk_subclasses): New.
4362 (walk_type): Treat GTY structs that have a "desc" as being the
4363 root of an inheritance hierarchy. Generate a switch on it
4364 within the marking function which walks all subclasses, adding
4365 cases for them via walk_subclasses. For subclasses, visit all
4366 fields of the type (including inherited ones).
4367 (write_func_for_structure): Don't write fns for subclasses, only
4368 for the ultimate base class within an inheritance hierarchy.
4369 Subclasses-marking will be handled by the base class marking functions.
4370 (write_types): Likewise.
4371 (write_local_func_for_structure): Likewise.
4372 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
4373 a "tag" option (and are thus concrete subclasses).
4374 (write_root): Use the marker function for the ultimate base class.
4375 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
4376 (opts_have): Add declaration.
4378 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
4380 * lra-spills.c (lra_final_code_change): Remove useless move insns
4381 originated from moves of pseudos.
4383 2013-10-28 Jeff Law <law@redhat.com>
4385 * lower-subreg.c (resolve_simple_move): Fix comment typo.
4387 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
4389 * df-scan.c (df_collection_rec): Adjust.
4390 (copy_defs): New constant.
4391 (copy_uses): Likewise.
4392 (copy_eq_uses): Likewise.
4393 (copy_mw): Likewise.
4394 (copy_all): Likewise.
4395 (df_insn_rescan): Adjust.
4396 (df_notes_rescan): Likewise.
4397 (df_swap_refs): Likewise.
4398 (df_sort_and_compress_refs): Likewise.
4399 (df_sort_and_compress_mws): Likewise.
4400 (df_install_refs): Likewise.
4401 (df_install_mws): Likewise.
4402 (df_refs_add_to_chains): Add flags parameter controlling which vectors
4404 (df_bb_refs_record): Adjust.
4405 (df_record_entry_block_defs): Likewise.
4406 (df_record_exit_block_defs): Likewise.
4407 (df_refs_verify): Likewise.
4408 (df_mws_verify): Likewise.
4409 (df_insn_refs_verify): Likewise.
4410 (df_bb_verify): Likewise.
4411 * ipa-pure-const.c (finish_state): Remove.
4412 (propagate): Adjust.
4413 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
4414 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
4415 var-tracking.c: Adjust.
4416 * vec.c (stack_vecs): Remove.
4417 (register_stack_vec): Likewise.
4418 (stack_vec_register_index): Likewise.
4419 (unregister_stack_vec): Likewise.
4420 * vec.h (struct va_stack): Remove.
4421 (struct vec<T, A, vl_ptr>): Specialize as
4422 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
4423 allocation strategy compatable with the vl_ptr layout.
4424 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
4425 specialization anyway.
4426 (class stack_vec): New class.
4427 (vec_stack_alloc): Remove.
4428 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
4430 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
4431 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4432 Sergey Lega <sergey.s.lega@intel.com>
4433 Anna Tikhonova <anna.tikhonova@intel.com>
4434 Ilya Tocar <ilya.tocar@intel.com>
4435 Andrey Turetskiy <andrey.turetskiy@intel.com>
4436 Ilya Verbin <ilya.verbin@intel.com>
4437 Kirill Yukhin <kirill.yukhin@intel.com>
4438 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4440 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
4441 (*prefetch_avx512pf_<mode>): New.
4442 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
4443 (avx512f_maskcmp<mode>3): Ditto.
4444 (vashrv16si3): Ditto.
4446 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
4447 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4448 Sergey Lega <sergey.s.lega@intel.com>
4449 Anna Tikhonova <anna.tikhonova@intel.com>
4450 Ilya Tocar <ilya.tocar@intel.com>
4451 Andrey Turetskiy <andrey.turetskiy@intel.com>
4452 Ilya Verbin <ilya.verbin@intel.com>
4453 Kirill Yukhin <kirill.yukhin@intel.com>
4454 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4456 * config/i386/i386.md (any_truncate): New.
4457 (trunsuffix): Ditto.
4458 * config/i386/predicates.md (const_8_to_9_operand): New.
4459 (const_10_to_11_operand): Ditto.
4460 (const_12_to_13_operand): Ditto.
4461 (const_14_to_15_operand): Ditto.
4462 (const_16_to_19_operand): Ditto.
4463 (const_20_to_23_operand): Ditto.
4464 (const_24_to_27_operand): Ditto.
4465 (const_28_to_31_operand): Ditto.
4466 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
4467 (cvtusi2<ssescalarmodesuffix>32): New.
4468 (cvtusi2<ssescalarmodesuffix>64): Ditto.
4469 (ufloatv16siv16sf2): Ditto.
4470 (avx512f_fix_notruncv16sfv16si): Ditto.
4471 (avx512f_ufix_notruncv16sfv16si): Ditto.
4472 (avx512f_vcvtss2usi): Ditto.
4473 (avx512f_vcvtss2usiq): Ditto.
4474 (avx512f_vcvttss2usi): Ditto.
4475 (avx512f_vcvttss2usiq): Ditto.
4476 (avx512f_vcvtsd2usi): Ditto.
4477 (avx512f_vcvtsd2usiq): Ditto.
4478 (avx512f_vcvttsd2usi): Ditto.
4479 (avx512f_vcvttsd2usiq): Ditto.
4480 (ufloatv8siv8df): Ditto.
4481 (avx512f_cvtdq2pd512_2): Ditto.
4482 (avx512f_cvtpd2dq512): Ditto.
4483 (avx512f_ufix_notruncv8dfv8si): Ditto.
4484 (avx512f_cvtpd2ps512): Ditto.
4485 (vec_unpacks_lo_v16sf): Ditto.
4486 (vec_unpacks_hi_v16sf): Ditto.
4487 (vec_unpacks_float_hi_v16si): Ditto.
4488 (vec_unpacks_float_lo_v16si): Ditto.
4489 (avx512f_unpckhps512): Ditto.
4490 (avx512f_unpcklps512): Ditto.
4491 (avx512f_movshdup512): Ditto.
4492 (avx512f_movsldup512): Ditto.
4493 (vec_extract_lo_v32hi): Ditto.
4494 (vec_extract_hi_v32hi): Ditto.
4495 (vec_extract_lo_v64qi): Ditto.
4496 (vec_extract_hi_v64qi): Ditto.
4497 (avx512f_unpckhpd512): Ditto.
4498 (avx512f_movddup512): Ditto.
4499 (avx512f_unpcklpd512): Ditto.
4500 (*avx512f_unpcklpd512): Ditto.
4501 (avx512f_shufps512_1): Ditto.
4502 (avx512f_shufpd512_1): Ditto.
4503 (avx512f_interleave_highv8di): Ditto.
4504 (avx512f_interleave_lowv8di): Ditto.
4505 (PMOV_DST_MODE): Ditto.
4506 (pmov_src_mode): Ditto.
4507 (pmov_src_lower): Ditto.
4509 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
4510 (*avx512f_<code>v8div16qi2): Ditto.
4511 (*avx512f_<code>v8div16qi2_store): Ditto.
4512 (vec_widen_umult_even_v16si): Ditto.
4513 (*vec_widen_umult_even_v16si): Ditto.
4514 (vec_widen_smult_even_v16si): Ditto.
4515 (*vec_widen_smult_even_v16si): Ditto.
4516 (avx512f_interleave_highv16si): Ditto.
4517 (avx512f_interleave_lowv16si): Ditto.
4518 (avx512f_<code>v16qiv16si2): Ditto.
4519 (avx512f_<code>v16hiv16si2): Ditto.
4520 (avx512f_<code>v8qiv8di2): Ditto.
4521 (avx512f_<code>v8hiv8di2): Ditto.
4522 (avx512f_<code>v8siv8di2): Ditto.
4523 (avx512cd_maskb_vec_dupv8di): Ditto.
4524 (avx512cd_maskw_vec_dupv16si): Ditto.
4525 (avx512f_vcvtph2ps512): Ditto.
4526 (avx512f_vcvtps2ph512): Ditto.
4527 (VEC_EXTRACT_MODE): Extened with wider modes.
4528 (VEC_PERM_AVX2): Ditto.
4529 (VEC_PERM_CONST): Ditto.
4531 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
4533 * config/arc/arc.c (arc_ccfsm_post_advance):
4534 Add comment about TYPE_RETURN.
4536 2013-10-28 Bin Cheng <bin.cheng@arm.com>
4538 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
4539 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
4540 (strip_offset): Convert offset to unsigned number.
4542 2013-10-27 Tom de Vries <tom@codesourcery.com>
4544 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
4545 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
4547 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
4549 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
4550 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
4551 (TARGET_BIG_ENDIAN): New macro.
4552 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
4553 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
4554 * config/sh/sh.c: Likewise.
4555 * config/sh/sh.h: Likewise.
4557 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
4559 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
4560 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
4562 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
4564 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
4566 (cmpstr_t, cmpstrsi): Fix formatting.
4568 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
4571 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
4572 addressing, do not use general_operand for memory operands.
4574 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
4577 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4578 * lra-spills.c (lra_final_code_change): Remove useless move insns.
4580 2013-10-26 Jeff Law <law@redhat.com>
4582 * predict.c (PRED_MUDFLAP): Remove.
4583 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
4585 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
4586 (OBJS): Remove tree-nomudflap.o
4587 (GTFILES): Remove tree-mudflap.c
4588 * builtins.c (expand_builtin_alloc): Remove mudflap support.
4589 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
4590 (mfwrap_spec, mflib_spec): Likewise.
4591 (cpp_unique_options, cc1_options, static_specs): Likewise.
4592 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
4593 * passes.def: Likewise.
4594 * toplev.c (compile_file, process_options): Likewise.
4595 * tree-inline.c (copy_tree_r): Likewise.
4596 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
4597 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
4598 (build_constant_desc, output_constant_def_contents): Likewise.
4599 (categorize_decl_for_section): Likewise.
4600 * tree-mudflap.c: Removed.
4601 * tree-mudflap.h: Removed.
4602 * tree-nomudflap.c: Removed.
4603 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
4604 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
4605 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
4606 * config/sol2.h (MFLIB_SPEC): Likewise.
4607 * doc/install.texi: Remove mudflap references.
4608 * doc/passes.texi: Similarly.
4609 * doc/sourcebuild.texi: Similarly.
4610 * doc/invoke.texi: Remove mudlfap related options.
4612 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4614 PR rtl-optimization/58759
4615 * lra-constraints.c (lra_constraints): Remove wrong condition to
4616 remove insn setting up an equivalent pseudo.
4618 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4620 * config/rs6000/rs6000-protos.h
4621 (rs6000_secondary_memory_needed_mode): New prototype.
4622 * config/rs6000/rs6000.c: Include ira.h.
4623 (TARGET_LRA_P): Redefine.
4624 (rs6000_legitimate_offset_address_p): Call
4625 legitimate_constant_pool_address_p in strict mode for LRA.
4626 (rs6000_legitimate_address_p): Ditto.
4627 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
4628 (rs6000_emit_move): Add LRA version of code to generate load/store
4630 (rs6000_secondary_memory_needed_mode): New.
4631 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
4632 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
4633 constants, memory, and FP registers.
4634 (rs6000_lra_p): New.
4635 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
4636 * config/rs6000/rs6000.opt (mlra): New option.
4637 * lra-spills.c (lra_final_code_change): Remove useless move insns.
4639 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
4641 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
4642 has_single_use () and not do the conversion if has_single_use ()
4643 returns false for the multiplication result.
4645 2013-10-25 David Malcolm <dmalcolm@redhat.com>
4647 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
4648 to "NODE", since this works on a "tree", not an
4650 (CONSTANT_CLASS_P): Likewise.
4653 (INDIRECT_REF_P): Likewise.
4654 (REFERENCE_CLASS_P): Likewise.
4655 (COMPARISON_CLASS_P): Likewise.
4656 (UNARY_CLASS_P): Likewise.
4657 (BINARY_CLASS_P): Likewise.
4658 (STATEMENT_CLASS_P): Likewise.
4659 (VL_EXP_CLASS_P): Likewise.
4660 (EXPRESSION_CLASS_P): Likewise.
4661 (IS_TYPE_OR_DECL_P): Likewise.
4663 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
4665 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
4666 ADDR_EXPR in the defining statement.
4668 2013-10-25 Richard Biener <rguenther@suse.de>
4670 PR tree-optimization/58626
4671 * tree-loop-distribution.c (enum rdg_dep_type): Remove
4672 anti_dd, output_dd and input_dd.
4673 (struct rdg_edge): Remove level and relation members.
4674 (RDGE_LEVEL, RDGE_RELATION): Remove.
4675 (dot_rdg_1): Adjust.
4676 (create_rdg_edge_for_ddr): Remove.
4677 (create_rdg_edges_for_scalar): Adjust.
4678 (create_edge_for_control_dependence): Likewise.
4679 (create_rdg_edges): Split into ...
4680 (create_rdg_flow_edges): ... this
4681 (create_rdg_cd_edges): ... and this.
4683 (build_rdg): Likewise, do not compute data dependences or
4685 (pg_add_dependence_edges): New function.
4687 (distribute_loop): First apply all non-dependence based
4688 partition mergings. Then compute dependences between partitions
4689 and merge and order partitions according to them.
4691 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
4693 PR rtl-optimization/58831
4694 * alias.c (init_alias_analysis): At the beginning of each iteration,
4695 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
4697 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
4699 * recog.c (search_ofs): New static variable moved from...
4700 (peep2_find_free_register): ...here.
4701 (peephole2_optimize): Initialize it.
4703 2013-10-25 Tobias Burnus <burnus@net-b.de>
4705 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
4707 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
4709 * config/i386/i386.h (TARGET_MPX): New define.
4710 (TARGET_MPX_P): Ditto.
4712 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4714 * config/i386/constraints.md (B): New.
4717 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
4718 * config/i386/i386-modes.def (BND32): New.
4720 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
4721 * config/i386/i386.c (isa_opts): Add mmpx.
4722 (regclass_map): Add bound registers.
4723 (dbx_register_map): Likewise.
4724 (dbx64_register_map): Likewise.
4725 (svr4_dbx_register_map): Likewise.
4727 (ix86_option_override_internal): Support MPX ISA.
4728 (ix86_conditional_register_usage): Support bound registers.
4729 (print_reg): Likewise.
4730 (ix86_code_end): Add MPX bnd prefix.
4731 (output_set_got): Likewise.
4732 (ix86_output_call_insn): Likewise.
4733 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
4734 (ix86_print_operand_punct_valid_p): Likewise.
4735 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
4737 (ix86_class_likely_spilled_p): Add bound regs support.
4738 (ix86_hard_regno_mode_ok): Likewise.
4739 (x86_order_regs_for_local_alloc): Likewise.
4740 (ix86_bnd_prefixed_insn_p): New.
4741 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
4742 (FIXED_REGISTERS): Add bound registers.
4743 (CALL_USED_REGISTERS): Likewise.
4744 (REG_ALLOC_ORDER): Likewise.
4745 (HARD_REGNO_NREGS): Likewise.
4747 (VALID_BND_REG_MODE): New.
4748 (FIRST_BND_REG): New.
4749 (LAST_BND_REG): New.
4750 (reg_class): Add BND_REGS.
4751 (REG_CLASS_NAMES): Likewise.
4752 (REG_CLASS_CONTENTS): Likewise.
4754 (ANY_BND_REG_P): New.
4756 (HI_REGISTER_NAMES): Add bound registers.
4757 * config/i386/i386.md (UNSPEC_BNDMK): New.
4758 (UNSPEC_BNDMK_ADDR): New.
4759 (UNSPEC_BNDSTX): New.
4760 (UNSPEC_BNDLDX): New.
4761 (UNSPEC_BNDLDX_ADDR): New.
4762 (UNSPEC_BNDCL): New.
4763 (UNSPEC_BNDCU): New.
4764 (UNSPEC_BNDCN): New.
4765 (UNSPEC_MPX_FENCE): New.
4768 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
4769 (length_immediate): Likewise.
4770 (prefix_0f): Likewise.
4772 (prefix_rep): Check for bnd prefix.
4773 (length_nobnd): New.
4774 (length): Use length_nobnd if specified.
4779 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
4782 (simple_return_internal): Likewise.
4783 (simple_return_pop_internal): Likewise.
4784 (*indirect_jump): Add MPX bnd prefix.
4785 (*tablejump_1): Likewise.
4786 (simple_return_internal_long): Likewise.
4787 (simple_return_indirect_internal): Likewise.
4791 (*mov<mode>_internal_mpx): New.
4792 (<mode>_<bndcheck>): New.
4793 (*<mode>_<bndcheck>): New.
4798 * config/i386/predicates.md (lea_address_operand): Rename to...
4799 (address_no_seg_operand): ... this.
4800 (address_mpx_no_base_operand): New.
4801 (address_mpx_no_index_operand): New.
4802 (bnd_mem_operator): New.
4803 * config/i386/i386.opt (mmpx): New.
4804 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
4805 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
4807 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4809 * mode-classes.def (MODE_POINTER_BOUNDS): New.
4810 * tree.def (POINTER_BOUNDS_TYPE): New.
4811 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
4812 (POINTER_BOUNDS_MODE): New.
4813 (make_pointer_bounds_mode): New.
4814 * machmode.h (POINTER_BOUNDS_MODE_P): New.
4815 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
4816 (layout_type): Support POINTER_BOUNDS_TYPE.
4817 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
4818 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
4819 (type_contains_placeholder_1): Likewise.
4820 * tree.h (POINTER_BOUNDS_TYPE_P): New.
4821 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
4822 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
4824 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4826 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
4827 mode of address computation when calling memory_address_addr_space.
4829 2013-08-24 Richard Henderson <rth@twiddle.net>
4832 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
4833 instead of create_convert_operand_to.
4834 (maybe_emit_sync_lock_test_and_set): Likewise.
4835 (expand_atomic_compare_and_swap): Likewise.
4836 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
4838 2013-08-24 Sriraman Tallam <tmsriram@google.com>
4840 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
4841 Change param_is to use the struct and not the pointer to the struct.
4843 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
4845 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
4846 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4847 Move to targhooks.c.
4848 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
4849 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4850 Relocate from gimplify.c.
4851 * targhooks.h: Add 2 prototypes.
4852 * tree.h: Delete 2 prototypes.
4854 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4856 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
4857 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
4859 2013-10-24 Cong Hou <congh@google.com>
4861 * convert.c (convert_to_real): Guard those unsafe math function
4862 convertions with flag_unsafe_math_optimizations. Handle sqrt()
4865 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
4868 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
4870 (cgraph_create_edge): Use the new argument.
4871 (cgraph_create_indirect_edge): Likewise.
4873 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
4875 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
4877 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
4878 changes statep->state.
4880 2013-10-24 Tobias Burnus <burnus@net-b.de>
4883 * tree-cfg.c (replace_loop_annotate): New function.
4884 (execute_build_cfg): Call it.
4885 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
4886 * internal-fn.c (expand_ANNOTATE): New function.
4887 * internal-fn.def (ANNOTATE): Define as new internal function.
4888 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
4889 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
4890 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
4891 * doc/extend.texi (Pragmas): Document #pragma ivdep.
4892 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
4894 2013-10-17 Ian Bolton <ian.bolton@arm.com>
4895 Marcus Shawcroft <marcus.shawcroft@arm.com>
4897 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
4898 Special case reload SP+C into none GENERAL_REGS.
4900 2013-10-24 Michael Matz <matz@suse.de>
4902 * gengtype.c (is_file_equal): Check that files will be same length.
4904 2013-10-25 Christian Bruel <christian.bruel@st.com>
4906 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
4907 * config/sh/t-sh (sh-mem.o): New rule.
4908 * config/sh/sh-mem.cc (expand_block_move): Moved here.
4909 (sh_expand_cmpstr): New function.
4910 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
4911 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
4912 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
4913 (rotlhi3_8): Rename.
4915 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4917 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
4918 * Makefile.in (CXXDEPMODE): Assign and change users.
4919 (CCDEPMODE): Delete.
4920 * configure: Regenerate.
4922 2013-10-23 David Malcolm <dmalcolm@redhat.com>
4924 * gengtype-parse.c (require_without_advance): New.
4925 (type): For GTY-marked types that are not GTY((user)), parse any
4926 base classes, requiring them to be single-inheritance, and not
4927 be templates. For non-GTY-marked types and GTY((user)),
4928 continue to skip over any C++ inheritance specification.
4929 * gengtype-state.c (state_writer::write_state_struct_type):
4930 Write base class of type (if any).
4931 (read_state_struct_type): Read base class of type (if any).
4932 * gengtype.c (new_structure): Add a "base_class" parameter.
4933 (create_optional_field_): Update for new parameter to new_structure.
4934 (adjust_field_rtx_def): Likewise.
4935 (adjust_field_tree_exp): Likewise.
4936 * gengtype.h (struct type): Add "base_class" field to the s
4938 (new_structure): Add "base" parameter.
4940 2013-10-23 Sriraman Tallam <tmsriram@google.com>
4943 * config/i386/i386.c (ix86_option_override_internal):
4944 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
4945 (ix86_valid_target_attribute_tree):
4946 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
4947 Change TARGET_SSE to TARGET_SSE_P (opts->...)
4949 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4951 * tree-ssa-loop.h: Remove include files.
4952 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4953 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
4954 * graphite-clast-to-gimple.c: Likewise.
4955 * graphite-scop-detection.c: Likewise.
4956 * graphite-sese-to-poly.c: Likewise.
4957 * ipa-inline-analysis.c: Likewise.
4958 * ipa-pure-const.c: Likewise.
4959 * loop-init.c: Likewise.
4960 * passes.c: Likewise.
4961 * predict.c: Likewise.
4962 * tree-cfg.c: Likewise.
4963 * tree-cfgcleanup.c: Likewise.
4964 * tree-chrec.c: Likewise.
4965 * tree-data-ref.c: Likewise.
4966 * tree-loop-distribution.c: Likewise.
4967 * tree-parloops.c: Likewise.
4968 * tree-predcom.c: Likewise.
4969 * tree-scalar-evolution.c: Likewise.
4970 * tree-ssa-address.c: Likewise.
4971 * tree-ssa.c: Likewise.
4972 * tree-ssa-dce.c: Likewise.
4973 * tree-ssa-loop.c: Likewise.
4974 * tree-ssa-loop-im.c: Likewise.
4975 * tree-ssa-loop-ivcanon.c: Likewise.
4976 * tree-ssa-loop-ivopts.c: Likewise.
4977 * tree-ssa-loop-manip.c: Likewise.
4978 * tree-ssa-loop-niter.c: Likewise.
4979 * tree-ssa-loop-prefetch.c: Likewise.
4980 * tree-ssa-loop-unswitch.c: Likewise.
4981 * tree-ssa-reassoc.c: Likewise.
4982 * tree-vect-data-refs.c: Likewise.
4983 * tree-vect-loop.c: Likewise.
4984 * tree-vect-loop-manip.c: Likewise.
4985 * tree-vectorizer.c: Likewise.
4986 * tree-vect-stmts.c: Likewise.
4987 * tree-vrp.c: Likewise.
4989 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4991 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
4993 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4995 PR tree-optimization/58775
4996 PR tree-optimization/58791
4997 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
4998 (insert_stmt_after): Rewritten, don't move the stmt, but really
5000 (get_stmt_uid_with_default): Remove.
5001 (build_and_add_sum): Use insert_stmt_after and
5002 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
5003 (update_range_test): Set uid on stmts added by
5004 force_gimple_operand_gsi. Don't immediately modify statements
5005 in inter-bb optimization, just update oe->op values.
5006 (optimize_range_tests): Return bool whether any changed have been made.
5007 (update_ops): New function.
5008 (struct inter_bb_range_test_entry): New type.
5009 (maybe_optimize_range_tests): Perform statement changes here.
5010 (not_dominated_by, appears_later_in_bb, get_def_stmt,
5011 ensure_ops_are_available): Remove.
5012 (find_insert_point): Rewritten.
5013 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
5014 return LHS of the (new resp. old) stmt. Don't call
5015 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
5016 instead of last, move new stmt at the right place.
5017 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
5018 (negate_value): Likewise. Set uids.
5019 (break_up_subtract_bb): Initialize uids.
5020 (reassociate_bb): Adjust rewrite_expr_tree caller.
5021 (do_reassoc): Don't call renumber_gimple_stmt_uids.
5023 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
5026 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
5027 TARGET_32BIT final condition.
5028 (mulsi3_internal2 and splitter): Same.
5030 2013-10-23 Jeff Law <law@redhat.com>
5032 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
5033 through joiner blocks with abnormal outgoing edges.
5035 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
5036 Add parameter JOINERS, to allow/disallow threading through joiner
5038 (thread_block): New. Call thread_block_1.
5039 (mark_threaded_blocks): Remove code to filter out certain cases
5040 of threading through joiner blocks.
5041 (thread_through_all_blocks): Document how we can have a dangling
5042 edge AUX field and clear it.
5044 2013-10-23 Ian Lance Taylor <iant@google.com>
5046 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
5047 (C++ Dialect Options): Likewise.
5048 (Optimize Options): Likewise.
5050 2013-10-23 Tom de Vries <tom@codesourcery.com>
5052 PR tree-optimization/58805
5053 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
5055 2013-10-23 Jakub Jelinek <jakub@redhat.com>
5057 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
5058 sequence based on get_range_info returned range.
5060 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
5062 * tree-ssa.h: Remove all #include's
5063 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
5064 * alias.c: Move required includes from tree-ssa.h.
5066 * builtins.c: Likewise.
5067 * calls.c: Likewise.
5068 * cfgexpand.c: Likewise.
5069 * cfghooks.c: Likewise.
5070 * cfgloop.c: Likewise.
5071 * cfgloopmanip.c: Likewise.
5072 * cgraph.c: Likewise.
5073 * cgraphbuild.c: Likewise.
5074 * cgraphclones.c: Likewise.
5075 * cgraphunit.c: Likewise.
5077 * except.c: Likewise.
5079 * final.c: Likewise.
5080 * fold-const.c: Likewise.
5081 * ggc-page.c: Likewise.
5082 * gimple-builder.c: Likewise.
5083 * gimple-fold.c: Likewise.
5084 * gimple-iterator.c: Likewise.
5085 * gimple-low.c: Likewise.
5086 * gimple-pretty-print.c: Likewise.
5087 * gimple-ssa-strength-reduction.c: Likewise.
5088 * gimple-streamer-in.c: Likewise.
5089 * gimple-streamer-out.c: Likewise.
5090 * gimplify.c: Likewise.
5091 * graphite-blocking.c: Likewise.
5092 * graphite-clast-to-gimple.c: Likewise.
5093 * graphite-dependences.c: Likewise.
5094 * graphite-interchange.c: Likewise.
5095 * graphite-optimize-isl.c: Likewise.
5096 * graphite-poly.c: Likewise.
5097 * graphite-scop-detection.c: Likewise.
5098 * graphite-sese-to-poly.c: Likewise.
5099 * graphite.c: Likewise.
5100 * ipa-cp.c: Likewise.
5101 * ipa-inline-analysis.c: Likewise.
5102 * ipa-inline-transform.c: Likewise.
5103 * ipa-inline.c: Likewise.
5104 * ipa-prop.c: Likewise.
5105 * ipa-pure-const.c: Likewise.
5106 * ipa-reference.c: Likewise.
5107 * ipa-split.c: Likewise.
5108 * ipa-utils.c: Likewise.
5109 * loop-init.c: Likewise.
5110 * lto-cgraph.c: Likewise.
5111 * lto-section-in.c: Likewise.
5112 * lto-section-out.c: Likewise.
5113 * lto-streamer-in.c: Likewise.
5114 * lto-streamer-out.c: Likewise.
5115 * lto-streamer.c: Likewise.
5116 * omp-low.c: Likewise.
5117 * passes.c: Likewise.
5118 * predict.c: Likewise.
5119 * print-tree.c: Likewise.
5120 * profile.c: Likewise.
5122 * targhooks.c: Likewise.
5123 * tracer.c: Likewise.
5124 * trans-mem.c: Likewise.
5125 * tree-call-cdce.c: Likewise.
5126 * tree-cfg.c: Likewise.
5127 * tree-cfgcleanup.c: Likewise.
5128 * tree-chrec.c: Likewise.
5129 * tree-complex.c: Likewise.
5130 * tree-data-ref.c: Likewise.
5131 * tree-dfa.c: Likewise.
5132 * tree-eh.c: Likewise.
5133 * tree-emutls.c: Likewise.
5134 * tree-if-conv.c: Likewise.
5135 * tree-inline.c: Likewise.
5136 * tree-into-ssa.c: Likewise.
5137 * tree-loop-distribution.c: Likewise.
5138 * tree-mudflap.c: Likewise.
5139 * tree-nested.c: Likewise.
5140 * tree-nrv.c: Likewise.
5141 * tree-object-size.c: Likewise.
5142 * tree-outof-ssa.c: Likewise.
5143 * tree-parloops.c: Likewise.
5144 * tree-phinodes.c: Likewise.
5145 * tree-predcom.c: Likewise.
5146 * tree-pretty-print.c: Likewise.
5147 * tree-profile.c: Likewise.
5148 * tree-scalar-evolution.c: Likewise.
5149 * tree-sra.c: Likewise.
5150 * tree-ssa-address.c: Likewise.
5151 * tree-ssa-alias.c: Likewise.
5152 * tree-ssa-ccp.c: Likewise.
5153 * tree-ssa-coalesce.c: Likewise.
5154 * tree-ssa-copy.c: Likewise.
5155 * tree-ssa-copyrename.c: Likewise.
5156 * tree-ssa-dce.c: Likewise.
5157 * tree-ssa-dom.c: Likewise.
5158 * tree-ssa-dse.c: Likewise.
5159 * tree-ssa-forwprop.c: Likewise.
5160 * tree-ssa-ifcombine.c: Likewise.
5161 * tree-ssa-live.c: Likewise.
5162 * tree-ssa-loop-ch.c: Likewise.
5163 * tree-ssa-loop-im.c: Likewise.
5164 * tree-ssa-loop-ivcanon.c: Likewise.
5165 * tree-ssa-loop-ivopts.c: Likewise.
5166 * tree-ssa-loop-manip.c: Likewise.
5167 * tree-ssa-loop-niter.c: Likewise.
5168 * tree-ssa-loop-prefetch.c: Likewise.
5169 * tree-ssa-loop-unswitch.c: Likewise.
5170 * tree-ssa-loop.c: Likewise.
5171 * tree-ssa-math-opts.c: Likewise.
5172 * tree-ssa-operands.c: Likewise.
5173 * tree-ssa-phiopt.c: Likewise.
5174 * tree-ssa-phiprop.c: Likewise.
5175 * tree-ssa-pre.c: Likewise.
5176 * tree-ssa-propagate.c: Likewise.
5177 * tree-ssa-reassoc.c: Likewise.
5178 * tree-ssa-sccvn.c: Likewise.
5179 * tree-ssa-sink.c: Likewise.
5180 * tree-ssa-strlen.c: Likewise.
5181 * tree-ssa-structalias.c: Likewise.
5182 * tree-ssa-tail-merge.c: Likewise.
5183 * tree-ssa-ter.c: Likewise.
5184 * tree-ssa-threadedge.c: Likewise.
5185 * tree-ssa-threadupdate.c: Likewise.
5186 * tree-ssa-uncprop.c: Likewise.
5187 * tree-ssa-uninit.c: Likewise.
5188 * tree-ssa.c: Likewise.
5189 * tree-ssanames.c: Likewise.
5190 * tree-stdarg.c: Likewise.
5191 * tree-streamer-in.c: Likewise.
5192 * tree-switch-conversion.c: Likewise.
5193 * tree-tailcall.c: Likewise.
5194 * tree-vect-data-refs.c: Likewise.
5195 * tree-vect-generic.c: Likewise.
5196 * tree-vect-loop-manip.c: Likewise.
5197 * tree-vect-loop.c: Likewise.
5198 * tree-vect-patterns.c: Likewise.
5199 * tree-vect-slp.c: Likewise.
5200 * tree-vect-stmts.c: Likewise.
5201 * tree-vectorizer.c: Likewise.
5202 * tree-vrp.c: Likewise.
5205 * value-prof.c: Likewise.
5206 * var-tracking.c: Likewise.
5207 * varpool.c: Likewise.
5208 * vtable-verify.c: Likewise.
5210 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5212 * config/tilegx/tilegx.c: Include "tree.h".
5214 2013-10-23 Jakub Jelinek <jakub@redhat.com>
5216 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
5217 the info, not after it.
5218 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
5219 dump_ssaname_info call.
5220 (pp_gimple_stmt_1): Adjust caller.
5221 (dump_phi_nodes): Likewise. Don't print "# " here.
5223 2013-10-22 Jan Hubicka <jh@suse.cz>
5225 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
5227 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
5228 * i386.c (expand_small_movmem_or_setmem): New function.
5229 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
5231 (alg_usable_p): Add support for value ranges; cleanup.
5232 (ix86_expand_set_or_movmem): Add support for misaligned moves.
5234 2013-10-22 Sterling Augustine <saugustine@google.com>
5236 * doc/invoke.texi: Document -ggnu-pubnames.
5237 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
5239 * dwarf2out.c: Include gdb/gdb-index.h.
5240 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
5241 debug_generate_pub_sections.
5242 (is_java, output_pubtables, output_pubname): New functions.
5243 (include_pubname_in_output): Handle debug_generate_pub_sections at
5245 (size_of_pubnames): Use new local space_for_flags based on
5246 debug_generate_pub_sections.
5247 (output_pubnames): Unify pubnames and pubtypes output logic.
5248 Genericize comments. Call output_pubname.
5249 (dwarf2out_finish): Move logic to output_pubnames and call it.
5251 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
5254 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
5255 Remove CCCmode handling.
5256 <case LTU>: Return 'c' suffix for CCCmode.
5257 <case GEU>: Return 'nc' suffix for CCCmode.
5258 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
5259 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
5260 (*sub<mode>3_cc_overflow): Ditto.
5261 (*subsi3_zext_cc_overflow): Ditto.
5263 2013-10-22 Steve Ellcey <sellcey@mips.com>
5265 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
5266 (AND (NOT OP1) (NOT OP2)).
5268 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
5270 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
5271 meaning of merge-high and merge-low masks for little endian; avoid
5272 use of vector-pack masks for little endian for mismatched modes.
5274 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5276 * config/tilepro/tilepro.c: Include "tree.h".
5278 2013-10-22 Andreas Schwab <schwab@suse.de>
5280 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
5283 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
5285 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
5286 [Ed Smith-Rowland]): New entries.
5287 ([Stephen M. Webb]): Update.
5289 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
5291 * tree-ssa-ter.h: Remove duplicate copy of file contents.
5293 2013-10-21 Marek Polacek <polacek@redhat.com>
5296 * fold-const.c (fold_range_test): Return 0 if the type is not
5299 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
5301 * system.h: Move hwint.h include further down.
5302 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
5303 gcc_checking_asserts.
5304 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
5306 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
5308 Fix volatile issues in optimize_bit_field_compare.
5309 * fold-const.c (optimize_bit_field_compare): Bail out if
5310 lvolatilep or rvolatilep.
5312 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
5314 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
5315 and get_inner_reference returning different pmode for non-volatile
5316 bit-field members dependent on flag_strict_volatile_bitfields.
5317 * stor-layout.c (layout_decl): Remove special handling of
5318 flag_strict_volatile_bitfields.
5319 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
5320 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
5322 2013-10-21 Paulo Matos <pmatos@broadcom.com>
5324 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
5327 2013-10-21 Jeff Law <law@redhat.com>
5329 * tree-ssa-threadedge.c (thread_through_normal_block): New
5330 argument VISITED. Remove VISISTED as a local variable. When we
5331 have a threadable jump, verify the destination of the jump has not
5333 (thread_across_edge): Allocate VISITED bitmap once at function
5334 scope and use it throughout. Make sure to set appropriate bits in
5335 VISITED for E (start of jump thread path).
5336 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
5337 through a joiner if any edge on the path has a recorded jump thread.
5339 2013-10-21 Ian Lance Taylor <iant@google.com>
5341 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
5342 don't imply that attributes can solve all problems.
5343 (Directory Options): Fix typo.
5345 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5347 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
5348 extend and extend_add.
5350 2013-10-21 Richard Biener <rguenther@suse.de>
5352 PR tree-optimization/58794
5353 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
5354 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
5356 2013-10-21 Richard Biener <rguenther@suse.de>
5359 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
5360 to (T) X for sign-changing conversions (or no conversion).
5362 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
5364 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
5366 2013-10-20 Jan Hubicka <jh@suse.cz>
5368 * config/i386/i386-tune.def: Add comment; organize into categories
5370 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5372 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
5373 argument. Update function comment.
5374 (expand_set_or_movmem_via_rep): New function combining
5375 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
5376 (expand_movmem_via_rep_mov): Remove.
5377 expand_setmem_via_rep_stos): Remove.
5378 (expand_movmem_epilogue): Update calls correspondingly.
5379 (expand_setmem_epilogue_via_loop): Likewise.
5381 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
5382 (expand_set_or_movmem_prologue): New function combining
5383 expand_movmem_prologue and expand_setmem_prologue.
5384 (expand_movmem_prologue): Remove.
5385 (expand_setmem_prologue): Remove.
5386 (expand_set_or_movmem_constant_prologue): New function combining
5387 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
5388 (expand_constant_movmem_prologue): Remove.
5389 (expand_constant_setmem_prologue): Remove.
5390 (promote_duplicated_reg): Allow vector-const0 value.
5391 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
5392 and ix86_expand_setmem.
5393 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
5394 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
5396 2013-10-21 Diego Novillo <dnovillo@google.com>
5398 * asan.c: Include tree.h
5399 * bb-reorder.c: Likewise.
5400 * cfgcleanup.c: Likewise.
5401 * cfgloopmanip.c: Likewise.
5402 * data-streamer-in.c: Likewise.
5403 * data-streamer-out.c: Likewise.
5404 * data-streamer.c: Likewise.
5405 * dwarf2cfi.c: Likewise.
5406 * graphite-blocking.c: Likewise.
5407 * graphite-clast-to-gimple.c: Likewise.
5408 * graphite-dependences.c: Likewise.
5409 * graphite-interchange.c: Likewise.
5410 * graphite-optimize-isl.c: Likewise.
5411 * graphite-poly.c: Likewise.
5412 * graphite-scop-detection.c: Likewise.
5413 * graphite-sese-to-poly.c: Likewise.
5414 * graphite.c: Likewise.
5415 * ipa-devirt.c: Likewise.
5416 * ipa-profile.c: Likewise.
5419 * loop-init.c: Likewise.
5420 * loop-unroll.c: Likewise.
5421 * lower-subreg.c: Likewise.
5422 * lto/lto-object.c: Likewise.
5423 * recog.c: Likewise.
5424 * reginfo.c: Likewise.
5425 * tree-loop-distribution.c: Likewise.
5426 * tree-parloops.c: Likewise.
5427 * tree-ssa-strlen.c: Likewise.
5428 * tree-streamer.c: Likewise.
5429 * value-prof.c: Likewise.
5430 * target-globals.c: Likewise.
5431 * expr.h: Include tree-core.h instead of tree.h.
5432 * gimple.h: Likewise.
5433 * ipa-prop.h: Likewise.
5434 * ipa-utils.h: Likewise.
5435 * lto-streamer.h: Likewise.
5436 * streamer-hooks.h: Likewise.
5437 * ipa-reference.h: Include cgraph.h instead of tree.h.
5438 * cgraph.h: Include basic-block.h instead of tree.h.
5439 * tree-streamer.h: Do not include tree.h.
5440 * genattrtab.c (write_header): Generate inclusion of tree.h.
5441 * genautomata.c (main): Likewise.
5442 * genemit.c: Likewise.
5443 * genopinit.c: Likewise.
5444 * genoutput.c (output_prologue): Likewise.
5445 * genpeep.c: Likewise.
5447 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5449 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
5451 (vec_unpacku_hi_v8hi): Likewise.
5452 (vec_unpacku_lo_v16qi): Likewise.
5453 (vec_unpacku_lo_v8hi): Likewise.
5455 2013-10-20 Jan Hubicka <jh@suse.cz>
5457 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
5458 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
5459 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
5461 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
5463 * config/mips/mips.h (ISA_HAS_WSBH): Define.
5464 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
5466 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
5468 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
5471 * system.h: Undef m_slot.
5473 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5475 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
5476 all elements for both endian flavors.
5478 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
5481 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
5482 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
5483 and SI_REG for 64bit SYSV ABI targets.
5485 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
5487 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
5488 to multi_reg_return. Clarify that we are skipping USEs of multiple
5489 return registers. Use bool type where appropriate.
5491 2013-10-18 Jan Hubicka <jh@suse.cz>
5493 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
5495 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
5496 (X86_TUNE_PUSH_MEMORY): Likewise.
5497 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
5498 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
5499 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
5501 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
5502 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
5504 (ix86_option_override_internal): Update to use tune features instead
5507 2013-10-18 Cong Hou <congh@google.com>
5509 PR tree-optimization/58508
5510 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
5511 statement that contains data refs with zero-step.
5513 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5515 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
5516 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
5517 * gimple-low.c (gimple_check_call_arg,
5518 gimple_check_call_matching_types): Move to cgraph.c.
5519 * gimple-low.h: Remove prototype.
5520 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
5521 Relocate from gimple-low.c.
5522 * cgraph.h: Add prototype. Don't include basic-block.h.
5523 * gimplify.c: Add gimple-low to include list.
5524 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
5525 * tree-eh.c: Add gimple-low to include list.
5526 * tree-nested.c: Likewise.
5527 * cfgexpand.c: Add tree-ssa-address.h to include list.
5529 * gimple-fold.c: Likewise.
5530 * gimple-ssa-strength-reduction.c: Likewise.
5531 * trans-mem.c: Likewise.
5532 * tree-mudflap.c: Likewise.
5533 * tree-ssa-loop-ivopts.c: Likewise.
5534 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
5535 (degenerate_phi_result): Move to tree-phinodes.c.
5536 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
5537 * tree-ssa-threadedge.c: Likewise.
5538 * tree-vrp.c: Likewise.
5539 * tree-phinodes.c (degenerate_phi_result): Relocate here.
5540 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
5541 * tree-phinodes.h (degenerate_phi_result): Add prototype.
5542 * tree-ssa-copy.c: Include tree-ssa-dom.h.
5543 * tree-ssa-forwprop.c: Likewise.
5544 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
5545 pass_data_cleanup_cfg_post_optimizing,
5546 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
5547 * tree-optimize.c: Delete File.
5548 * graphite.c: Include tree-cfgcleanup.h.
5549 * passes.c: Likewise.
5550 * tree-cfg.c: Likewise.
5551 * tree-profile.c: Likewise.
5552 * tree-ssa-dse.c: Likewise.
5553 * tree-ssa-loop-ivcanon.c: Likewise.
5554 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
5555 * tree-outof-ssa.c: Include sbitmap.h.
5556 * tree-ssa-live.c: Likewise.
5557 * tree-ssa-propagate.c: Likewise.
5558 * tree-ssa-structalias.c: Likewise.
5559 * tree-stdarg.c: Likewise.
5560 * Makefile.in (OBJS): Delete tree-optimize.o.
5561 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
5562 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
5563 * varasm.c: Include basic-block.h.
5564 * cfgloop.h: Include function.h instead of basic-block.h
5565 (bb_loop_depth): Move to cfgloop.c.
5566 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
5568 2013-10-18 Teresa Johnson <tejohnson@google.com>
5570 * predict.c (probably_never_executed): Compare frequency-based
5571 count to number of training runs.
5572 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
5574 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5576 * config/arm/arm.c (cortexa9_extra_costs): New table.
5577 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
5579 2013-10-18 Jeff Law <law@redhat.com>
5581 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
5583 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
5584 (register_jump_thread): Add "registered_jump_thread" debug
5586 * dbgcnt.def (registered_jump_thread): New debug counter.
5588 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5590 * config/rs6000/rs6000.c: Include cgraph.h.
5592 2013-10-18 Teresa Johnson <tejohnson@google.com>
5594 * tree-ssa-tail-merge.c (replace_block_by): Update edge
5595 weights during merging.
5597 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5599 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
5600 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
5601 * tree-ssa-operands.h: Remove prototype.
5602 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
5603 * gimple.c (virtual_operand_p): Relocate from gimple.c.
5604 * gimple.h: Add prototype.
5605 * gimple-ssa.h: Include tree-ssa-operands.h.
5606 * tree-dump.c: Add tree-cfg.h to include list.
5607 * tree-ssa-alias.c: Add ipa-reference.h to include list.
5608 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
5609 * config/i386/i386.c: Don't include tree-flow.h.
5610 * config/rs6000/rs6000.c: Likewise.
5612 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5614 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
5616 2013-10-18 Richard Biener <rguenther@suse.de>
5618 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
5619 or TYPE_UNSIGNED of integral types.
5620 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
5621 NULL_TREE for zero-precision integral types.
5623 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
5625 * config/aarch64/arm_neon.h
5626 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
5628 2013-10-17 Sriraman Tallam <tmsriram@google.com>
5631 * opth-gen.awk: Define target_flags_explicit.
5633 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5635 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
5636 fields to the reg_addr array that describes the valid addressing
5637 mode for any register, general purpose registers, floating point
5638 registers, and Altivec registers.
5639 (FIRST_RELOAD_REG_CLASS): Likewise.
5640 (LAST_RELOAD_REG_CLASS): Likewise.
5641 (struct reload_reg_map_type): Likewise.
5642 (reload_reg_map_type): Likewise.
5643 (RELOAD_REG_VALID): Likewise.
5644 (RELOAD_REG_MULTIPLE): Likewise.
5645 (RELOAD_REG_INDEXED): Likewise.
5646 (RELOAD_REG_OFFSET): Likewise.
5647 (RELOAD_REG_PRE_INCDEC): Likewise.
5648 (RELOAD_REG_PRE_MODIFY): Likewise.
5649 (reg_addr): Likewise.
5650 (mode_supports_pre_incdec_p): New helper functions to say whether
5651 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
5652 (mode_supports_pre_modify_p): Likewise.
5653 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
5654 print the valid address mode bits for each mode.
5655 (rs6000_debug_print_mode): Likewise.
5656 (rs6000_debug_reg_global): Likewise.
5657 (rs6000_setup_reg_addr_masks): New function to set up the address
5658 mask bits for each type.
5659 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
5660 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
5661 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
5662 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
5663 PRE_MODIFY are supported.
5664 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
5665 registers, instead of {src,dest}_av_p.
5666 (rs6000_print_options_internal): Tweak the debug output slightly.
5668 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
5670 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
5673 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
5675 * tree-flow.h (struct omp_region): Move to omp-low.c.
5676 Remove omp_ prototypes and variables.
5677 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
5678 (copy_var_decl): Relocate prototype from tree-flow.h.
5679 * gimple.c (copy_var_decl): Relocate from omp-low.c.
5680 * tree.h: Move prototype to omp-low.h.
5681 * omp-low.h: New File. Relocate prototypes here.
5682 * omp-low.c (struct omp_region): Make local here.
5683 (root_omp_region): Make static.
5684 (copy_var_decl) Move to gimple.c.
5685 (new_omp_region): Make static.
5686 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
5687 * tree-cfg.c: Include omp-low.h.
5688 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
5689 * gimplify.c: Include omp-low.h.
5690 * tree-parloops.c: Likewise.
5692 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
5694 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
5695 operands are in memory, prefer to force non-matched operand 1 to
5698 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5701 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
5702 restrict TImode addresses to single indirect registers if both
5703 -mquad-memory and -mvsx-timode are used.
5704 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
5705 we should emit load/store quad. Remove using %y for quad memory
5708 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
5709 constraints to allow load/store quad on machines where TImode is
5710 not allowed in VSX registers. Use 'n' instead of 'F' constraint
5711 for TImode to load integer constants.
5713 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5715 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
5717 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
5719 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
5720 handling of STACK_REG.
5722 2013-10-17 Richard Biener <rguenther@suse.de>
5724 PR tree-optimization/58143
5725 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
5727 (rewrite_to_defined_overflow): Likewise.
5728 (move_computations_dom_walker::before_dom): Rewrite stmts
5729 with undefined signed overflow that are not always executed
5730 into unsigned arithmetic.
5732 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5735 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
5736 explicit isa flag to be an options variable, instead of using
5737 global_options_set. Remove define from rs6000.h.
5738 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
5740 * config/rs6000/rs6000.c (rs6000_option_override_internal):
5741 Initialize rs6000_isa_flags_explicit.
5742 (rs6000_function_specific_save): Add gcc_options* parameter, so
5743 that the powerpc builds after the 2013-10-15 changes.
5744 (rs6000_function_specific_restore): Likewise.
5746 2013-10-16 DJ Delorie <dj@redhat.com>
5748 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
5749 op is a REG before checking REGNO.
5750 (rl78_alloc_physical_registers): Verify pattern is a SET before
5753 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5755 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
5757 (vec_unpacks_lo_v4sf): Likewise.
5758 (vec_unpacks_float_hi_v4si): Likewise.
5759 (vec_unpacks_float_lo_v4si): Likewise.
5760 (vec_unpacku_float_hi_v4si): Likewise.
5761 (vec_unpacku_float_lo_v4si): Likewise.
5763 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5765 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
5766 (vsx_concat_v2sf): Likewise.
5768 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
5770 * config/aarch64/aarch64.md
5771 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
5773 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
5775 PR tree-optimization/58697
5776 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
5777 estimated_loop_iterations_int.
5778 (max_stmt_executions_int): Call get_max_loop_iterations_int.
5779 (get_max_loop_iterations_int): New. HWINT version of
5780 get_max_loop_iterations.
5781 * cfgloop.h: Add prototypes.
5782 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
5783 * loop-unroll.c (decide_peel_once_rolling): Call
5784 get_estimated_loop_iterations_int.
5785 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
5786 * tree-ssa-loop-niter.h: Tweak prototypes.
5788 2013-10-16 David Malcolm <dmalcolm@redhat.com>
5790 * gengtype-parse.c (struct_field_seq): Ignore access-control
5791 keywords ("public:" etc).
5793 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
5795 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
5796 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
5798 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
5800 * config/arm/arm.opt (mlra): New option.
5801 * config/arm/arm.c (arm_lra_p): New function.
5802 (TARGET_LRA_P): Define.
5804 2013-10-16 Paulo Matos <pmatos@broadcom.com>
5806 * tree-core.h (tree_code_name): Remove.
5807 * tree.h (get_tree_code_name): New prototype.
5808 * tree.c (tree_code_name): Make static.
5809 (get_tree_code_name): New function.
5810 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
5811 tree_class_check_failed, tree_range_check_failed,
5812 tree_not_class_check_failed, omp_clause_check_failed,
5813 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
5814 wrapper get_tree_code_name instead of calling tree_code_name directly.
5815 * tree-vrp.c (dump_asserts_for): Likewise.
5816 * tree-dump.c (dequeue_and_dump): Likewise.
5817 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
5818 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
5819 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
5820 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
5821 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
5822 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
5823 dump_gimple_omp_for): Likewise.
5824 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
5825 * tree-ssa-pre.c (print_pre_expr): Likewise.
5826 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
5827 * print-tree.c (print_node_brief, print_node): Likewise.
5828 * gimple.c (gimple_check_failed): Likewise.
5829 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
5830 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
5831 * config/mep/mep.c (mep_validate_vliw): Likewise.
5832 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
5833 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
5835 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
5837 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
5840 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
5842 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
5843 (MULTILIB_MATCHES): Add multilib for -march=v8.
5845 2013-10-15 Sriraman Tallam <tmsriram@google.com>
5848 * optc-save-gen.awk: Add extra parameter to the save and restore
5850 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
5851 * tree.c (build_optimization_node): New parameter. Add extra parameter
5852 to call to cl_optimization_save.
5853 (build_target_option_node): New parameter. Add extra parameter
5854 to call to cl_target_option_save.
5855 * tree.h (build_optimization_node): New parameter.
5856 (build_target_option_node): New parameter.
5857 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
5858 build_optimization_node and build_target_option_node.
5859 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
5860 (handle_pragma_push_options): Ditto.
5861 * toplev.c (process_options): Ditto.
5862 * opts.c (init_options_struct): Check for opts_set non-null.
5863 * target.def (target_option.save): New parameter.
5864 (target_option.restore): New parameter.
5865 * tm.texi: Generate.
5866 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
5867 (ix86_pragma_target_parse): Ditto.
5868 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
5870 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
5871 to build_optimization_node and build_target_option_node.
5872 (rs6000_valid_attribute_p): Ditto.
5873 (rs6000_pragma_target_parse): Ditto.
5874 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
5876 * config/i386/i386.h:
5877 TARGET_64BIT_P: New Macro
5878 TARGET_MMX_P: New Macro.
5879 TARGET_3DNOW_P: New Macro.
5880 TARGET_3DNOW_A_P: New Macro.
5881 TARGET_SSE_P: New Macro.
5882 TARGET_SSE2_P: New Macro.
5883 TARGET_SSE3_P: New Macro.
5884 TARGET_SSSE3_P: New Macro.
5885 TARGET_SSE4_1_P: New Macro.
5886 TARGET_SSE4_2_P: New Macro.
5887 TARGET_AVX_P: New Macro.
5888 TARGET_AVX2_P: New Macro.
5889 TARGET_AVX512F_P: New Macro.
5890 TARGET_AVX512PF_P: New Macro.
5891 TARGET_AVX512ER_P: New Macro.
5892 TARGET_AVX512CD_P: New Macro.
5893 TARGET_FMA_P: New Macro.
5894 TARGET_SSE4A_P: New Macro.
5895 TARGET_FMA4_P: New Macro.
5896 TARGET_XOP_P: New Macro.
5897 TARGET_LWP_P: New Macro.
5898 TARGET_ABM_P: New Macro.
5899 TARGET_BMI_P: New Macro.
5900 TARGET_BMI2_P: New Macro.
5901 TARGET_LZCNT_P: New Macro.
5902 TARGET_TBM_P: New Macro.
5903 TARGET_POPCNT_P: New Macro.
5904 TARGET_SAHF_P: New Macro.
5905 TARGET_MOVBE_P: New Macro.
5906 TARGET_CRC32_P: New Macro.
5907 TARGET_AES_P: New Macro.
5908 TARGET_PCLMUL_P: New Macro.
5909 TARGET_CMPXCHG16B_P: New Macro.
5910 TARGET_FSGSBASE_P: New Macro.
5911 TARGET_RDRND_P: New Macro.
5912 TARGET_F16C_P: New Macro.
5913 TARGET_RTM_P: New Macro.
5914 TARGET_HLE_P: New Macro.
5915 TARGET_RDSEED_P: New Macro.
5916 TARGET_PRFCHW_P: New Macro.
5917 TARGET_ADX_P: New Macro.
5918 TARGET_FXSR_P: New Macro.
5919 TARGET_XSAVE_P: New Macro.
5920 TARGET_XSAVEOPT_P: New Macro.
5921 TARGET_LP64_P: New Macro.
5922 TARGET_X32_P: New Macro.
5923 TARGET_FPMATH_DEFAULT_P: New Macro.
5924 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
5925 * config/i386/i386.c (ix86_option_override_internal): New parameters.
5927 Change ix86_tune_string to access opts->x_ix86_tune_string.
5928 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
5929 Change ix86_arch_string to access opts->x_ix86_arch_string.
5930 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
5931 Change ix86_pmode to access opts->x_ix86_pmode.
5932 Change ix86_abi to access opts->x_ix86_abi.
5933 Change ix86_cmodel to access opts->x_ix86_cmodel.
5934 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
5935 Change ix86_isa_flags_explicit to access
5936 opts->x_ix86_isa_flags_explicit.
5937 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
5938 Change ix86_regparm to access opts->x_ix86_regparm.
5939 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
5940 Change ix86_preferred_stack_boundary_arg to access
5941 opts->x_ix86_preferred_stack_boundary_arg.
5942 Change ix86_force_align_arg_pointer to access
5943 opts->x_ix86_force_align_arg_pointer.
5944 Change ix86_incoming_stack_boundar_arg to access
5945 opts->x_ix86_incoming_stack_boundar_arg.
5946 Change ix86_fpmath to access opts->x_ix86_fpmath.
5947 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
5948 Change ix86_recip_name to access opts->x_ix86_recip_name.
5949 Change ix86_stack_protector_guard to access
5950 opts->x_ix86_stack_protector_guard.
5951 Change ix86_tune_memcpy_strategy to access
5952 opts->x_ix86_tune_memcpy_strategy.
5953 Change ix86_tune_memset_strategy to access
5954 opts->x_ix86_tune_memset_strategy.
5955 Change global_options to access opts.
5956 Change global_options_set to access opts_set.
5957 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
5958 Change TARGET_MMX to TARGET_MMX_P (opts->...).
5959 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
5960 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
5961 Change TARGET_SSE to TARGET_SSE_P (opts->...).
5962 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
5963 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
5964 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
5965 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
5966 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
5967 Change TARGET_AVX to TARGET_AVX_P (opts->...).
5968 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
5969 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
5970 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
5971 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
5972 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
5973 Change TARGET_FMA to TARGET_FMA_P (opts->...).
5974 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
5975 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
5976 Change TARGET_XOP to TARGET_XOP_P (opts->...).
5977 Change TARGET_LWP to TARGET_LWP_P (opts->...).
5978 Change TARGET_ABM to TARGET_ABM_P (opts->...).
5979 Change TARGET_BMI to TARGET_BMI_P (opts->...).
5980 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
5981 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
5982 Change TARGET_TBM to TARGET_TBM_P (opts->...).
5983 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
5984 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
5985 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
5986 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
5987 Change TARGET_AES to TARGET_AES_P (opts->...).
5988 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
5989 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
5990 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
5991 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
5992 Change TARGET_F16C to TARGET_F16C_P (opts->...).
5993 Change TARGET_RTM to TARGET_RTM_P (opts->...).
5994 Change TARGET_HLE to TARGET_HLE_P (opts->...).
5995 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
5996 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
5997 Change TARGET_ADX to TARGET_ADX_P (opts->...).
5998 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
5999 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
6000 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
6001 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
6002 Change TARGET_X32 to TARGET_X32_P (opts->...).
6003 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
6004 Change TARGET_FLOAT_RETURNS_IN_80387 to
6005 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
6006 (ix86_function_specific_save): New parameter. Use opts-> fields
6007 to replace global fields.
6008 (ix86_function_specific_restore): Ditto.
6009 (ix86_valid_target_attribute_inner_p): New parameters.
6011 Fix call to ix86_handle_option and set_option.
6012 (ix86_valid_target_attribute_tree): New parameters.
6013 Change global_options to access opts.
6014 Change global_options_set to access opts_set.
6015 Fix call to ix86_valid_target_attribute_inner_p.
6016 Change ix86_tune_string to access opts->x_ix86_tune_string.
6017 Change ix86_arch_string to access opts->x_ix86_arch_string.
6018 Change ix86_fpmath to access opts->x_ix86_fpmath
6019 Fix call to ix86_option_override_internal.
6020 Fix call to ix86_add_new_builtins.
6021 Fix calls to build_optimization_node and build_target_option_node.
6022 (ix86_valid_target_attribute_p): Remove access to global_options.
6023 Use new gcc_options structure func_options.
6024 Fix call to ix86_valid_target_attribute_tree.
6025 Fix call to build_optimization_node.
6026 (get_builtin_code_for_version): Fix call to
6027 ix86_valid_target_attribute_tree.
6029 2013-10-15 David Malcolm <dmalcolm@redhat.com>
6031 * Makefile.in (PICFLAG): New.
6032 (enable_host_shared): New.
6033 (INTERNAL_CFLAGS): Use PICFLAG.
6034 (LIBIBERTY): Use pic build of libiberty.a if configured with
6035 --enable-host-shared.
6036 * configure.ac: Add --enable-host-shared, setting up new
6038 * configure: Regenerate.
6039 * doc/install.texi (--enable-shared): Add note contrasting it with ...
6040 (--enable-host-shared): New option.
6042 2013-10-15 Richard Biener <rguenther@suse.de>
6044 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
6045 for built-in functions.
6047 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
6049 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
6050 (optimize_range_tests_1): New function,
6051 extracted from optimize_range_tests.
6052 (optimize_range_tests_xor): Similarly.
6053 (optimize_range_tests_diff): New function.
6054 (optimize_range_tests): Use optimize_range_tests_1.
6056 2013-10-15 Cong Hou <congh@google.com>
6058 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
6059 requirement of the reduction pattern so that one operand of the
6060 reduction operation can come from outside of the loop.
6062 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6064 * config/arm/neon-schedgen.ml: Remove.
6065 * config/arm/cortex-a9-neon.md: Remove comment regarding
6068 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6070 * config/arm/types: Remove old neon types.
6072 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6074 * config/arm/cortex-a7.md
6075 (cortex_a7_neon_type): New.
6076 (cortex_a7_neon_mul): Update for new types.
6077 (cortex_a7_neon_mla): Likewise.
6078 (cortex_a7_neon): Likewise.
6080 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6082 * config/arm/cortex-a15-neon.md
6083 (cortex_a15_neon_type): New,
6085 (cortex_a15_neon_int_1): Remove.
6086 (cortex_a15_neon_int_2): Likewise.
6087 (cortex_a15_neon_int_3): Likewise.
6088 (cortex_a15_neon_int_4): Likewise.
6089 (cortex_a15_neon_int_5): Likewise.
6090 (cortex_a15_neon_vqneg_vqabs): Likewise.
6091 (cortex_a15_neon_vmov): Likewise.
6092 (cortex_a15_neon_vaba): Likewise.
6093 (cortex_a15_neon_vaba_qqq): Likewise.
6094 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6095 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6096 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
6098 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6099 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6100 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
6101 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6102 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6103 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6104 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6105 (cortex_a15_neon_shift_1): Likewise.
6106 (cortex_a15_neon_shift_2): Likewise.
6107 (cortex_a15_neon_shift_3): Likewise.
6108 (cortex_a15_neon_vshl_ddd): Likewise.
6109 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6110 (cortex_a15_neon_vsra_vrsra): Likewise.
6111 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6112 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6113 (cortex_a15_neon_bp_3cycle): Likewise.
6114 (cortex_a15_neon_ldm_2): Likewise.
6115 (cortex_a15_neon_stm_2): Likewise.
6116 (cortex_a15_neon_mcr): Likewise.
6117 (cortex_a15_neon_mrc): Likewise.
6118 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6119 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6120 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6121 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6122 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6123 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6124 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6125 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6126 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6127 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6128 (cortex_a15_neon_bp_simple): Likewise.
6129 (cortex_a15_neon_bp_2cycle): Likewise.
6130 (cortex_a15_neon_bp_3cycle): Likewise.
6131 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6132 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6133 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6134 (cortex_a15_neon_vld2_4_regs): Likewise.
6135 (cortex_a15_neon_vld3_vld4): Likewise.
6136 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6137 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6138 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
6139 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
6140 (cortex_a15_neon_vst3_vst4): Rename to...
6141 (cortex_a15_neon_vst4): This, update for new attributes.
6142 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
6143 (cortex_a15_neon_vld3_vld4_lane): Likewise.
6144 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6145 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6146 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6147 (cortex_a15_neon_ldm_2): Likewise.
6148 (cortex_a15_neon_stm_2): Likewise.
6149 (cortex_a15_neon_mcr): Likewise.
6150 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6151 (cortex_a15_neon_mrc): Likewise.
6152 (cortex_a15_neon_mrrc): Likewise.
6154 (cortex_a15_neon_abd): New.
6155 (cortex_a15_neon_abd_q): Likewise.
6156 (cortex_a15_neon_aba): Likewise.
6157 (cortex_a15_neon_aba_q): Likewise.
6158 (cortex_a15_neon_acc): Likewise.
6159 (cortex_a15_neon_acc_q): Likewise.
6160 (cortex_a15_neon_arith_basic): Likewise.
6161 (cortex_a15_neon_arith_complex): Likewise.
6162 (cortex_a15_neon_multiply): Likewise.
6163 (cortex_a15_neon_multiply_q): Likewise.
6164 (cortex_a15_neon_mla): Likewise.
6165 (cortex_a15_neon_mla_q): Likewise.
6166 (cortex_a15_neon_sat_mla_long): Likewise.
6167 (cortex_a15_neon_shift_acc): Likewise.
6168 (cortex_a15_neon_shift_imm_basic): Likewise.
6169 (cortex_a15_neon_shift_imm_complex): Likewise.
6170 (cortex_a15_neon_shift_reg_basic): Likewise.
6171 (cortex_a15_neon_shift_reg_basic_q): Likewise.
6172 (cortex_a15_neon_shift_reg_complex): Likewise.
6173 (cortex_a15_neon_shift_reg_complex_q): Likewise.
6174 (cortex_a15_neon_fp_negabs): Likewise
6175 (cortex_a15_neon_fp_arith): Likewise
6176 (cortex_a15_neon_fp_arith_q): Likewise
6177 (cortex_a15_neon_fp_cvt_int): Likewise
6178 (cortex_a15_neon_fp_cvt_int_q): Likewise
6179 (cortex_a15_neon_fp_cvt_16): Likewise
6180 (cortex_a15_neon_fp_mul): Likewise
6181 (cortex_a15_neon_fp_mul_q): Likewise
6182 (cortex_a15_neon_fp_mla): Likewise
6183 (cortex_a15_neon_fp_mla_q): Likewise
6184 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
6185 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
6186 (cortex_a15_neon_bitops): Likewise.
6187 (cortex_a15_neon_bitops_q): Likewise.
6188 (cortex_a15_neon_from_gp): Likewise.
6189 (cortex_a15_neon_from_gp_q): Likewise.
6190 (cortex_a15_neon_tbl3_tbl4): Likewise.
6191 (cortex_a15_neon_zip_q): Likewise.
6192 (cortex_a15_neon_to_gp): Likewise.
6193 (cortex_a15_neon_load_a): Likewise.
6194 (cortex_a15_neon_load_b): Likewise.
6195 (cortex_a15_neon_load_c): Likewise.
6196 (cortex_a15_neon_load_d): Likewise.
6197 (cortex_a15_neon_load_e): Likewise.
6198 (cortex_a15_neon_load_f): Likewise.
6199 (cortex_a15_neon_store_a): Likewise.
6200 (cortex_a15_neon_store_b): Likewise.
6201 (cortex_a15_neon_store_c): Likewise.
6202 (cortex_a15_neon_store_d): Likewise.
6203 (cortex_a15_neon_store_e): Likewise.
6204 (cortex_a15_neon_store_f): Likewise.
6205 (cortex_a15_neon_store_g): Likewise.
6206 (cortex_a15_neon_store_h): Likewise.
6207 (cortex_a15_vfp_to_from_gp): Likewise.
6209 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6211 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
6213 (cortex_a9_neon_vshl_ddd): Remove.
6214 (cortex_a9_neon_vst3_vst4): Likewise.
6215 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6217 (cortex_a9_neon_bit_ops_q): New.
6219 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
6220 (cortex_a9_neon_int_2): Likewise.
6221 (cortex_a9_neon_int_3): Likewise.
6222 (cortex_a9_neon_int_4): Likewise.
6223 (cortex_a9_neon_int_5): Likewise.
6224 (cortex_a9_neon_vqneg_vqabs): Likewise.
6225 (cortex_a9_neon_vmov): Likewise.
6226 (cortex_a9_neon_vaba): Likewise.
6227 (cortex_a9_neon_vaba_qqq): Likewise.
6228 (cortex_a9_neon_shift_1): Likewise.
6229 (cortex_a9_neon_shift_2): Likewise.
6230 (cortex_a9_neon_shift_3): Likewise.
6231 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6232 (cortex_a9_neon_vsra_vrsra): Likewise.
6233 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6234 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6235 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
6237 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6238 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6239 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
6241 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6242 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6243 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
6244 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6245 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
6246 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
6247 (cortex_a9_neon_fp_vsum): Likewise.
6248 (cortex_a9_neon_fp_vmul_ddd): Likewise.
6249 (cortex_a9_neon_fp_vmul_qqd): Likewise.
6250 (cortex_a9_neon_fp_vmla_ddd): Likewise.
6251 (cortex_a9_neon_fp_vmla_qqq): Likewise.
6252 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
6253 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
6254 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6255 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6256 (cortex_a9_neon_bp_simple): Likewise.
6257 (cortex_a9_neon_bp_2cycle): Likewise.
6258 (cortex_a9_neon_bp_3cycle): Likewise.
6259 (cortex_a9_neon_ldr): Likewise.
6260 (cortex_a9_neon_str): Likewise.
6261 (cortex_a9_neon_vld1_1_2_regs): Likewise.
6262 (cortex_a9_neon_vld1_3_4_regs): Likewise.
6263 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6264 (cortex_a9_neon_vld2_4_regs): Likewise.
6265 (cortex_a9_neon_vld3_vld4): Likewise.
6266 (cortex_a9_neon_vld1_vld2_lane): Likewise.
6267 (cortex_a9_neon_vld3_vld4_lane): Likewise.
6268 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6269 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6270 (cortex_a9_neon_vst1_3_4_regs): Likewise.
6271 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
6272 (cortex_a9_neon_vst1_vst2_lane): Likewise.
6273 (cortex_a9_neon_vst3_vst4_lane): Likewise.
6274 (cortex_a9_neon_mcr): Likewise.
6275 (cortex_a9_neon_mcr_2_mcrr): Likewise.
6276 (cortex_a9_neon_mrc): Likewise.
6277 (cortex_a9_neon_mrrc): Likewise.
6279 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6281 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
6283 (cortex_a8_neon_vshl_ddd): Remove.
6284 (cortex_a8_neon_vst3_vst4): Likewise.
6285 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6287 (cortex_a8_neon_bit_ops_q): New.
6289 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
6290 (cortex_a8_neon_int_2): Likewise..
6291 (cortex_a8_neon_int_3): Likewise.
6292 (cortex_a8_neon_int_5): Likewise.
6293 (cortex_a8_neon_vqneg_vqabs): Likewise.
6294 (cortex_a8_neon_int_4): Likewise.
6295 (cortex_a8_neon_vaba): Likewise.
6296 (cortex_a8_neon_vaba_qqq): Likewise.
6297 (cortex_a8_neon_shift_1): Likewise.
6298 (cortex_a8_neon_shift_2): Likewise.
6299 (cortex_a8_neon_shift_3): Likewise.
6300 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6301 (cortex_a8_neon_vsra_vrsra): Likewise.
6302 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6303 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6304 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
6306 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6307 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6308 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
6310 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6311 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6312 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6313 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6314 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6315 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6316 (cortex_a8_neon_fp_vsum): Likewise.
6317 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6318 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6319 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6320 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6321 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6322 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6323 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6324 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6325 (cortex_a8_neon_bp_simple): Likewise.
6326 (cortex_a8_neon_bp_2cycle): Likewise.
6327 (cortex_a8_neon_bp_3cycle): Likewise.
6328 (cortex_a8_neon_ldr): Likewise.
6329 (cortex_a8_neon_str): Likewise.
6330 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6331 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6332 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6333 (cortex_a8_neon_vld2_4_regs): Likewise.
6334 (cortex_a8_neon_vld3_vld4): Likewise.
6335 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6336 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6337 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6338 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6339 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6340 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6341 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6342 (cortex_a8_neon_mcr): Likewise.
6343 (cortex_a8_neon_mcr_2_mcrr): Likewise.
6344 (cortex_a8_neon_mrc): Likewise.
6345 (cortex_a8_neon_mrrc): Likewise.
6347 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6349 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
6351 * config/aarch64/aarch64-simd.md (neon_type): Remove.
6352 (aarch64_simd_dup<mode>): Add "type" attribute.
6353 (aarch64_dup_lane<mode>): Likewise.
6354 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
6355 (*aarch64_simd_mov<mode>): Likewise.
6356 (aarch64_simd_mov_from_<mode>low): Likewise.
6357 (aarch64_simd_mov_from_<mode>high): Likewise.
6358 (orn<mode>3): Likewise.
6359 (bic<mode>3): Likewise.
6360 (add<mode>3): Likewise.
6361 (sub<mode>3): Likewise.
6362 (mul<mode>3): Likewise.
6363 (*aarch64_mul3_elt<mode>): Likewise.
6364 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
6365 (*aarch64_mul3_elt_to_128df): Likewise.
6366 (*aarch64_mul3_elt_to_64v2df): Likewise.
6367 (neg<mode>2): Likewise.
6368 (abs<mode>2): Likewise.
6369 (abd<mode>_3): Likewise.
6370 (aba<mode>_3): Likewise.
6371 (fabd<mode>_3): Likewise.
6372 (*fabd_scalar<mode>3): Likewise.
6373 (and<mode>3): Likewise.
6374 (ior<mode>3): Likewise.
6375 (xor<mode>3): Likewise.
6376 (one_cmpl<mode>2): Likewise.
6377 (aarch64_simd_vec_set<mode>): Likewise.
6378 (aarch64_simd_lshr<mode>): Likewise.
6379 (aarch64_simd_ashr<mode>): Likewise.
6380 (aarch64_simd_imm_shl<mode>): Likewise.
6381 (aarch64_simd_reg_sshl<mode): Likewise.
6382 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
6383 (aarch64_simd_reg_shl<mode>_signed): Likewise.
6384 (aarch64_simd_vec_setv2di): Likewise.
6385 (aarch64_simd_vec_set<mode>): Likewise.
6386 (aarch64_mla<mode>): Likewise.
6387 (*aarch64_mla_elt<mode>): Likewise.
6388 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6389 (aarch64_mls<mode>): Likewise.
6390 (*aarch64_mls_elt<mode>): Likewise.
6391 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6392 (<su><maxmin><mode>3): Likewise.
6393 (move_lo_quad_<mode>): Likewise.
6394 (aarch64_simd_move_hi_quad_<mode>): Likewise.
6395 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
6396 (vec_pack_trunc_<mode>): Likewise.
6397 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
6398 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
6399 (*aarch64_<su>mlal_lo<mode>): Likewise.
6400 (*aarch64_<su>mlal_hi<mode>): Likewise.
6401 (*aarch64_<su>mlsl_lo<mode>): Likewise.
6402 (*aarch64_<su>mlsl_hi<mode>): Likewise.
6403 (*aarch64_<su>mlal<mode>): Likewise.
6404 (*aarch64_<su>mlsl<mode>): Likewise.
6405 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
6406 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
6407 (add<mode>3): Likewise.
6408 (sub<mode>3): Likewise.
6409 (mul<mode>3): Likewise.
6410 (div<mode>3): Likewise.
6411 (neg<mode>2): Likewise.
6412 (abs<mode>2): Likewise.
6413 (fma<mode>4): Likewise.
6414 (*aarch64_fma4_elt<mode>): Likewise.
6415 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6416 (*aarch64_fma4_elt_to_128df): Likewise.
6417 (*aarch64_fma4_elt_to_64v2df): Likewise.
6418 (fnma<mode>4): Likewise.
6419 (*aarch64_fnma4_elt<mode>): Likewise.
6420 (*aarch64_fnma4_elt_<vswap_width_name><mode>
6421 (*aarch64_fnma4_elt_to_128df): Likewise.
6422 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6423 (<frint_pattern><mode>2): Likewise.
6424 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
6425 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
6426 (vec_unpacks_lo_v4sf): Likewise.
6427 (aarch64_float_extend_lo_v2df): Likewise.
6428 (vec_unpacks_hi_v4sf): Likewise.
6429 (aarch64_float_truncate_lo_v2sf): Likewise.
6430 (aarch64_float_truncate_hi_v4sf): Likewise.
6431 (aarch64_vmls<mode>): Likewise.
6432 (<su><maxmin><mode>3): Likewise.
6433 (<maxmin_uns><mode>3): Likewise.
6434 (reduc_<sur>plus_<mode>): Likewise.
6435 (reduc_<sur>plus_v2di): Likewise.
6436 (reduc_<sur>plus_v2si): Likewise.
6437 (reduc_<sur>plus_<mode>): Likewise.
6438 (aarch64_addpv4sf): Likewise.
6439 (clz<mode>2): Likewise.
6440 (reduc_<maxmin_uns>_<mode>): Likewise.
6441 (reduc_<maxmin_uns>_v2di): Likewise.
6442 (reduc_<maxmin_uns>_v2si): Likewise.
6443 (reduc_<maxmin_uns>_<mode>): Likewise.
6444 (reduc_<maxmin_uns>_v4sf): Likewise.
6445 (aarch64_simd_bsl<mode>_internal): Likewise.
6446 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
6447 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
6448 (aarch64_get_lane<mode>): Likewise.
6449 (*aarch64_combinez<mode>): Likewise.
6450 (aarch64_combine<mode>): Likewise.
6451 (aarch64_simd_combine<mode>): Likewise.
6452 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
6453 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
6454 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
6455 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
6456 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
6457 (aarch64_<sur>h<addsub><mode>): Likewise.
6458 (aarch64_<sur><addsub>hn<mode>): Likewise.
6459 (aarch64_<sur><addsub>hn2<mode>): Likewise.
6460 (aarch64_pmul<mode>): Likewise.
6461 (aarch64_<su_optab><optab><mode>): Likewise.
6462 (aarch64_<sur>qadd<mode>): Likewise.
6463 (aarch64_sqmovun<mode>): Likewise.
6464 (aarch64_<sur>qmovn<mode>): Likewise.
6465 (aarch64_s<optab><mode>): Likewise.
6466 (aarch64_sq<r>dmulh<mode>): Likewise.
6467 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6468 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6469 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6470 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
6471 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6472 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6473 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
6474 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
6475 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6476 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6477 (aarch64_sqdmull<mode>): Likewise.
6478 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6479 (aarch64_sqdmull_n<mode>): Likewise.
6480 (aarch64_sqdmull2<mode>_internal): Likewise.
6481 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6482 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6483 (aarch64_<sur>shl<mode>): Likewise.
6484 (aarch64_<sur>q<r>shl<mode>
6485 (aarch64_<sur>shll_n<mode>): Likewise.
6486 (aarch64_<sur>shll2_n<mode>): Likewise.
6487 (aarch64_<sur>shr_n<mode>): Likewise.
6488 (aarch64_<sur>sra_n<mode>): Likewise.
6489 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
6490 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
6491 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
6492 (aarch64_cm<optab><mode>): Likewise.
6493 (aarch64_cm<optab>di): Likewise.
6494 (aarch64_cm<optab><mode>): Likewise.
6495 (aarch64_cm<optab>di): Likewise.
6496 (aarch64_cmtst<mode>): Likewise.
6497 (aarch64_cmtstdi): Likewise.
6498 (aarch64_cm<optab><mode>): Likewise.
6499 (*aarch64_fac<optab><mode>): Likewise.
6500 (aarch64_addp<mode>): Likewise.
6501 (aarch64_addpdi): Likewise.
6502 (sqrt<mode>2): Likewise.
6503 (vec_load_lanesoi<mode>): Likewise.
6504 (vec_store_lanesoi<mode>): Likewise.
6505 (vec_load_lanesci<mode>): Likewise.
6506 (vec_store_lanesci<mode>): Likewise.
6507 (vec_load_lanesxi<mode>): Likewise.
6508 (vec_store_lanesxi<mode>): Likewise.
6509 (*aarch64_mov<mode>): Likewise.
6510 (aarch64_ld2<mode>_dreg): Likewise.
6511 (aarch64_ld2<mode>_dreg): Likewise.
6512 (aarch64_ld3<mode>_dreg): Likewise.
6513 (aarch64_ld3<mode>_dreg): Likewise.
6514 (aarch64_ld4<mode>_dreg): Likewise.
6515 (aarch64_ld4<mode>_dreg): Likewise.
6516 (aarch64_tbl1<mode>): Likewise.
6517 (aarch64_tbl2v16qi): Likewise.
6518 (aarch64_combinev16qi): Likewise.
6519 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
6520 (aarch64_st2<mode>_dreg): Likewise.
6521 (aarch64_st2<mode>_dreg): Likewise.
6522 (aarch64_st3<mode>_dreg): Likewise.
6523 (aarch64_st3<mode>_dreg): Likewise.
6524 (aarch64_st4<mode>_dreg): Likewise.
6525 (aarch64_st4<mode>_dreg): Likewise.
6526 (*aarch64_simd_ld1r<mode>): Likewise.
6527 (aarch64_frecpe<mode>): Likewise.
6528 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
6529 (aarch64_frecps<mode>): Likewise.
6531 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6533 * config/arm/iterators.md (V_elem_ch): New.
6535 (VQH_type): Likewise.
6536 * config/arm/arm.md (is_neon_type): New.
6537 (conds): Use is_neon_type.
6538 (anddi3_insn): Update type attribute.
6539 (xordi3_insn): Likewise.
6540 (one_cmpldi2): Likewise.
6541 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
6542 * config/arm/neon.md (neon_mov): Update type attribute.
6543 (*movmisalign<mode>_neon_store): Likewise.
6544 (*movmisalign<mode>_neon_load): Likewise.
6545 (vec_set<mode>_internal): Likewise.
6546 (vec_set<mode>_internal): Likewise.
6547 (vec_setv2di_internal): Likewise.
6548 (vec_extract<mode>): Likewise.
6549 (vec_extract<mode>): Likewise.
6550 (vec_extractv2di): Likewise.
6551 (*add<mode>3_neon): Likewise.
6552 (adddi3_neon): Likewise.
6553 (*sub<mode>3_neon): Likewise.
6554 (subdi3_neon): Likewise.
6555 (fma<VCVTF:mode>4): Likewise.
6556 (fma<VCVTF:mode>4_intrinsic): Likewise.
6557 (*fmsub<VCVTF:mode>4): Likewise.
6558 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6559 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6560 (ior<mode>3): Likewise.
6561 (and<mode>3): Likewise.
6562 (orn<mode>3_neon): Likewise.
6563 (orndi3_neon): Likewise.
6564 (bic<mode>3_neon): Likewise.
6565 (bicdi3_neon): Likewise.
6566 (xor<mode>3): Likewise.
6567 (one_cmpl<mode>2): Likewise.
6568 (abs<mode>2): Likewise.
6569 (neg<mode>2): Likewise.
6570 (negdi2_neon): Likewise.
6571 (*umin<mode>3_neon): Likewise.
6572 (*umax<mode>3_neon): Likewise.
6573 (*smin<mode>3_neon): Likewise.
6574 (*smax<mode>3_neon): Likewise.
6575 (vashl<mode>3): Likewise.
6576 (vashr<mode>3_imm): Likewise.
6577 (vlshr<mode>3_imm): Likewise.
6578 (ashl<mode>3_signed): Likewise.
6579 (ashl<mode>3_unsigned): Likewise.
6580 (neon_load_count): Likewise.
6581 (ashldi3_neon_noclobber): Likewise.
6582 (ashldi3_neon): Likewise.
6583 (signed_shift_di3_neon): Likewise.
6584 (unsigned_shift_di3_neon): Likewise.
6585 (ashrdi3_neon_imm_noclobber): Likewise.
6586 (lshrdi3_neon_imm_noclobber): Likewise.
6587 (<shift>di3_neon): Likewise.
6588 (widen_ssum<mode>3): Likewise.
6589 (widen_usum<mode>3): Likewise.
6590 (quad_halves_<code>v4si): Likewise.
6591 (quad_halves_<code>v4sf): Likewise.
6592 (quad_halves_<code>v8hi): Likewise.
6593 (quad_halves_<code>v16qi): Likewise.
6594 (reduc_splus_v2di): Likewise.
6595 (neon_vpadd_internal<mode>): Likewise.
6596 (neon_vpsmin<mode>): Likewise.
6597 (neon_vpsmax<mode>): Likewise.
6598 (neon_vpumin<mode>): Likewise.
6599 (neon_vpumax<mode>): Likewise.
6600 (*ss_add<mode>_neon): Likewise.
6601 (*us_add<mode>_neon): Likewise.
6602 (*ss_sub<mode>_neon): Likewise.
6603 (*us_sub<mode>_neon): Likewise.
6604 (neon_vadd<mode>_unspec): Likewise.
6605 (neon_vaddl<mode>): Likewise.
6606 (neon_vaddw<mode>): Likewise.
6607 (neon_vhadd<mode>): Likewise.
6608 (neon_vqadd<mode>): Likewise.
6609 (neon_vaddhn<mode>): Likewise.
6610 (neon_vmul<mode>): Likewise.
6611 (neon_vfms<VCVTF:mode>): Likewise.
6612 (neon_vmlal<mode>): Likewise.
6613 (neon_vmls<mode>): Likewise.
6614 (neon_vmlsl<mode>): Likewise.
6615 (neon_vqdmulh<mode>): Likewise.
6616 (neon_vqdmlal<mode>): Likewise.
6617 (neon_vqdmlsl<mode>): Likewise.
6618 (neon_vmull<mode>): Likewise.
6619 (neon_vqdmull<mode>): Likewise.
6620 (neon_vsub<mode>_unspec): Likewise.
6621 (neon_vsubl<mode>): Likewise.
6622 (neon_vsubw<mode>): Likewise.
6623 (neon_vqsub<mode>): Likewise.
6624 (neon_vhsub<mode>): Likewise.
6625 (neon_vsubhn<mode>): Likewise.
6626 (neon_vceq<mode>): Likewise.
6627 (neon_vcge<mode>): Likewise.
6628 (neon_vcgeu<mode>): Likewise.
6629 (neon_vcgt<mode>): Likewise.
6630 (neon_vcgtu<mode>): Likewise.
6631 (neon_vcle<mode>): Likewise.
6632 (neon_vclt<mode>): Likewise.
6633 (neon_vcage<mode>): Likewise.
6634 (neon_vcagt<mode>): Likewise.
6635 (neon_vtst<mode>): Likewise.
6636 (neon_vabd<mode>): Likewise.
6637 (neon_vabdl<mode>): Likewise.
6638 (neon_vaba<mode>): Likewise.
6639 (neon_vabal<mode>): Likewise.
6640 (neon_vmax<mode>): Likewise.
6641 (neon_vmin<mode>): Likewise.
6642 (neon_vpaddl<mode>): Likewise.
6643 (neon_vpadal<mode>): Likewise.
6644 (neon_vpmax<mode>): Likewise.
6645 (neon_vpmin<mode>): Likewise.
6646 (neon_vrecps<mode>): Likewise.
6647 (neon_vrsqrts<mode>): Likewise.
6648 (neon_vqabs<mode>): Likewise.
6649 (neon_vqneg<mode>): Likewise.
6650 (neon_vcls<mode>): Likewise.
6651 (clz<mode>2): Likewise.
6652 (popcount<mode>2): Likewise.
6653 (neon_vrecpe<mode>): Likewise.
6654 (neon_vrsqrte<mode>): Likewise.
6655 (neon_vget_lane<mode>_sext_internal): Likewise.
6656 (neon_vget_lane<mode>_zext_internal): Likewise.
6657 (neon_vdup_n<mode>): Likewise.
6658 (neon_vdup_n<mode>): Likewise.
6659 (neon_vdup_nv2di): Likewise.
6660 (neon_vdup_lane<mode>_interal): Likewise.
6661 (*neon_vswp<mode>): Likewise.
6662 (neon_vcombine<mode>): Likewise.
6663 (float<mode><V_cvtto>2): Likewise.
6664 (floatuns<mode><V_cvtto>2): Likewise.
6665 (fix_trunc<mode><V_cvtto>2): Likewise.
6666 (fixuns_trunc<mode><V_cvtto>2
6667 (neon_vcvt<mode>): Likewise.
6668 (neon_vcvt<mode>): Likewise.
6669 (neon_vcvtv4sfv4hf): Likewise.
6670 (neon_vcvtv4hfv4sf): Likewise.
6671 (neon_vcvt_n<mode>): Likewise.
6672 (neon_vcvt_n<mode>): Likewise.
6673 (neon_vmovn<mode>): Likewise.
6674 (neon_vqmovn<mode>): Likewise.
6675 (neon_vqmovun<mode>): Likewise.
6676 (neon_vmovl<mode>): Likewise.
6677 (neon_vmul_lane<mode>): Likewise.
6678 (neon_vmul_lane<mode>): Likewise.
6679 (neon_vmull_lane<mode>): Likewise.
6680 (neon_vqdmull_lane<mode>): Likewise.
6681 (neon_vqdmulh_lane<mode>): Likewise.
6682 (neon_vqdmulh_lane<mode>): Likewise.
6683 (neon_vmla_lane<mode>): Likewise.
6684 (neon_vmla_lane<mode>): Likewise.
6685 (neon_vmlal_lane<mode>): Likewise.
6686 (neon_vqdmlal_lane<mode>): Likewise.
6687 (neon_vmls_lane<mode>): Likewise.
6688 (neon_vmls_lane<mode>): Likewise.
6689 (neon_vmlsl_lane<mode>): Likewise.
6690 (neon_vqdmlsl_lane<mode>): Likewise.
6691 (neon_vext<mode>): Likewise.
6692 (neon_vrev64<mode>): Likewise.
6693 (neon_vrev32<mode>): Likewise.
6694 (neon_vrev16<mode>): Likewise.
6695 (neon_vbsl<mode>_internal): Likewise.
6696 (neon_vshl<mode>): Likewise.
6697 (neon_vqshl<mode>): Likewise.
6698 (neon_vshr_n<mode>): Likewise.
6699 (neon_vshrn_n<mode>): Likewise.
6700 (neon_vqshrn_n<mode>): Likewise.
6701 (neon_vqshrun_n<mode>): Likewise.
6702 (neon_vshl_n<mode>): Likewise.
6703 (neon_vqshl_n<mode>): Likewise.
6704 (neon_vqshlu_n<mode>): Likewise.
6705 (neon_vshll_n<mode>): Likewise.
6706 (neon_vsra_n<mode>): Likewise.
6707 (neon_vsri_n<mode>): Likewise.
6708 (neon_vsli_n<mode>): Likewise.
6709 (neon_vtbl1v8qi): Likewise.
6710 (neon_vtbl2v8qi): Likewise.
6711 (neon_vtbl3v8qi): Likewise.
6712 (neon_vtbl4v8qi): Likewise.
6713 (neon_vtbl1v16qi): Likewise.
6714 (neon_vtbl2v16qi): Likewise.
6715 (neon_vcombinev16qi): Likewise.
6716 (neon_vtbx1v8qi): Likewise.
6717 (neon_vtbx2v8qi): Likewise.
6718 (neon_vtbx3v8qi): Likewise.
6719 (neon_vtbx4v8qi): Likewise.
6720 (*neon_vtrn<mode>_insn): Likewise.
6721 (*neon_vzip<mode>_insn): Likewise.
6722 (*neon_vuzp<mode>_insn): Likewise.
6723 (neon_vld1<mode>): Likewise.
6724 (neon_vld1_lane<mode>): Likewise.
6725 (neon_vld1_lane<mode>): Likewise.
6726 (neon_vld1_dup<mode>): Likewise.
6727 (neon_vld1_dup<mode>): Likewise.
6728 (neon_vld1_dupv2di): Likewise.
6729 (neon_vst1<mode>): Likewise.
6730 (neon_vst1_lane<mode>): Likewise.
6731 (neon_vst1_lane<mode>): Likewise.
6732 (neon_vld2<mode>): Likewise.
6733 (neon_vld2<mode>): Likewise.
6734 (neon_vld2_lane<mode>): Likewise.
6735 (neon_vld2_lane<mode>): Likewise.
6736 (neon_vld2_dup<mode>): Likewise.
6737 (neon_vst2<mode>): Likewise.
6738 (neon_vst2<mode>): Likewise.
6739 (neon_vst2_lane<mode>): Likewise.
6740 (neon_vst2_lane<mode>): Likewise.
6741 (neon_vld3<mode>): Likewise.
6742 (neon_vld3qa<mode>): Likewise.
6743 (neon_vld3qb<mode>): Likewise.
6744 (neon_vld3_lane<mode>): Likewise.
6745 (neon_vld3_lane<mode>): Likewise.
6746 (neon_vld3_dup<mode>): Likewise.
6747 (neon_vst3<mode>): Likewise.
6748 (neon_vst3qa<mode>): Likewise.
6749 (neon_vst3qb<mode>): Likewise.
6750 (neon_vst3_lane<mode>): Likewise.
6751 (neon_vst3_lane<mode>): Likewise.
6752 (neon_vld4<mode>): Likewise.
6753 (neon_vld4qa<mode>): Likewise.
6754 (neon_vld4qb<mode>): Likewise.
6755 (neon_vld4_lane<mode>): Likewise.
6756 (neon_vld4_lane<mode>): Likewise.
6757 (neon_vld4_dup<mode>): Likewise.
6758 (neon_vst4<mode>): Likewise.
6759 (neon_vst4qa<mode>): Likewise.
6760 (neon_vst4qb<mode>): Likewise.
6761 (neon_vst4_lane<mode>): Likewise.
6762 (neon_vst4_lane<mode>): Likewise.
6763 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6764 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6765 (neon_vec_<US>mult_lo_<mode>): Likewise.
6766 (neon_vec_<US>mult_hi_<mode>): Likewise.
6767 (neon_vec_<US>shiftl_<mode>): Likewise.
6768 (neon_unpack<US>_<mode>): Likewise.
6769 (neon_vec_<US>mult_<mode>): Likewise.
6770 (vec_pack_trunc_<mode>): Likewise.
6771 (neon_vec_pack_trunc_<mode>): Likewise.
6772 (neon_vabd<mode>_2): Likewise.
6773 (neon_vabd<mode>_3): Likewise.
6775 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6777 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
6778 (load_pair): Update type attribute.
6779 (store_pair): Update type attribute.
6780 * config/aarch64/iterators.md (q): New.
6782 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6784 * config/arm/types.md: Add new types for Neon insns.
6786 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6787 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6788 Sergey Lega <sergey.s.lega@intel.com>
6789 Anna Tikhonova <anna.tikhonova@intel.com>
6790 Ilya Tocar <ilya.tocar@intel.com>
6791 Andrey Turetskiy <andrey.turetskiy@intel.com>
6792 Ilya Verbin <ilya.verbin@intel.com>
6793 Kirill Yukhin <kirill.yukhin@intel.com>
6794 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6796 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
6797 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
6798 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
6800 (srcp14<mode>): Ditto.
6801 (rsqrt14<mode>): Ditto.
6802 (rsqrt14<mode>): Ditto.
6803 (avx512f_vmscalef<mode>): Ditto.
6804 (avx512f_scalef<mode>): Ditto.
6805 (avx512f_getexp<mode>): Ditto.
6806 (avx512f_sgetexp<mode>): Ditto.
6807 (avx512f_fixupimm<mode>): Ditto.
6808 (avx512f_sfixupimm<mode>): Ditto.
6809 (avx512f_rndscale<mode>): Ditto.
6810 (*avx512er_exp2<mode>): Ditto.
6811 (*avx512er_rcp28<mode>): Ditto.
6812 (avx512er_rsqrt28<mode>): Ditto.
6813 (avx512f_getmant<mode>): Ditto.
6814 (avx512f_getmant<mode>): Ditto.
6815 (avx512f_rndscale<mode>): Fix formatting.
6817 2013-10-15 Martin Jambor <mjambor@suse.cz>
6819 * ipa-utils.h (ipa_edge_within_scc): Declare.
6820 * ipa-cp.c (edge_within_scc): Moved...
6821 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
6823 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6824 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6825 Sergey Lega <sergey.s.lega@intel.com>
6826 Anna Tikhonova <anna.tikhonova@intel.com>
6827 Ilya Tocar <ilya.tocar@intel.com>
6828 Andrey Turetskiy <andrey.turetskiy@intel.com>
6829 Ilya Verbin <ilya.verbin@intel.com>
6830 Kirill Yukhin <kirill.yukhin@intel.com>
6831 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6833 * config/i386/predicates.md (const_8_to_15_operand): New.
6834 (const_16_to_31_operand): Ditto.
6835 * config/i386/sse.md (V8FI): New.
6837 (reduc_splus_v8df): Ditto.
6838 (reduc_splus_v16sf): Ditto.
6839 (avx512f_vextract<shuffletype>32x4_1): Ditto.
6840 (vec_extract_hi_<mode>): Ditto.
6841 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
6842 (vec_set_lo_<mode>): Ditto.
6843 (vec_set_hi_<mode>): Ditto.
6844 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
6845 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
6846 (avx512f_pshufd_1): Ditto.
6847 (avx512f_broadcast<mode>): Ditto.
6848 (avx512f_broadcast<mode>): Ditto.
6849 (define_split): Split vec_extract_lo into move.
6850 (ssequartermode): Ditto.
6851 (ssedoublemode): Extened with wider modes.
6852 (vec_extract_lo_<mode>): Ditto.
6854 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6855 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6856 Sergey Lega <sergey.s.lega@intel.com>
6857 Anna Tikhonova <anna.tikhonova@intel.com>
6858 Ilya Tocar <ilya.tocar@intel.com>
6859 Andrey Turetskiy <andrey.turetskiy@intel.com>
6860 Ilya Verbin <ilya.verbin@intel.com>
6861 Kirill Yukhin <kirill.yukhin@intel.com>
6862 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6864 * config/i386/predicates.md (register_or_constm1_operand): New.
6865 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
6866 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
6867 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
6868 UNSPEC_SCATTER_PREFETCH
6870 (avx512f_ucmp<mode>3): Ditto.
6871 (avx512f_vternlog<mode>): Ditto.
6872 (avx512f_align<mode>): Ditto.
6873 (<shift_insn><mode>3): Ditto.
6874 (avx512f_<rotate>v<mode>): Ditto.
6875 (avx512f_<rotate><mode>): Ditto.
6876 (avx512f_eq<mode>3): Ditto.
6877 (avx512f_eq<mode>3_1): Ditto.
6878 (avx512f_gt<mode>3): Ditto.
6879 (avx512f_testm<mode>3): Ditto.
6880 (avx512f_testnm<mode>3): Ditto.
6881 (avx512pf_gatherpf<mode>): Ditto.
6882 (*avx512pf_gatherpf<mode>_mask): Ditto.
6883 (*avx512pf_gatherpf<mode>): Ditto.
6884 (avx512pf_scatterpf<mode>): Ditto.
6885 (*avx512pf_scatterpf<mode>_mask): Ditto.
6886 (*avx512pf_scatterpf<mode>): Ditto.
6887 (avx512f_vec_dup_gpr<mode>): Ditto.
6888 (clz<mode>2): Ditto.
6889 (conflict<mode>): Ditto.
6890 (REDUC_SMINMAX_MODE): Extened with wider modes.
6891 (reduc_<code>_<mode>): Ditto.
6892 (vlshr<mode>3): Ditto.
6893 (vashl<mode>3): Ditto.
6895 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6896 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6897 Sergey Lega <sergey.s.lega@intel.com>
6898 Anna Tikhonova <anna.tikhonova@intel.com>
6899 Ilya Tocar <ilya.tocar@intel.com>
6900 Andrey Turetskiy <andrey.turetskiy@intel.com>
6901 Ilya Verbin <ilya.verbin@intel.com>
6902 Kirill Yukhin <kirill.yukhin@intel.com>
6903 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6905 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
6908 (avx512fmaskmode): Ditto.
6909 (bcstscalarsuff): Ditto.
6910 (avx512f_blendm<mode>): Ditto.
6911 (cmp_imm_predicate): Ditto.
6912 (avx512f_cmp<mode>3): Ditto.
6913 (avx512f_vec_dup<mode>): Ditto.
6914 (avx512f_vec_dup_mem<mode>): Ditto.
6915 (avx512f_vpermi2var<mode>3): Ditto.
6916 (avx512f_vpermt2var<mode>3): Ditto.
6917 (vec_init<mode>): Ditto.
6918 (avx512f_gathersi<mode>): Ditto.
6919 (*avx512f_gathersi<mode>): Ditto.
6920 (*avx512f_gathersi<mode>_2): Ditto.
6921 (avx512f_gatherdi<mode>): Ditto.
6922 (*avx512f_gatherdi<mode>): Ditto.
6923 (*avx512f_gatherdi<mode>_2): Ditto.
6924 (avx512f_scattersi<mode>): Ditto.
6925 (*avx512f_scattersi<mode>): Ditto.
6926 (avx512f_scatterdi<mode>): Ditto.
6927 (*avx512f_scatterdi<mode>): Ditto.
6928 (sseintprefix): Extened with wider modes.
6929 (VEC_GATHER_IDXSI): Ditto.
6930 (VEC_GATHER_IDXDI): Ditto.
6931 (VEC_GATHER_SRCDI): Ditto.
6933 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
6934 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6936 * config/arm/t-aprofile: New file.
6937 * config.gcc: Handle --with-multilib-list option.
6939 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
6941 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
6944 2013-10-15 Richard Biener <rguenther@suse.de>
6946 * tree-loop-distribution.c (build_empty_rdg): Inline into
6948 (rdg_flag_vertex): Inline into single user.
6949 (rdg_flag_vertex_and_dependent): Likewise.
6950 (build_rdg_partition_for_vertex): Remove processed bitmap.
6951 (rdg_build_partitions): Simplify.
6953 2013-10-15 Richard Biener <rguenther@suse.de>
6955 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
6956 Restructure forwarding through conversions and copies to
6957 avoid performing copy-propagation the wrong way. Adjust
6958 recursion invocations.
6959 (forward_propagate_addr_expr): Add argument stating if we
6960 are recursing from a single-use.
6961 (ssa_forward_propagate_and_combine): Adjust.
6963 2013-10-14 David Malcolm <dmalcolm@redhat.com>
6965 * dumpfile.h (gcc::dump_manager): New class, to hold state
6966 relating to dumpfile management.
6967 (get_dump_file_name): Remove in favor of method of dump_manager.
6968 (dump_initialized_p): Likewise.
6969 (dump_start): Likewise.
6970 (dump_finish): Likewise.
6971 (dump_switch_p): Likewise.
6972 (dump_register): Likewise.
6973 (get_dump_file_info): Likewise.
6974 * context.c (gcc::context::context): Construct the dump_manager
6976 * context.h (gcc::context::get_dumps): New.
6977 (gcc::context::m_dumps): New.
6978 * coverage.c (coverage_init): Port to dump_manager API.
6979 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
6980 (extra_dump_files_in_use): Likewise.
6981 (extra_dump_files_alloced): Likewise.
6982 (gcc::dump_manager::dump_manager): New.
6983 (dump_register): Convert to...
6984 (gcc::dump_manager::dump_register): ...method, replacing
6985 function-static next_dump with m_next_dump field.
6986 (get_dump_file_info): Convert to...
6987 (gcc::dump_manager::get_dump_file_info): ...method.
6988 (get_dump_file_name): Convert to...
6989 (gcc::dump_manager::get_dump_file_name): ...method.
6990 (dump_start): Convert to...
6991 (gcc::dump_manager::dump_start): ...method.
6992 (dump_finish): Convert to...
6993 (gcc::dump_manager::dump_finish): ...method.
6994 (dump_begin): Replace body with...
6995 (gcc::dump_manager::dump_begin): ...new method.
6996 (dump_phase_enabled_p): Convert to...
6997 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6998 (dump_phase_enabled_p): Convert to...
6999 (gcc::dump_manager::dump_phase_enabled_p): ...method.
7000 (dump_initialized_p): Convert to...
7001 (gcc::dump_manager::dump_initialized_p): ...method.
7002 (dump_flag_name): Replace body with...
7003 (gcc::dump_manager::dump_flag_name): ...new method.
7004 (dump_enable_all): Convert to...
7005 (gcc::dump_manager::dump_enable_all): ...new method.
7006 (opt_info_enable_passes): Convert to...
7007 (gcc::dump_manager::opt_info_enable_passes): ...new method.
7008 (dump_switch_p_1): Convert to...
7009 (gcc::dump_manager::dump_switch_p_1): ...new method.
7010 (dump_switch_p): Convert to...
7011 (gcc::dump_manager::dump_switch_p): ...new method.
7012 (opt_info_switch_p): Port to dump_manager API.
7013 (enable_rtl_dump_file): Likewise.
7014 * opts-global.c (handle_common_deferred_options): Port to new
7016 * passes.c (pass_manager::finish_optimization_passes): Likewise.
7017 (pass_manager::register_one_dump_file): Likewise.
7018 (pass_manager::register_pass): Likewise.
7019 (pass_init_dump_file): Likewise.
7020 (pass_fini_dump_file): Likewise.
7021 * statistics.c (statistics_early_init): Likewise.
7023 2013-10-14 Richard Biener <rguenther@suse.de>
7025 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
7026 iterative_hash_canonical_type, gimple_canonical_type_hash,
7027 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
7028 gimple_register_canonical_type, print_gimple_types_stats,
7029 free_gimple_type_tables): Move to lto/lto.c
7030 (gt-gimple.h): Do not include.
7031 * gimple.h (gimple_register_canonical_type,
7032 print_gimple_types_stats, free_gimple_type_tables): Remove.
7033 * Makefile.in (GTFILES): Remove gimple.c.
7035 2013-10-14 Travis Snoozy <quandary@remstate.com>
7038 * config/msp430/msp430.c (msp430_option_override): Correct thinko
7039 scanning for msp430x targets.
7041 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
7044 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
7045 (registers_ok_for_ldd_peep): Move around.
7046 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
7047 * config/sparc/sparc.md (widening peepholes): Use it.
7049 2013-10-14 Richard Biener <rguenther@suse.de>
7053 * gimple.c (iterative_hash_canonical_type): Make sure to
7054 record the hash into the correct hashtable slot.
7056 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
7058 PR rtl-optimization/58662
7059 * combine.c (try_combine): Take into account death nodes on I2 when
7060 splitting a PARALLEL of two independent SETs. Fix dump message.
7062 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
7065 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
7066 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
7067 * config/sh/t-sh (sh_treg_combine.o): New entry.
7068 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
7069 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
7070 (register_sh_passes): New function. Register sh_treg_combine pass.
7071 (sh_option_override): Invoke it.
7072 (sh_canonicalize_comparison): Handle op0_preserve_value.
7073 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
7074 opportunities. Canonicalize branch condition.
7075 (nott): Allow only if pseudos can be created for non-SH2A.
7077 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
7080 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
7081 (ix86_expand_movmem): Replace copy_addr_to_reg with
7082 ix86_copy_addr_to_reg.
7083 (ix86_expand_setmem): Likewise.
7085 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
7087 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
7088 provided by ix86_fp_compare_mode instead of CCFPUmode.
7090 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
7092 * config/aarch64/arm_neon.h
7093 (vtbx<1,3>_<psu>8): Fix register constriants.
7095 2013-10-11 Jeff Law <law@redhat.com>
7097 PR tree-optimization/58640
7098 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
7099 threading paths that cross over two loop entry points.
7101 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7103 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
7104 handle vector float as well.
7105 (*vsx_le_perm_load_v4si): Likewise.
7106 (*vsx_le_perm_store_v2di): Likewise.
7107 (*vsx_le_perm_store_v4si): Likewise.
7109 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7111 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
7112 directly to circumvent subtract from splat{31} workaround.
7113 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
7115 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
7116 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
7117 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
7118 (altivec_vperm_<mode>): Convert to define_insn_and_split to
7119 separate big and little endian logic.
7120 (*altivec_vperm_<mode>_internal): New define_insn.
7121 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
7122 separate big and little endian logic.
7123 (*altivec_vperm_<mode>_uns_internal): New define_insn.
7124 (vec_permv16qi): Add little endian logic.
7126 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
7128 * doc/extend.texi (returns_nonnull): Remove arguments.
7130 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7131 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7132 Sergey Lega <sergey.s.lega@intel.com>
7133 Anna Tikhonova <anna.tikhonova@intel.com>
7134 Ilya Tocar <ilya.tocar@intel.com>
7135 Andrey Turetskiy <andrey.turetskiy@intel.com>
7136 Ilya Verbin <ilya.verbin@intel.com>
7137 Kirill Yukhin <kirill.yukhin@intel.com>
7138 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7140 * config/i386/sse.md (VI48F_256_512): New.
7141 (avx2_permvar<mode>): Change to ...
7142 (<avx2_avx512f>_permvar<mode>): This.
7144 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7145 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7146 Sergey Lega <sergey.s.lega@intel.com>
7147 Anna Tikhonova <anna.tikhonova@intel.com>
7148 Ilya Tocar <ilya.tocar@intel.com>
7149 Andrey Turetskiy <andrey.turetskiy@intel.com>
7150 Ilya Verbin <ilya.verbin@intel.com>
7151 Kirill Yukhin <kirill.yukhin@intel.com>
7152 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7154 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
7155 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
7156 * config/i386/sse.md (VI4_AVX): New.
7157 (sf2simodelower): Ditto.
7158 (sse2_cvtps2dq): Change to ...
7159 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
7161 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7162 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7163 Sergey Lega <sergey.s.lega@intel.com>
7164 Anna Tikhonova <anna.tikhonova@intel.com>
7165 Ilya Tocar <ilya.tocar@intel.com>
7166 Andrey Turetskiy <andrey.turetskiy@intel.com>
7167 Ilya Verbin <ilya.verbin@intel.com>
7168 Kirill Yukhin <kirill.yukhin@intel.com>
7169 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7171 * config/i386/sse.md (V_512): New.
7173 (vcond<V_512:mode><VF_512:mode>): Ditto.
7174 (vcond<V_512:mode><VI_512:mode>): Ditto.
7175 (vcondu<V_512:mode><VI_512:mode>): Ditto.
7177 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7178 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7179 Sergey Lega <sergey.s.lega@intel.com>
7180 Anna Tikhonova <anna.tikhonova@intel.com>
7181 Ilya Tocar <ilya.tocar@intel.com>
7182 Andrey Turetskiy <andrey.turetskiy@intel.com>
7183 Ilya Verbin <ilya.verbin@intel.com>
7184 Kirill Yukhin <kirill.yukhin@intel.com>
7185 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7187 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
7188 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
7190 (fma<mode>4): Removed condition.
7191 (fms<mode>4): Ditto.
7192 (fnma<mode>4): Ditto.
7193 (fnms<mode>4): Ditto.
7194 (fma4i_fmadd_<mode>): Ditto.
7195 (*fma_fmadd_<mode>): Ditto.
7196 (*fma_fmsub_<mode>): Ditto.
7197 (*fma_fnmadd_<mode>): Ditto.
7198 (*fma_fnmsub_<mode>): Ditto.
7199 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
7200 (*fma_fmaddsub_<mode>): Ditto.
7201 (*fma_fmsubadd_<mode>): Ditto.
7202 (*fmai_fmadd_<mode>): Ditto.
7203 (*fmai_fmsub_<mode>): Ditto.
7204 (*fmai_fnmadd_<mode>): Ditto.
7205 (*fmai_fnmsub_<mode>): Ditto.
7207 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7208 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7209 Sergey Lega <sergey.s.lega@intel.com>
7210 Anna Tikhonova <anna.tikhonova@intel.com>
7211 Ilya Tocar <ilya.tocar@intel.com>
7212 Andrey Turetskiy <andrey.turetskiy@intel.com>
7213 Ilya Verbin <ilya.verbin@intel.com>
7214 Kirill Yukhin <kirill.yukhin@intel.com>
7215 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7217 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
7218 (VI124_256): Changed to ...
7219 (VI124_256_48_512): This.
7220 (ssepackmode): Extended with wider modes.
7221 (<code><mode>3): Changed iterator.
7222 (*avx2_<code><mode>3): Ditto.
7223 (vec_pack_trunc_<mode>): Ditto.
7225 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7226 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7227 Sergey Lega <sergey.s.lega@intel.com>
7228 Anna Tikhonova <anna.tikhonova@intel.com>
7229 Ilya Tocar <ilya.tocar@intel.com>
7230 Andrey Turetskiy <andrey.turetskiy@intel.com>
7231 Ilya Verbin <ilya.verbin@intel.com>
7232 Kirill Yukhin <kirill.yukhin@intel.com>
7233 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7235 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
7236 (VI8F_256_512): Ditto.
7237 (abs<mode>2): Changed iterator.
7238 (avx2_perm<mode>): Changed to ...
7239 (<avx2_avx512f>_perm<mode>): This.
7240 (avx2_perm<mode>_1): Changed to ...
7241 (<avx2_avx512f>_perm<mode>_1): This.
7243 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7244 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7245 Sergey Lega <sergey.s.lega@intel.com>
7246 Anna Tikhonova <anna.tikhonova@intel.com>
7247 Ilya Tocar <ilya.tocar@intel.com>
7248 Andrey Turetskiy <andrey.turetskiy@intel.com>
7249 Ilya Verbin <ilya.verbin@intel.com>
7250 Kirill Yukhin <kirill.yukhin@intel.com>
7251 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7253 * config/i386/sse.md (VI48_AVX512F): New.
7254 (VI48_AVX2): Changed to ...
7255 (VI48_AVX2_48_AVX512F): This.
7256 (avx2_ashrv<mode>): Changed to ...
7257 (<avx2_avx512f>_ashrv<mode>): This.
7258 (avx2_<shift_insn>v<mode>): Changed to ...
7259 (<avx2_avx512f>_<shift_insn>v<mode>): This.
7261 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7262 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7263 Sergey Lega <sergey.s.lega@intel.com>
7264 Anna Tikhonova <anna.tikhonova@intel.com>
7265 Ilya Tocar <ilya.tocar@intel.com>
7266 Andrey Turetskiy <andrey.turetskiy@intel.com>
7267 Ilya Verbin <ilya.verbin@intel.com>
7268 Kirill Yukhin <kirill.yukhin@intel.com>
7269 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7271 * config/i386/sse.md (VI4_AVX512F): New.
7272 (VI8_AVX2_AVX512F): Ditto.
7273 (mul<mode>3): Extended with wider modes.
7274 (*<sse4_1_avx2>_mul<mode>3): Ditto.
7275 (mul<mode>3): Ditto.
7276 (vec_widen_<s>mult_odd_<mode>): Ditto.
7278 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7279 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7280 Sergey Lega <sergey.s.lega@intel.com>
7281 Anna Tikhonova <anna.tikhonova@intel.com>
7282 Ilya Tocar <ilya.tocar@intel.com>
7283 Andrey Turetskiy <andrey.turetskiy@intel.com>
7284 Ilya Verbin <ilya.verbin@intel.com>
7285 Kirill Yukhin <kirill.yukhin@intel.com>
7286 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7288 * config/i386/sse.md (VI2_AVX512F): New.
7289 (VI124_AVX512F): Ditto.
7290 (sseunpackmode): Extended with wider modes.
7291 (sseunpackfltmode): Ditto.
7292 (vec_unpacks_float_hi_<mode>): Ditto.
7293 (vec_unpacks_float_lo_<mode>): Ditto.
7294 (vec_unpacku_float_hi_<mode>): Ditto.
7295 (vec_unpacku_float_lo_<mode>): Ditto.
7296 (vec_unpacks_lo_<mode>): Ditto.
7297 (vec_unpacks_hi_<mode>): Ditto.
7298 (vec_unpacku_lo_<mode>): Ditto.
7299 (vec_unpacku_hi_<mode>): Ditto.
7301 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7302 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7303 Sergey Lega <sergey.s.lega@intel.com>
7304 Anna Tikhonova <anna.tikhonova@intel.com>
7305 Ilya Tocar <ilya.tocar@intel.com>
7306 Andrey Turetskiy <andrey.turetskiy@intel.com>
7307 Ilya Verbin <ilya.verbin@intel.com>
7308 Kirill Yukhin <kirill.yukhin@intel.com>
7309 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7311 * config/i386/i386.md (multdiv): New.
7312 (multdiv_mnemonic): Ditto.
7313 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
7314 (<sse>_vm<multdiv_mnemonic><mode>3): This.
7315 (<sse>_vmdiv<mode>3): Removed.
7317 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7318 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7319 Sergey Lega <sergey.s.lega@intel.com>
7320 Anna Tikhonova <anna.tikhonova@intel.com>
7321 Ilya Tocar <ilya.tocar@intel.com>
7322 Andrey Turetskiy <andrey.turetskiy@intel.com>
7323 Ilya Verbin <ilya.verbin@intel.com>
7324 Kirill Yukhin <kirill.yukhin@intel.com>
7325 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7327 * config/i386/sse.md (V): Extended with wider modes.
7329 (ssehalfvecmode): Ditto.
7331 (ssepackfltmode): Ditto.
7332 (avx_vec_concat<mode>): Ditto.
7333 (V_256_512): New iterator.
7334 (VF2_512_256): Ditto.
7335 (si2dfmode): New attribute.
7336 (si2dfmodelower): Ditto.
7338 (concat_tg_mode): Ditto.
7339 (floatv4siv4df2): Changed to ...
7340 (float<si2dfmodelower><mode>2): This.
7341 (avx_cvtps2pd256): Changed to ...
7342 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
7343 (vec_pack_trunc_v4df): Changed to ...
7344 (vec_pack_trunc_<mode>): This.
7345 (avx_vpermil<mode>): Changed to ...
7346 (<sse2_avx_avx512f>_vpermil<mode>): This.
7347 (<fixsuffix>fix_truncv8dfv8si2): New.
7348 (vec_pack_sfix_trunc_v8df): Ditto.
7349 (avx512f_rndscale<mode>): Ditto.
7350 (avx512f_roundpd512): Ditto.
7351 (vec_pack_ufix_trunc_<mode>): Updated iterator.
7353 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7354 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7355 Sergey Lega <sergey.s.lega@intel.com>
7356 Anna Tikhonova <anna.tikhonova@intel.com>
7357 Ilya Tocar <ilya.tocar@intel.com>
7358 Andrey Turetskiy <andrey.turetskiy@intel.com>
7359 Ilya Verbin <ilya.verbin@intel.com>
7360 Kirill Yukhin <kirill.yukhin@intel.com>
7361 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7363 * config/i386/i386.md (any_fix): New iterator.
7364 (fixsuffix): New attribute.
7365 * config/i386/sse.md (VF1): Extened with wider modes.
7369 (sseintvecmodelower): Ditto.
7370 (ssescalarmode): Ditto.
7371 (ssescalarnum): Ditto.
7373 (ssexmmmode): Ditto.
7374 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
7375 (<sse>_rcp<mode>2): Change iterator.
7376 (rsqrt<mode>2): Ditto.
7377 (<sse>_rsqrt<mode>2): Ditto.
7378 (avx2_vec_dup<mode>): Ditto.
7379 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
7380 (round<mode>2_sfix): Ditto.
7381 (avx2_pbroadcast<mode>): Ditto.
7382 (*andnot<mode>3): Handle XI mode.
7383 (*<code><mode>3): Ditto.
7384 (AVXTOSSEMODE): Removed.
7385 (avx_vpermil<mode>): Changed to ...
7386 (<sse2_avx_avx512f>_vpermil<mode>): This.
7388 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7389 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7390 Sergey Lega <sergey.s.lega@intel.com>
7391 Anna Tikhonova <anna.tikhonova@intel.com>
7392 Ilya Tocar <ilya.tocar@intel.com>
7393 Andrey Turetskiy <andrey.turetskiy@intel.com>
7394 Ilya Verbin <ilya.verbin@intel.com>
7395 Kirill Yukhin <kirill.yukhin@intel.com>
7396 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7398 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
7399 (<sse>_comi): Ditto.
7400 (<sse>_ucomi): Ditto.
7401 (sse_cvtss2siq_2): Ditto.
7402 (sse2_cvtsd2si): Ditto.
7403 (sse2_cvtsd2siq): Ditto.
7404 (sse2_cvttsd2si): Ditto.
7405 (sse2_cvttsd2siq): Ditto.
7406 (<shift_insn><mode>3): Ditto.
7407 (sse2_cvtsi2sdq): Update constraint and prefix.
7408 (sse_cvtsi2ss): Update prefix.
7409 (sse_cvtsi2ssq): Ditto.
7411 2013-10-11 Jakub Jelinek <jakub@redhat.com>
7413 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
7414 ignore internal calls.
7416 2013-10-11 Richard Biener <rguenther@suse.de>
7418 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
7419 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
7420 TREE_OVERFLOW is set.
7422 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
7424 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
7426 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
7427 * gimple.h: Likewise.
7429 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
7431 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
7434 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
7436 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
7437 * configure: Regenerate.
7439 2013-10-11 Jakub Jelinek <jakub@redhat.com>
7441 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
7442 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
7443 formatting fixes, use pp_colon instead of pp_character (..., ':'),
7444 similarly pp_right_paren.
7445 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
7446 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
7447 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
7448 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
7450 (omp_declare_simd_clauses_equal,
7451 omp_remove_redundant_declare_simd_attrs): New functions.
7452 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
7453 (walk_tree_1): Handle new OpenMP 4.0 clauses.
7454 * tree.h (OMP_LOOP_CHECK): Define.
7455 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
7456 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
7457 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
7458 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
7459 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
7460 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
7461 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
7462 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
7463 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
7464 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
7465 OMP_CLAUSE_SIMDLEN_EXPR): Define.
7466 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
7467 (omp_remove_redundant_declare_simd_attrs): New prototype.
7468 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
7469 GIMPLE_OMP_TEAMS): New codes.
7470 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
7471 * omp-low.c (struct omp_context): Add cancel_label and cancellable
7473 (target_nesting_level): New variable.
7474 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
7475 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
7476 for collapse > 1 static schedule unless ordered.
7477 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
7478 (determine_parallel_type): Adjust get_ws_args_for caller.
7479 (install_var_field): Handle mask & 4 for double indirection.
7480 (scan_sharing_clauses): Ignore shared clause on teams construct.
7481 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
7482 (create_omp_child_function): If inside target or declare target
7483 constructs, set "omp declare target" attribute on the child function.
7484 (find_combined_for): New function.
7485 (scan_omp_parallel): Handle combined loops.
7486 (scan_omp_target, scan_omp_teams): New functions.
7487 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
7488 restrictions and set ctx->cancellable for cancellable constructs.
7489 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
7490 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
7491 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
7492 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
7493 (omp_clause_aligned_alignment): New function.
7494 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
7495 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
7496 on teams constructs. Handle user defined reductions and new
7498 (lower_reduction_clauses): Don't set placeholder to address of ref
7499 if it has already the right type.
7500 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
7501 (expand_parallel_call): Use the new non-_start suffixed builtins,
7502 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
7503 and GOMP_parallel_end after the call.
7504 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
7505 (expand_omp_for_init_counts): Handle combined loops.
7506 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
7508 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
7509 end of cancellable loops.
7510 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7511 Likewise. Handle collapse > 1 loops.
7512 (expand_omp_simd): Handle combined loops.
7513 (expand_omp_for): Add inner_stmt argument, adjust callers of
7514 expand_omp_for* functions, use expand_omp_for_static*chunk even
7515 for collapse > 1 unless ordered.
7516 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
7517 of cancellable sections.
7518 (expand_omp_single): Remove need_barrier variable, just rely on
7519 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
7520 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
7521 (expand_omp_atomic_load, expand_omp_atomic_store,
7522 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
7523 (expand_omp_target): New function.
7524 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
7525 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
7526 (build_omp_regions_1): Immediately close region for
7527 GF_OMP_TARGET_KIND_UPDATE.
7528 (maybe_add_implicit_barrier_cancel): New function.
7529 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
7531 (lower_omp_single): Likewise. Add clobber after the barrier.
7532 (lower_omp_taskgroup): New function.
7533 (lower_omp_for): Handle combined loops. Adjust
7534 lower_rec_input_clauses caller. Handle cancellation.
7535 (lower_depend_clauses): New function.
7536 (lower_omp_taskreg): Lower depend clauses. Adjust
7537 lower_rec_input_clauses caller. Add clobber after the call. Handle
7539 (lower_omp_target, lower_omp_teams): New functions.
7540 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
7541 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
7542 and GOMP_cancellation_point calls.
7543 (lower_omp): Fold stmts inside of target region.
7544 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
7545 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
7546 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
7547 (BT_FN_VOID_OMPFN_PTR_UINT,
7548 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
7549 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
7550 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
7551 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
7552 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
7553 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
7554 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
7555 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
7556 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
7557 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
7558 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
7559 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
7560 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
7561 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
7562 BUILT_IN_GOMP_PARALLEL_END.
7563 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
7564 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
7565 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
7566 GF_OMP_FOR_KIND_DISTRIBUTE.
7567 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
7568 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
7569 (dump_gimple_omp_return): Print lhs if it has any.
7570 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
7571 gimple_omp_atomic_seq_cst_p.
7572 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
7573 and GIMPLE_OMP_TEAMS.
7574 * langhooks.c (lhd_omp_mappable_type): New function.
7575 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
7576 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
7578 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
7579 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
7580 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
7581 (struct gimplify_omp_ctx): Add combined_loop field.
7582 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
7583 on stmts inside of target region.
7584 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
7585 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
7586 ORT_TARGET and ORT_TARGET_DATA.
7587 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
7589 (omp_notice_threadprivate_variable): Complain about threadprivate
7590 variables in target region.
7591 (omp_notice_variable): Complain about vars with non-mappable type
7592 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
7593 (omp_check_private): Ignore ORT_TARGET* regions.
7594 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
7595 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
7596 (find_combined_omp_for): New function.
7597 (gimplify_omp_for): Handle gimplification of combined loops.
7598 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
7600 (gimplify_omp_target_update): New function.
7601 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
7602 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
7603 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
7604 (gimplify_body): If fndecl has "omp declare target" attribute, add
7605 implicit ORT_TARGET context around it.
7606 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
7607 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
7608 * tree-nested.c (convert_nonlocal_reference_stmt,
7609 convert_local_reference_stmt, convert_gimple_call): Handle
7610 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7611 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
7612 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
7613 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
7614 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
7615 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
7616 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
7617 BUILT_IN_GOMP_LOOP_END_CANCEL,
7618 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
7619 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
7620 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
7621 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
7622 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
7623 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
7624 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
7625 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
7626 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
7627 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
7628 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
7629 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
7630 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
7631 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
7632 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
7633 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7634 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
7635 gimple_build_omp_teams): New functions.
7636 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
7637 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
7638 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
7639 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7640 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
7641 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
7642 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
7643 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
7644 GF_OMP_ATOMIC_SEQ_CST): New.
7645 (gimple_build_omp_taskgroup, gimple_build_omp_target,
7646 gimple_build_omp_teams): New prototypes.
7647 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
7648 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7649 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
7650 GIMPLE_OMP_SINGLE as end of range.
7651 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
7652 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
7653 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
7654 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
7655 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
7656 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
7657 gimple_omp_target_kind, gimple_omp_target_set_kind,
7658 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
7659 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
7660 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
7661 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
7662 gimple_omp_teams_set_clauses): New inlines.
7663 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
7664 and GIMPLE_OMP_TASKGROUP.
7665 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
7666 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
7667 enum omp_clause_proc_bind_kind): New.
7668 (union omp_clause_subcode): Add depend_kind, map_kind and
7669 proc_bind_kind fields.
7670 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
7671 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7672 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
7673 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
7674 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
7676 2013-10-10 Teresa Johnson <tejohnson@google.com>
7678 * predict.c (tree_estimate_probability): Add new parameter
7679 for estimate_bb_frequencies.
7680 (estimate_bb_frequencies): Add new parameter to force estimation.
7681 (rebuild_frequencies): When max frequency in function is small,
7682 recompute counts from frequencies.
7683 * predict.h (estimate_bb_frequencies): New parameter.
7685 2013-10-10 David Malcolm <dmalcolm@redhat.com>
7687 * ipa-inline.c (ipa_inline): Fix leak of "order" when
7688 optimizations are disabled.
7690 2013-10-10 David Malcolm <dmalcolm@redhat.com>
7692 * coverage.c (coverage_finish): Fix leak of da_file_name.
7694 2013-10-10 Jan Hubicka <jh@suse.cz>
7696 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
7697 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
7700 2013-10-10 Jakub Jelinek <jakub@redhat.com>
7703 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
7704 if any labels are in FALLTHRU_BB, use a special label emitted
7705 immediately after the asm goto insn rather than label_rtx
7707 (expand_asm_stmt): Adjust caller.
7708 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
7709 edge if the last insn in predecessor is a jump with single successor,
7710 but it isn't simplejump_p.
7712 2013-10-10 Richard Biener <rguenther@suse.de>
7714 PR tree-optimization/58656
7715 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
7717 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7719 * gimplify.c: Include expr.h and tm_p.h for targets with special
7720 va-arg padding requirements.
7722 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7724 * tree-flow.h: Move some prototypes to gimple.h.
7725 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
7726 * gimple.h: Relocate some prototypes from tree-flow.h
7727 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
7729 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
7730 (build_va_arg_indirect_ref): Relocate and make static.
7731 (std_gimplify_va_arg_expr): Relocate here.
7732 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
7733 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
7735 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7737 * doc/md.texi: Document the mnemonic attribute.
7739 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7742 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
7743 (set_attr_alternative x ...) when searching for user defined
7746 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7748 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
7750 2013-10-09 Easwaran Raman <eraman@google.com>
7752 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
7753 * cfgexpand.c (defer_stack_allocation): ...use here
7754 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
7756 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
7758 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
7759 (operand_equal_for_value_replacement): New function, extracted from
7760 value_replacement and enhanced to catch more cases.
7761 (value_replacement): Use operand_equal_for_value_replacement.
7763 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7765 * loop-doloop.c (doloop_modify, doloop_optimize): Use
7766 get_max_loop_iterations.
7768 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7770 * config/arm/aarch-common.c (arm_early_load_addr_dep):
7771 Place comment above function.
7773 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7775 * tree-flow.h: Remove all remaining prototypes, enums and structs that
7776 are not related to tree-cfg.c.
7777 * tree-ssa-address.h: New file. Relocate prototypes.
7778 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
7779 (addr_for_mem_ref): New. Combine call to get_address_description and
7780 return addr_for_mem_ref.
7781 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
7782 * tree-ssa-live.h: Adjust prototypes.
7783 * passes.c: Include tree-ssa-live.h.
7784 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
7785 * graphite.c (graphite_transform_loops): Make static.
7786 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7787 make_pass_graphite, pass_data_graphite_transforms,
7788 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
7789 * ipa-pure-const.c (warn_function_noreturn): Make static.
7790 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7791 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7792 Relocate from tree-cfg.c
7793 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
7795 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7796 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7797 Move to ipa-pure-const.c.
7798 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
7799 Relocate from tree-optimize.c.
7800 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
7801 make_pass_fixup_cfg): Move to tree-cfg.c.
7802 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
7803 Relocate some prototypes.
7804 * tree-data-ref.h (tree_check_data_deps) Add prototype.
7805 * tree-dump.c (dump_function_to_file): Remove prototype.
7806 Add tree-flow.h to the include file.
7807 * tree-dump.h: Remove prototype.
7808 * tree-parloops.h: New File. Add prototypes.
7809 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
7810 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
7811 from tree-ssa-loop.c.
7812 * tree-predcom.c (run_tree_predictive_commoning,
7813 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
7814 Relocate here from tree-ssa-loop.c.
7815 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
7816 ssa_name_values.release ().
7817 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
7818 (ssa_name_values): Relocate from tree-flow.h.
7819 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
7820 * tree-ssa-loop.c (run_tree_predictive_commoning,
7821 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
7822 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7823 make_pass_graphite, pass_data_graphite_transforms,
7824 make_pass_graphite_transforms, gate_tree_parallelize_loops,
7825 tree_parallelize_loops, pass_data_parallelize_loops,
7826 make_pass_parallelize_loops): Move to other files.
7827 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
7829 * tree.h: Remove prototypes from tree-address.c.
7831 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7833 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
7834 (struct int_tree_map): Move to tree-hasher.h
7835 (SCALE, LABEL, PERCENT): Move to gimple.h
7836 * tree-flow-inline.h: Delete. Move functions to other files.
7837 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
7838 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
7839 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
7840 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
7841 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
7842 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
7843 * tree-hasher.h: Don't include tree-flow.h.
7844 (struct int_tree_map): Relocate from tree-flow.h.
7845 * tree-sra.c (contains_view_convert_expr_p): Relocate from
7846 tree-flow-inline.h and make static.
7847 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
7849 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
7850 tree-flow-inline.h and make static.
7851 * tree.h (is_global_var, may_be_aliased): Relocate from
7853 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
7854 * value-prof.c: No longer include tree-flow-inline.h.
7855 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
7857 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7859 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
7860 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
7861 (enum move_pos): Move to tree-ssa-loop-im.h
7862 * cfgloop.h: Move some prototypes.
7863 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
7864 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
7865 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
7866 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
7867 (loop_containing_stmt): Relocate from tree-flow-inline.h.
7868 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
7869 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
7870 (enum move_pos): Relocate here.
7871 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
7873 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
7874 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
7875 make_pass_lim): Relocate here from tree-ssa-loop.c.
7876 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
7878 (loop_edge_to_cancel, unloop_loops): Make static.
7879 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7880 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
7881 (tree_complete_unroll, gate_tree_complete_unroll,
7882 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
7883 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7884 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
7885 * tree-ssa-loop-ivopts.c: Remove local prototypes.
7886 (stmt_invariant_in_loop_p): Remove unused function.
7887 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
7888 * tree-ssa-loop-manip.h: New file. Add prototypes.
7889 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
7890 (gcov_type_to_double_int): Move to cfgloop.h.
7891 (double_int_cmp, bound_index,
7892 estimate_numbers_of_iterations_loop): Make static.
7893 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
7894 (max_loop_iterations): Factor out get_max_loop_iterations.
7895 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
7897 * tree-ssa-loop-niter.h: New file. Add prototypes.
7898 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
7899 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
7900 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
7901 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
7902 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
7903 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
7904 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
7905 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
7906 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
7907 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
7908 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7909 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
7910 pass_data_complete_unroll, make_pass_complete_unroll,
7911 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7912 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
7913 tree-ssa-loop-ivcanon.c.
7914 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
7915 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
7916 tree-ssa-loop-prefetch.c.
7917 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
7919 (get_lsm_tmp_name): Relocate and add suffix parameter.
7920 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
7921 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
7922 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
7923 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
7924 (get_estimated_loop_iterations): Factor out accessor from
7925 estimated_loop_iterations in tree-ssa-loop-niter.c.
7926 (get_max_loop_iterations): Factor out accessor from
7927 _max_loop_iterations in tree-ssa-niter.c.
7928 * loop-unroll.c (decide_unroll_constant_iterations,
7929 decide_unroll_runtime_iterations, decide_peel_simple,
7930 decide_unroll_stupid): Use new get_* accessors.
7932 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
7934 PR tree-optimization/20318
7935 * doc/extend.texi (returns_nonnull): New function attribute.
7936 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
7938 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
7939 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
7941 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
7944 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
7945 false if both components are bitfields.
7947 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7949 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
7950 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
7951 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
7952 * config/aarch64/aarch64.h
7953 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
7954 * config/aarch64/aarch64-simd-builtins.def
7955 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
7957 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7959 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
7961 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7963 * config/aarch64/arm_neon.h (vdiv_f64): Added.
7965 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7967 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
7968 (vneg_f64): New intrinsic.
7969 (vneg_s8): Asm replaced with C.
7970 (vneg_s16): Likewise.
7971 (vneg_s32): Likewise.
7972 (vneg_s64): New intrinsic.
7973 (vnegq_f32): Asm replaced with C.
7974 (vnegq_f64): Likewise.
7975 (vnegq_s8): Likewise.
7976 (vnegq_s16): Likewise.
7977 (vnegq_s32): Likewise.
7978 (vnegq_s64): Likewise.
7980 2013-10-09 Renlin Li <Renlin.Li@arm.com>
7982 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
7984 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7986 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
7987 packed stack special handling.
7988 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
7989 back to fixed stack slots for FPRs saved due to stdarg.
7991 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7993 * config/s390/s390.c (s390_frame_info): Restructure function.
7995 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7997 * config/s390/s390.c (struct s390_frame_layout): New field
7999 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
8000 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
8001 regs_ever_clobbered to char*.
8002 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
8003 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
8004 (s390_register_info): Enable FPR save slots. Move/Copy some
8005 functionality into ...
8006 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
8007 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
8009 (s390_frame_info): Do gpr slot allocation here now. stdarg does
8010 not imply a stack frame.
8011 (s390_init_frame_layout): Remove variable clobbered_regs.
8012 (s390_update_register_info): Remove function.
8013 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
8014 cannot be used for register renaming.
8015 (s390_hard_regno_scratch_ok): New function.
8016 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
8017 (s390_initial_elimination_offset): Change offset calculation of
8018 the return address pointer.
8019 (save_gprs): Deal with only r6 being saved from the call-saved regs.
8020 (restore_gprs): Set frame related flag.
8021 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
8022 (s390_emit_prologue): Call s390_register_info instead of
8023 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
8024 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
8025 (s390_optimize_prologue): Call s390_optimize_register_info.
8026 Try to remove also FPR slot save/restore INSNs. Remove frame
8027 related flags from restore INSNs.
8029 2013-10-08 DJ Delorie <dj@redhat.com>
8031 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
8034 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
8035 avoid conflict with the MI use of %c.
8036 * config/rl78/rl78-real.md: change %c to %C throughout.
8037 * config/rl78/rl78-virt.md: Likewise.
8039 2013-10-08 Jan Hubicka <jh@suse.cz>
8041 * config/i386/i386.c (ix86_option_override_internal): Switch
8042 to SSE math for -ffast-math when target ISA supports SSE2.
8044 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
8046 * tree-flow.h: Remove some prototypes.
8047 * tree.h: Remove some protypes, add a couple.
8048 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
8049 using_eh_for_cleanups_p): Add interface routines for front ends.
8050 * tree-eh.h: New file. Add protoptyes.
8051 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
8052 (add_stmt_to_eh_lp_fn): Make static.
8053 (lower_try_finally): Use new using_eh_for_cleanups_p.
8054 * emit-rtl.c: Include tree-eh.h.
8055 * gimple.h: Include tree-eh.h.
8057 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
8059 PR tree-optimization/58480
8060 * tree-vrp.c (infer_nonnull_range): New function.
8061 (infer_value_range): Call infer_nonnull_range.
8063 2013-10-08 Dehao Chen <dehao@google.com>
8065 PR tree-optimization/58619
8066 * tree-inline.c (copy_phis_for_bb): Combine location data
8069 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8072 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
8073 RTX_FRAME_RELATED_P on INSN.
8075 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8077 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
8078 (altivec_expand_vec_perm_const): Call it.
8080 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8082 * config/rs6000/vector.md (mov<mode>): Emit permuted move
8083 sequences for LE VSX loads and stores at expand time.
8084 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
8086 * config/rs6000/rs6000.c (rs6000_const_vec): New.
8087 (rs6000_gen_le_vsx_permute): New.
8088 (rs6000_gen_le_vsx_load): New.
8089 (rs6000_gen_le_vsx_store): New.
8090 (rs6000_gen_le_vsx_move): New.
8091 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
8092 (*vsx_le_perm_load_v4si): New.
8093 (*vsx_le_perm_load_v8hi): New.
8094 (*vsx_le_perm_load_v16qi): New.
8095 (*vsx_le_perm_store_v2di): New.
8096 (*vsx_le_perm_store_v4si): New.
8097 (*vsx_le_perm_store_v8hi): New.
8098 (*vsx_le_perm_store_v16qi): New.
8099 (*vsx_xxpermdi2_le_<mode>): New.
8100 (*vsx_xxpermdi4_le_<mode>): New.
8101 (*vsx_xxpermdi8_le_V8HI): New.
8102 (*vsx_xxpermdi16_le_V16QI): New.
8103 (*vsx_lxvd2x2_le_<mode>): New.
8104 (*vsx_lxvd2x4_le_<mode>): New.
8105 (*vsx_lxvd2x8_le_V8HI): New.
8106 (*vsx_lxvd2x16_le_V16QI): New.
8107 (*vsx_stxvd2x2_le_<mode>): New.
8108 (*vsx_stxvd2x4_le_<mode>): New.
8109 (*vsx_stxvd2x8_le_V8HI): New.
8110 (*vsx_stxvd2x16_le_V16QI): New.
8112 2013-10-07 Renlin Li <Renlin.Li@arm.com>
8114 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
8116 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8118 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
8119 loop to work also for 31bit ABI.
8120 Save the stack pointer for frame_size > 0.
8122 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8124 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
8125 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
8126 constraint letters from expanders.
8127 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
8128 retry count to general_operand.
8129 ("tabort"): Give operand 0 a mode.
8130 ("tabort_1"): Add mode and constraint letter for operand 0.
8131 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
8133 2013-10-04 Jeff Law <law@redhat.com>
8135 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
8137 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
8139 (thread_across_edge): Here. Call it.
8141 2013-10-04 Cary Coutant <ccoutant@google.com>
8143 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
8144 discarding a location list expression (or a piece of one).
8146 2013-10-03 Jan Hubicka <jh@suse.cz>
8148 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
8149 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
8150 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
8152 2013-10-03 Jan Hubicka <jh@suse.cz>
8154 * config/i386/i386.c (ix86_option_override_internal): Do not enable
8155 accumulate-outgoing-args when producing unwind info.
8157 2013-10-03 Wei Mi <wmi@google.com>
8159 * lra-constraints.c (insert_move_for_subreg): New function
8160 extracted from simplify_operand_subreg.
8161 (simplify_operand_subreg): Add reload for paradoxical subreg.
8163 2013-10-03 Rong Xu <xur@google.com>
8165 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
8166 the candidate of builtin_expect such that we should fix the
8167 size/time estimation.
8168 (estimate_function_body_sizes): Do the acutally size/time fix-up
8171 2013-10-03 Rong Xu <xur@google.com>
8173 * predict.c (tree_predict_by_opcode): Get the probability
8174 for builtin_expect from param builtin_expect_probability.
8175 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
8176 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
8177 * doc/invoke.texi: Add documentation for builtin-expect-probability.
8179 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
8182 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
8183 * calls.c (alloca_call_p): Use get_callee_fndecl.
8184 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
8185 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
8187 (vrp_visit_stmt): Remove duplicated code.
8189 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
8191 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
8192 ceildf2, btruncdf2, instead of vsx_* name.
8194 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
8195 iterators to only do V2DF and V4SF here. Move the DF code to
8196 rs6000.md where it is combined with SF mode. Replace <VSv> with
8197 just 'v' since only vector operations are handled with these insns
8198 after moving the DF support to rs6000.md.
8199 (vsx_sub<mode>3): Likewise.
8200 (vsx_mul<mode>3): Likewise.
8201 (vsx_div<mode>3): Likewise.
8202 (vsx_fre<mode>2): Likewise.
8203 (vsx_neg<mode>2): Likewise.
8204 (vsx_abs<mode>2): Likewise.
8205 (vsx_nabs<mode>2): Likewise.
8206 (vsx_smax<mode>3): Likewise.
8207 (vsx_smin<mode>3): Likewise.
8208 (vsx_sqrt<mode>2): Likewise.
8209 (vsx_rsqrte<mode>2): Likewise.
8210 (vsx_fms<mode>4): Likewise.
8211 (vsx_nfma<mode>4): Likewise.
8212 (vsx_copysign<mode>3): Likewise.
8213 (vsx_btrunc<mode>2): Likewise.
8214 (vsx_floor<mode>2): Likewise.
8215 (vsx_ceil<mode>2): Likewise.
8216 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
8217 (vsx_sminsf3): Likewise.
8218 (vsx_fmadf4): Likewise.
8219 (vsx_fmsdf4): Likewise.
8220 (vsx_nfmadf4): Likewise.
8221 (vsx_nfmsdf4): Likewise.
8222 (vsx_cmpdf_internal1): Likewise.
8224 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
8225 simpler to select whether a target has SPE or traditional floating
8226 point support in iterators.
8227 (TARGET_DF_SPE): Likewise.
8228 (TARGET_SF_FPR): Likewise.
8229 (TARGET_DF_FPR): Likewise.
8230 (TARGET_SF_INSN): Macros to say whether floating point support
8231 exists for a given operation for expanders.
8232 (TARGET_DF_INSN): Likewise.
8234 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
8235 combining of SF/DF mode operations, using both traditional and VSX
8243 (abs<mode>2): Combine SF/DF modes using traditional floating point
8244 instructions. Add support for using the upper DF registers with
8245 VSX support, and SF registers with power8-vector support. Update
8246 expanders for operations supported by both the SPE and traditional
8247 floating point units.
8248 (abs<mode>2_fpr): Likewise.
8249 (nabs<mode>2): Likewise.
8250 (nabs<mode>2_fpr): Likewise.
8251 (neg<mode>2): Likewise.
8252 (neg<mode>2_fpr): Likewise.
8253 (add<mode>3): Likewise.
8254 (add<mode>3_fpr): Likewise.
8255 (sub<mode>3): Likewise.
8256 (sub<mode>3_fpr): Likewise.
8257 (mul<mode>3): Likewise.
8258 (mul<mode>3_fpr): Likewise.
8259 (div<mode>3): Likewise.
8260 (div<mode>3_fpr): Likewise.
8261 (sqrt<mode>3): Likewise.
8262 (sqrt<mode>3_fpr): Likewise.
8263 (fre<Fs>): Likewise.
8264 (rsqrt<mode>2): Likewise.
8265 (cmp<mode>_fpr): Likewise.
8266 (smax<mode>3): Likewise.
8267 (smin<mode>3): Likewise.
8268 (smax<mode>3_vsx): Likewise.
8269 (smin<mode>3_vsx): Likewise.
8270 (negsf2): Delete SF operations that are merged with DF.
8277 (fmasf4_fpr): Likewise.
8278 (fmssf4_fpr): Likewise.
8279 (nfmasf4_fpr): Likewise.
8280 (nfmssf4_fpr): Likewise.
8281 (sqrtsf2): Likewise.
8282 (rsqrtsf_internal1): Likewise.
8283 (smaxsf3): Likewise.
8284 (sminsf3): Likewise.
8285 (cmpsf_internal1): Likewise.
8286 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
8287 (negdf2): Delete DF operations that are merged with SF.
8289 (nabsdf2): Likewise.
8295 (rsqrtdf_internal1): Likewise.
8296 (fmadf4_fpr): Likewise.
8297 (fmsdf4_fpr): Likewise.
8298 (nfmadf4_fpr): Likewise.
8299 (nfmsdf4_fpr): Likewise.
8300 (sqrtdf2): Likewise.
8301 (smaxdf3): Likewise.
8302 (smindf3): Likewise.
8303 (cmpdf_internal1): Likewise.
8304 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
8305 (btrunc<mode>2): Delete separate expander, and combine with the
8306 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
8307 (btrunc<mode>2_fpr): Likewise.
8308 (ceil<mode>2): Likewise.
8309 (ceil<mode>2_fpr): Likewise.
8310 (floor<mode>2): Likewise.
8311 (floor<mode>2_fpr): Likewise.
8312 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
8313 Add support for using the upper registers with VSX and
8314 power8-vector. Move insns to be closer to the define_expands. On
8315 VSX systems, prefer the traditional form of FMA over the VSX
8316 version, since the traditional form allows the target not to
8317 overlap with the inputs.
8318 (fms<mode>4_fpr): Likewise.
8319 (nfma<mode>4_fpr): Likewise.
8320 (nfms<mode>4_fpr): Likewise.
8322 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8323 Richard Earnshaw <richard.earnshaw@arm.com>
8325 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
8326 (struct mult_cost_table): Likewise.
8327 (struct mem_cost_table): Likewise.
8328 (struct fp_cost_table): Likewise.
8329 (struct vector_cost_table): Likewise.
8330 (cpu_cost_table): Likewise.
8331 * config/arm/arm.opt (mold-rts-costs): New option.
8332 (mnew-generic-costs): Likewise.
8333 * config/arm/arm.c (generic_extra_costs): New table.
8334 (cortexa15_extra_costs): Likewise.
8335 (arm_slowmul_tune): Use NULL as new costs.
8336 (arm_fastmul_tune): Likewise.
8337 (arm_strongarm_tune): Likewise.
8338 (arm_xscale_tune): Likewise.
8339 (arm_9e_tune): Likewise.
8340 (arm_v6t2_tune): Likewise.
8341 (arm_cortex_a5_tune): Likewise.
8342 (arm_cortex_a9_tune): Likewise.
8343 (arm_v6m_tune): Likewise.
8344 (arm_fa726te_tune): Likewise.
8345 (arm_cortex_a15_tune): Use cortex15_extra_costs.
8346 (arm_cortex_tune): Use generict_extra_costs.
8347 (shifter_op_p): New function.
8348 (arm_unspec_cost): Likewise.
8349 (LIBCALL_COST): Define.
8350 (arm_new_rtx_costs): New function.
8351 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
8352 table is available. Use old costs otherwise unless mnew-generic-costs
8354 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
8355 (cpu_cost_table): Declare.
8357 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
8360 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
8361 (*subs_mul_imm_<mode>)
8362 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
8363 (*sub_<shift>_<mode>)
8364 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
8365 Remove k constraint.
8367 2013-10-03 Ian Bolton <ian.bolton@arm.com>
8369 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
8371 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
8373 2013-10-02 Teresa Johnson <tejohnson@google.com>
8375 * predict.c (probably_never_executed): New function.
8376 (probably_never_executed_bb_p): Invoke probably_never_executed.
8377 (probably_never_executed_edge_p): Ditto.
8378 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
8379 Treat profile insanities conservatively.
8381 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
8383 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
8385 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
8387 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
8388 use it. Use smaller increase for scratch. Don't increase reject
8389 for early clobber scratch.
8390 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
8391 setting eliminated regs except setting fp from hfp.
8392 (lra_eliminate): Check lra_insn_recog_data on NULL.
8394 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
8397 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
8398 setting -mvsx-timode by default until the underlying problem is fixed.
8399 (RS6000_CPU, power7 defaults): Likewise.
8401 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8403 * config/x-linux (host-linux.o): Remove header dependencies.
8404 Use $(COMPILE) and $(POSTCOMPILE).
8405 * config/t-linux-android (linux-android.o): Ditto.
8407 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8409 * Makefile.in (expmed.o-warn): Remove.
8411 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8413 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
8414 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
8416 2013-10-02 Teresa Johnson <tejohnson@google.com>
8418 * dojump.c (do_jump_1): Divide probability between
8419 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
8421 2013-10-02 Tom Tromey <tromey@redhat.com>
8423 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
8425 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8427 * tree-flow.h: Remove some prototypes.
8428 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
8429 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
8430 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
8431 mark_virtual_phi_result_for_renaming): Relocate here.
8432 * tree-into-ssa.h: Add prototypes.
8433 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
8434 single_pred_before_succ_order.
8435 (blocks_in_phiopt_order): Rename and move to cfganal.c.
8436 (nonfreeing_call_p) Move to gimple.c.
8437 * cfganal.c (single_pred_before_succ_order): Move and renamed from
8439 * basic-block.h (single_pred_before_succ_order): Add prototype.
8440 * gimple.c (nonfreeing_call_p): Relocate here.
8441 * gimple.h: Add prototype.
8442 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
8443 * tree-ssa-dom.h: New file. Relocate prototypes here.
8444 * tree-ssa.h: Include tree-ssa-dom.h.
8446 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8448 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
8449 Use $(COMPILE) and $(POSTCOMPILE).
8451 * config/alpha/x-alpha (driver-alpha.o): Ditto.
8453 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8455 * tree-flow.h: Remove some prototypes.
8456 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
8457 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
8458 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
8459 propagate_tree_value*): Move from here to...
8460 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
8461 propagate_tree_value*): Relocate here.
8462 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
8463 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
8464 * gimple-fold.c: Remove gimple-fold.h from include list.
8465 * tree-vrp.c: Remove gimple-fold.h from include list.
8466 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
8467 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
8468 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
8469 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
8470 * sese.c: Add tree-ssa-propagate.h to include list.
8472 2013-10-02 Richard Biener <rguenther@suse.de>
8474 * tree-loop-distribution.c: Include tree-vectorizer.h for
8476 (enum partition_kind): Remove PKIND_REDUCTION.
8477 (struct partition_s): Remove has_writes member, add reduction_p member.
8478 (partition_alloc): Adjust.
8479 (partition_builtin_p): Likewise.
8480 (partition_has_writes): Remove.
8481 (partition_reduction_p): New function.
8482 (partition_merge_into): Likewise.
8483 (generate_code_for_partition): Commonize builtin partition
8485 (rdg_cannot_recompute_vertex_p): Remove.
8486 (already_processed_vertex_p): Likewise.
8487 (rdg_flag_vertex): Do not set has_writes.
8488 (classify_partition): Adjust.
8489 (rdg_build_partitions): Do not set has_writes, treat all
8490 partitions as useful.
8491 (distribute_loop): Record number of library calls generated. Adjust.
8492 (tree_loop_distribution): Report number of loops and library
8493 calls generated as opt-info.
8495 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8497 * tree-flow.h: Include new .h files. Move prototypes.
8498 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
8499 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
8500 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
8501 * tree-pretty-print.h: Add prototypes from tree-flow.h.
8502 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
8503 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
8504 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
8505 * tree.h (get_ref_base_and_extent): Move prototype out.
8506 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
8508 * gimple-low.h: New File. Add prototypes from tree-flow.h.
8509 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
8510 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
8511 * tree-scalar-evolution.c: Include tree.h.
8512 * sese.c: Include tree.h.
8513 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
8514 * dwarf2out.c: Include tree-dfa.h.
8515 * tree-chrec.c: Include tree.h.
8516 * tree-data-ref.c: Include tree.h.
8518 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
8520 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
8523 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8525 * config/t-sol2 (sol2-c.o): Remove header dependencies.
8526 Use $(COMPILE) and $(POSTCOMPILE).
8527 (sol2-cxx.o): Likewise.
8528 (sol2-stubs.o): Likewise.
8530 * config/x-solaris (host-solaris.o): Likewise.
8532 * config/sparc/t-sparc (sparc.o): Remove.
8533 (sparc-c.o): Remove header dependencies.
8534 Use $(COMPILE) and $(POSTCOMPILE).
8535 * config/sparc/x-sparc: Likewise.
8537 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
8539 * config/arc/arc-opts.h: Add 2013 to Copyright years.
8540 * config/arc/arc700.md: Likewise.
8541 * config/arc/arc-modes.def: Likewise.
8542 * config/arc/arc-simd.h: Likewise.
8543 * config/arc/t-arc-uClibc: Likewise.
8544 * config/arc/t-arc-newlib: Likewise.
8546 2013-10-02 Renlin Li <renlin.li@arm.com>
8548 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
8550 (aarch64_expand_epilogue): Likewise.
8552 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8553 Yufeng Zhang <yufeng.zhang@arm.com>
8555 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
8557 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
8558 'base_in' represent a conversion and legal_cast_p_1 holds; set
8559 'base_in' with the returned value from get_unwidened.
8561 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8563 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
8564 plus_constant is not used.
8566 2013-10-01 Wei Mi <wmi@google.com>
8568 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
8569 * config/i386/i386.md: Add define_peephole2 to
8570 break partial reg stall for cvtss2sd/cvtsd2ss.
8572 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8574 * config/arc/arc.c (pass_arc_ifcvt::clone):
8575 Update for ctxt_ -> m_ctxt change.
8577 2013-10-01 Jeff Law <law@redhat.com>
8579 * tree-ssa-threadupdate.c (struct redirection_data): Delete
8580 outgoing_edge and intermediate_edge fields. Instead store the path.
8581 (redirection_data::hash): Hash on the last edge's destination index.
8582 (redirection_data::equal): Check the entire thread path.
8583 (lookup_redirectio_data): Corresponding changes.
8584 (create_edge_and_update_destination_phis): Likewise.
8585 (thread_single_edge): Likewise.
8587 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8588 Diego Novillo <dnovillo@google.com>
8590 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
8591 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
8592 (vld32wh_insn, vld32wl_insn): Delete commented-out old
8593 versions of these patterns.
8595 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
8596 (__builtin_arc_aligned): Likewise.
8598 * config/arc/arc.md: Expand adc_0 comment stating the intended
8599 purpose and why it isn't ready.
8600 Replace commented out call_value_via_label_mixed with a
8601 plain comment about bl_s.
8603 * config/arc/arc.c (stdio.h): Don't include directly.
8604 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
8605 Remove if (1) condition.
8606 (arc_encode_section_info): Fix comment.
8608 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8610 * config/arc/arc.c (arc_conditional_register_usage):
8611 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
8612 Also set reg_alloc_order for DMA config regs.
8614 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8615 Jeremy Bennett <jeremy.bennett@embecosm.com>
8617 * doc/install.texi (--with-cpu): Mention ARC.
8618 (arc-*-elf32): New paragraph.
8619 (arc-linux-uclibc): Likewise.
8620 * doc/md.texi (Machine Constraints): Add ARC part.
8621 * doc/invoke.texi: (menu): Add ARC Options.
8622 (Machine Dependent Options) <ARC Options>: Add synopsis.
8623 (node ARC Options): Add.
8624 * doc/extend.texi (long_call / short_call attribute): Add ARC.
8625 (ARC Built-in Functions): New section defining
8626 generic ARC built-in functions.
8627 (ARC SIMD Built-in Functions): New section defining SIMD specific
8629 (Declaring Attributes of Functions): Extended
8630 description of short_call and long_call attributes for ARC and
8631 added index entries.
8633 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
8634 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
8635 Joern Rennecke <joern.rennecke@embecosm.com>
8636 Muhammad Khurram Riaz <khurram.riaz@arc.com>
8637 Brendan Kehoe <brendan@zen.org>
8638 Michael Eager <eager@eagercon.com>
8639 Simon Cook <simon.cook@embecosm.com>
8640 Jeremy Bennett <jeremy.bennett@embecosm.com>
8642 * config/arc, common/config/arc: New directories.
8644 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8645 Brendan Kehoe <brendan@zen.org>
8646 Simon Cook <simon.cook@embecosm.com>
8648 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
8650 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
8652 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
8653 * tree-ssa-coalesce.h: New. Move prototype to here.
8654 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
8655 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
8656 (gimple_can_coalesce_p): Move to...
8657 * gimple.c (gimple_can_coalesce_p): Here.
8659 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
8661 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
8662 (dump_decl_set): Move to gimple.c.
8663 * gimple.h: Don't include tree-ssa-operands.h.
8664 (dump_decl_set): Add prototype.
8665 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
8666 Move to gimple-ssa.h.
8667 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
8668 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
8669 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
8670 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
8671 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
8673 (dump_decl_set): Relocate here.
8674 * gimple-ssa.h: New file.
8675 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
8676 Relocate from gimple.h.
8677 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
8678 * tree-ssa-operands.c (swap_ssa_operands): Rename from
8679 swap_tree_operands and remove non-ssa path.
8680 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
8681 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
8683 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
8684 vect_is_simple_reduction_1): Use swap_ssa_operands.
8685 * tree-flow.h: Move various prototypes to tree-phinodes.h.
8686 (enum need_phi_state): Move to tree-into-ssa.c.
8687 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
8688 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
8689 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
8690 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
8691 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
8692 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
8693 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
8694 num_imm_uses): Move to ssa-iterators.h.
8695 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
8696 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
8698 (op_iter_done, op_iter_next_def, op_iter_next_tree,
8699 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
8700 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
8701 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
8702 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
8703 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
8704 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
8705 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
8706 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
8707 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
8708 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
8709 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
8710 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
8711 (set_phi_nodes): Move to tree-phinodes.h.
8712 * tree-ssa-operands.h (enum ssa_op_iter_type,
8713 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
8714 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
8715 (dump_decl_set): Remove prototype.
8716 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
8717 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
8718 (set_phi_nodes): Relocate from tree-flow-inline.h.
8719 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
8721 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
8722 include list. Temporarily add gimple.h to include list.
8723 * ssa-iterators.h: New file.
8724 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
8725 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
8726 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
8727 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
8728 Relocate from tree-ssa-operands.h.
8729 (delink_imm_use, link_imm_use_to_list, link_imm_use,
8730 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
8731 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
8732 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
8733 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
8734 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
8735 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
8736 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
8737 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
8738 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
8739 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
8740 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
8741 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
8742 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
8743 Relocate from tree-flow-inline.h.
8744 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
8746 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
8749 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
8750 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
8751 Insert '\t' to output template.
8752 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
8753 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
8754 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
8755 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
8757 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
8759 * doc/install.texi (Host/target specific installation notes for GCC):
8760 Put @anchor before @heading.
8761 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
8762 Use @email for email addresses.
8764 2013-10-01 Jeff Law <law@redhat.com>
8766 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
8767 a vec. Only delete the path if we create one without successfully
8768 registering a jump thread.
8769 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
8771 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
8772 (paths): New vector of jump threading paths.
8773 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
8774 (THREAD_PATH): New accessor macro for the entire thread path.
8775 (lookup_redirection_data): Get intermediate and final outgoing edge
8776 from the thread path.
8777 (create_edge_and_update_destination_phis): Copy the threading path.
8778 (ssa_fix_duplicate_block_edges): Get edges and block types from the
8779 jump threading path.
8780 (ssa_redirect_edges): Get edges and block types from the jump threading
8781 path. Free the path vector.
8782 (thread_block): Get edges from the jump threading path. Look at the
8783 entire path to see if we thread to a loop exit. If we cancel a jump
8784 thread request, then free the path vector.
8785 (thread_single_edge): Get edges and block types from the jump threading
8786 path. Free the path vector.
8787 (thread_through_loop_header): Get edges and block types from the jump
8788 threading path. Free the path vector.
8789 (mark_threaded_blocks): Iterate over the vector of paths and store
8790 the path on the appropriate edge. Get edges and block types from the
8791 jump threading path.
8792 (mark_threaded_blocks): Get edges and block types from the jump
8793 threading path. Free the path vector.
8794 (thread_through_all_blocks): Use the vector of paths rather than
8795 a vector of 3-edge sets.
8796 (register_jump_thread): Accept pointer to a path vector rather
8797 than the path vector itself. Store the path vector for later use.
8800 2013-10-01 Jakub Jelinek <jakub@redhat.com>
8801 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8804 * config/s390/s390.c (s390_split_branches): Modify check for table
8806 (s390_chunkify_start): Rearrange table jump insn check in order to
8807 deal with compare and branch insns correctly.
8809 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
8813 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
8814 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
8815 define_insn_and_split.
8816 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
8817 (shiftsi3_compare): New pattern.
8819 * config/arm/unspecs.md (UNSPEC_RRX): New.
8821 2013-10-01 Alan Modra <amodra@gmail.com>
8823 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
8824 casing inout operands.
8826 2013-10-01 Richard Biener <rguenther@suse.de>
8828 PR tree-optimization/58553
8829 * tree-loop-distribution.c (struct partition_s): Add niter member.
8830 (classify_partition): Populate niter member for the partition
8831 and properly identify whether the relevant store happens before
8832 or after the loop exit.
8833 (generate_memset_builtin): Use niter member from the partition.
8834 (generate_memcpy_builtin): Likewise.
8836 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8838 * vec.h (vec_prefix, vec): Prefix member names with "m_".
8839 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
8841 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8843 * basic-block.h (edge_list): Prefix member names with "m_".
8844 * context.h (context): Likewise.
8845 * domwalk.h (dom_walker): Likewise.
8846 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
8847 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
8848 * hash-table.h (hash_table): Likewise.
8849 * machmode.h (bit_field_mode_iterator): Likewise.
8850 * pass_manager.h (pass_list): Likewise.
8851 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
8852 * tree-pass.h (pass_data): Likewise.
8853 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
8854 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
8855 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
8856 * asan.c (pass_data_asan): Update accordingly.
8857 * cfganal.c (control_dependences::find_control_dependence): Likewise.
8858 (control_dependences::control_dependences): Likewise.
8859 (control_dependences::~control_dependences): Likewise.
8860 (control_dependences::~control_dependences): Likewise.
8861 (control_dependences::get_edges_dependent_on): Likewise.
8862 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
8863 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
8864 * context.c (gcc::context::context): Likewise.
8865 * cprop.c (pass_rtl_cprop::clone): Likewise.
8866 * domwalk.c (dom_walker::walk): Likewise.
8867 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
8868 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
8869 * mode-switching.c (pass_mode_switching::clone): Likewise.
8870 * passes.c (opt_pass::opt_pass): Likewise.
8871 (pass_manager::pass_manager): Likewise.
8872 * predict.c (pass_strip_predict_hints::clone): Likewise.
8873 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
8874 (pass_split_all_insns::clone): Likewise.
8875 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
8877 (bit_field_mode_iterator::next_mode): Likewise.
8878 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
8879 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
8880 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
8881 * tree-complex.c (pass_lower_complex::clone): Likewise.
8882 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
8883 * tree-object-size.c (pass_object_sizes::clone): Likewise.
8884 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
8885 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
8886 (pass_fold_builtins::clone): Likewise.
8887 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
8888 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
8889 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
8890 (pass_cd_dce::clone): Likewise.
8891 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
8892 (pass_phi_only_cprop::clone): Likewise.
8893 * tree-ssa-dse.c (pass_dse::clone): Likewise.
8894 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
8895 * tree-ssa-loop.c (pass_lim::clone): Likewise.
8896 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
8897 * tree-ssa-pre.c (pass_fre::clone): Likewise.
8898 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
8899 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
8900 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
8901 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
8902 * tree-vrp.c (pass_vrp::clone): Likewise.
8903 * tsan.c (pass_tsan::clone): Likewise.
8905 2013-09-30 Jakub Jelinek <jakub@redhat.com>
8908 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
8909 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
8912 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
8913 optimization, punt if sign_bit_p looked through any zero extension.
8915 2013-09-30 Teresa Johnson <tejohnson@google.com>
8917 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
8918 Update redirected out edge count in joiner case.
8919 (ssa_redirect_edges): Common the joiner and non-joiner cases
8920 so that joiner case gets profile updates.
8922 2013-09-30 Richard Biener <rguenther@suse.de>
8924 PR tree-optimization/58554
8925 * tree-loop-distribution.c (classify_partition): Require
8926 unconditionally executed stores for memcpy and memset recognition.
8927 (tree_loop_distribution): Calculate dominance info.
8929 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
8931 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
8932 (NO_PROFILE_COUNTERS): Likewise.
8933 (PROFILE_HOOK): Likewise.
8934 (FUNCTION_PROFILER): Likewise.
8935 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
8937 2013-09-30 Iain Sandoe <iain@codesourcery.com>
8939 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
8940 calls and defines in TARGET_MACHO conditional.
8941 (load_macho_picbase_di): Likewise.
8942 (reload_macho_picbase): Likewise.
8943 (reload_macho_picbase_si): Likewise.
8944 (reload_macho_picbase_di): Likewise.
8945 (nonlocal_goto_receiver): Likewise.
8947 2013-09-30 Nick Clifton <nickc@redhat.com>
8949 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
8950 that use the MSP430X ISA.
8951 (msp430_option_override): Scan -mmcu command line option for any
8952 MCU name that supports the MSP430X ISA.
8953 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
8954 -mmcu options which enable the MSP430X ISA.
8956 2013-09-30 Richard Biener <rguenther@suse.de>
8959 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
8960 before looking for setjmp-like calls.
8962 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8965 * config/darwin-protos.h (machopic_get_function_picbase): New.
8966 * config/darwin.c (machopic_get_function_picbase): New.
8967 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
8968 label for a new func. (load_macho_picbase_di): Likewise.
8969 (reload_macho_picbase): New expand.
8970 (reload_macho_picbase_si): New insn.
8971 (reload_macho_picbase_di): New insn.
8972 (nonlocal_goto_receiver): New define and split.
8973 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
8974 (unspecv enum): Add UNSPECV_NLGR.
8976 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8978 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
8979 that altivec registers are correctly sized on Darwin.
8981 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8983 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
8984 darwin-driver.o): Use COMPILE and POSTCOMPILE.
8985 * config/x-darwin (host-darwin.o): Likewise.
8986 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
8987 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
8988 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
8990 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
8992 * doc/invoke.texi: Fix usage of @tie{} command.
8994 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
8996 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
8998 2013-09-28 Jan Hubicka <jh@suse.cz>
9000 * config/i386/x86-tune.def: Add documentation for each of the options;
9003 2013-09-28 Jan Hubicka <jh@suse.cz>
9005 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
9007 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
9008 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
9009 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
9010 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
9011 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
9013 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
9015 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
9016 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
9017 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
9018 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
9019 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
9020 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
9021 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
9022 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
9023 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
9024 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
9025 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
9026 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
9027 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
9028 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
9029 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
9030 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
9031 gimple.h, godump.c, graphite-clast-to-gimple.c,
9032 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
9033 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
9034 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
9035 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
9036 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
9037 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
9038 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
9039 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
9040 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
9041 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
9042 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
9043 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
9044 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
9045 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
9046 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
9047 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
9048 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
9049 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
9050 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
9051 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
9052 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
9053 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
9054 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
9055 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
9056 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
9057 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
9058 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
9059 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
9060 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
9061 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
9062 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
9063 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
9064 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
9065 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
9066 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
9067 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
9068 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
9069 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
9070 whitespace before "(".
9072 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
9074 * expr.h (extract_bit_field): Remove packedp parameter.
9075 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
9076 from forward declaration.
9077 (store_split_bit_field): Remove packedp arg from calls to
9078 extract_fixed_bit_field.
9079 (extract_bit_field_1): Remove packedp parameter and packedp
9080 argument from recursive calls and calls to extract_fixed_bit_field.
9081 (extract_bit_field): Remove packedp parameter and corresponding
9082 arg to extract_bit_field_1.
9083 (extract_fixed_bit_field): Remove packedp parameter. Remove code
9085 (extract_split_bit_field): Remove packedp arg from call to
9086 extract_fixed_bit_field.
9087 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
9088 (copy_blkmode_from_reg): Likewise.
9089 (copy_blkmode_to_reg): Likewise.
9090 (read_complex_part): Likewise.
9091 (store_field): Likewise.
9092 (expand_expr_real_1): Likewise.
9093 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
9094 to extract_bit_field.
9095 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
9096 call to extract_bit_field.
9097 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
9098 call to extract_bit_field.
9099 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
9100 and special packedp behavior from -fstrict-volatile-bitfields
9103 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9105 * lra-eliminations.c (init_elim_table): Guard value_p.
9107 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
9109 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
9110 DFmode, DImode, and SFmode in the upper VSX registers based on the
9111 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
9112 if -mpower8-vector. Combine -mvsx-timode handling with the rest
9113 of the VSX register handling.
9115 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
9117 (zero_extendsidi2_lfiwzx): Add support for loading into the
9118 Altivec registers with -mpower8-vector. Use wu/wv constraints to
9119 only do VSX memory options on Altivec registers.
9120 (extendsidi2_lfiwax): Likewise.
9121 (extendsfdf2_fpr): Likewise.
9122 (mov<mode>_hardfloat, SF/SD modes): Likewise.
9123 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
9124 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
9125 (movdi_internal64): Likewise.
9127 2013-09-27 Xinliang David Li <davidxl@google.com>
9129 * opts.c (finish_options): Adjust parameters
9130 according to vect cost model.
9131 (common_handle_option): Set dynamic vect cost
9133 targhooks.c (default_add_stmt_cost): Compute stmt cost
9135 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
9136 Use helper function.
9137 * tree-vectorizer.h (unlimited_cost_model): New function.
9138 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
9139 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
9141 (vect_enhance_data_refs_alignment): Ditto.
9142 * flag-types.h: New enum.
9143 * common/config/i386/i386-common.c (ix86_option_init_struct):
9144 No need to initialize vect_cost_model flag.
9145 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
9148 2013-09-27 Diego Novillo <dnovillo@google.com>
9150 * gimple.h (enum ssa_mode): Remove.
9152 2013-09-27 Paulo Matos <pmatos@broadcom.com>
9154 * cfgloop.h (number_of_loops): Fix typo in check for null.
9156 2013-09-27 Jakub Jelinek <jakub@redhat.com>
9159 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
9160 are children of outermost saved_cfun's loop, and set it up to
9161 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
9162 if loop != loop0 and SESE region contains bbs that belong to loop0.
9164 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
9166 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
9167 (binary_scale_code_p, get_base_term, get_index_term): New functions.
9168 (set_address_segment, set_address_base, set_address_index)
9169 (set_address_disp): Accept the argument unconditionally.
9170 (baseness): Remove must_be_base_p and must_be_index_p checks.
9171 (decompose_normal_address): Classify as much as possible in the
9174 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
9176 * cse.c (count_reg_usage): Handle INT_LIST.
9177 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
9178 * reginfo.c (reg_scan_mark_refs): Likewise.
9179 * reload1.c (eliminate_regs_1): Likewise.
9181 2013-09-27 Iain Sandoe <iain@codesourcery.com>
9184 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
9187 2013-09-27 Richard Biener <rguenther@suse.de>
9189 PR tree-optimization/58459
9190 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
9191 restriction not propagating into loops.
9193 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
9195 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
9196 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
9197 * doc/tree-ssa.texi (Walking use-def chains): Delete.
9199 2013-09-26 Richard Biener <rguenther@suse.de>
9201 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
9203 2013-09-26 Richard Biener <rguenther@suse.de>
9205 * alias.h (component_uses_parent_alias_set): Rename to ...
9206 (component_uses_parent_alias_set_from): ... this.
9207 * alias.c (component_uses_parent_alias_set): Rename to ...
9208 (component_uses_parent_alias_set_from): ... this and return
9210 (reference_alias_ptr_type_1): Use the result from
9211 component_uses_parent_alias_set_from instead of stripping
9212 components one at a time.
9213 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
9215 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
9217 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
9218 Move prototypes to...
9219 * tree-ssa-ter.h: New File. Move prototypes here.
9220 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
9221 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
9222 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
9223 from is_replaceable_p.
9224 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
9225 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
9226 (process_replaceable): Use ter_is_replaceable_p.
9227 (find_replaceable_in_bb): Use ter_is_replaceable_p.
9228 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
9229 newly refactored ssa_is_replaceable_p.
9230 * cfgexpand.c: Include tree-outof-ssa.h.
9231 * ssaexpand.h: Delete.
9233 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
9235 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
9236 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
9238 (create_gimple_tmp): Delete.
9239 (get_expr_type, build_assign, build_type_cast): Move to...
9240 * gimple-builder.c: New File.
9241 (get_expr_type): Relocate from gimple.c.
9242 (build_assign, build_type_cast): Change to only create ssanames.
9243 * gimple.h: Move prototypes to...
9244 * gimple-builder.h: New File. Here.
9245 * tree-ssa.h: And here.
9246 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
9247 count_uses_and_derefs): Relocate from gimple.c.
9248 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
9249 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
9250 * tree-ssa-math-opts (execute_cse_reciprocals): Use
9251 gimple_replace_ssa_lhs.
9252 * asan.c: Include gimple-builder.h.
9253 * Makefile.in: Add gimple-builder.o.
9255 2013-09-26 Richard Biener <rguenther@suse.de>
9257 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
9258 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
9259 (loe_visit_block): Use gcc_checking_assert.
9260 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
9261 gimple_assign_ssa_name_copy_p.
9262 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
9265 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
9267 * config/rs6000/t-rs6000 (rs6000.o): Remove.
9268 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
9270 2013-09-26 Richard Biener <rguenther@suse.de>
9272 PR tree-optimization/58539
9273 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
9274 the fact that debug statements are not taking part in loop-closed
9277 2013-09-26 Nick Clifton <nickc@redhat.com>
9279 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
9280 time warning message.
9281 (msp430_print_operand_raw): Delete unused letter parameter.
9282 (TARGET_PRINT_OPERAND_ADDRESS): Define.
9283 (msp430_print_operand_address): New function.
9284 (msp430_print_operand): Move address printing code from here to
9286 * config/msp430/msp430.md (movsipsi2): Add comment in generated
9288 (zero_extendpsisi2): Likewise.
9289 (extendpsisi2): New pattern.
9290 (andneghi3): New pattern.
9292 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
9294 * config/aarch64/aarch64.opt (mlra): New option.
9295 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
9296 (TARGET_LRA_P): Define.
9298 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
9300 * expr.c (expand_assignment): Remove obsolete comment.
9302 2013-09-25 Jeff Law <law@redhat.com>
9304 * tree-flow.h (thread_through_all_blocks): Prototype moved into
9305 tree-ssa-threadupdate.h.
9306 (register_jump_thread): Similarly.
9307 * tree-ssa-threadupdate.h: New header file.
9308 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
9309 * tree-vrp.c: Likewise.
9310 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
9311 (thread_around_empty_blocks): Change type of path vector argument to
9312 an edge,type pair from just an edge. Initialize both elements when
9313 appending to a jump threading path. Tweak references to elements
9315 (thread_across_edge): Similarly. Release memory for the elements
9317 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
9318 (dump_jump_thread_path): New function broken out from
9319 register_jump_thread.
9320 (register_jump_thread): Use dump_jump_thread_path. Change type of
9321 path vector entries. Search the path for NULL edges and dump
9322 the path if one is found. Tweak the conversion of path to 3-edge
9323 form to use the block copy type information embedded in the path.
9325 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
9327 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
9329 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
9330 Vladimir Makarov <vmakarov@redhat.com>
9332 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
9333 from the least significant bit.
9334 (strip_address_mutations): Add bitfield operations handling.
9335 (must_be_index_p): Add shifting and rotate operations handling.
9336 (set_address_base): Use must_be_base_p predicate.
9337 (set_address_index): Use must_be_index_p predicate.
9339 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
9340 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9341 Sergey Lega <sergey.s.lega@intel.com>
9342 Anna Tikhonova <anna.tikhonova@intel.com>
9343 Ilya Tocar <ilya.tocar@intel.com>
9344 Andrey Turetskiy <andrey.turetskiy@intel.com>
9345 Ilya Verbin <ilya.verbin@intel.com>
9346 Kirill Yukhin <kirill.yukhin@intel.com>
9347 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9349 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
9351 (ix86_expand_vector_move_misalign): Support new unaligned load and
9352 stores and use new names.
9353 (CODE_FOR_sse2_storedqu): Rename to ...
9354 (CODE_FOR_sse2_storedquv16qi): ... this.
9355 (CODE_FOR_sse2_loaddqu): Rename to ...
9356 (CODE_FOR_sse2_loaddquv16qi): ... this.
9357 (CODE_FOR_avx_loaddqu256): Rename to ...
9358 (CODE_FOR_avx_loaddquv32qi): ... this.
9359 (CODE_FOR_avx_storedqu256): Rename to ...
9360 (CODE_FOR_avx_storedquv32qi): ... this.
9361 * config/i386/i386.md (fpint_logic): New.
9362 * config/i386/sse.md (VMOVE): Extend for AVX512.
9366 (VI_UNALIGNED_LOADSTORE): Ditto.
9367 (sse2_avx_avx512f): Ditto.
9368 (sse2_avx2): Extend for AVX512.
9369 (sse4_1_avx2): Ditto.
9370 (avx2_avx512f): New.
9371 (sse): Extend for AVX512.
9374 (avxsizesuffix): Ditto.
9375 (sseintvecmode): Ditto.
9377 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
9378 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
9379 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
9380 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
9381 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
9382 (<sse2_avx_avx512f>_storedqu<mode): ... this.
9383 (<sse>_movnt<mode>): Replace constraint "x" with "v".
9384 (STORENT_MODE): Extend for AVX512.
9385 (*absneg<mode>2): Replace constraint "x" with "v".
9386 (*mul<mode>3): Ditto.
9387 (*ieee_smin<mode>3): Ditto.
9388 (*ieee_smax<mode>3): Ditto.
9389 (avx_cmp<mode>3): Replace VF with VF_128_256.
9390 (*<sse>_maskcmp<mode>3_comm): Ditto.
9391 (<sse>_maskcmp<mode>3): Ditto.
9392 (<sse>_andnot<mode>3): Extend for AVX512.
9393 (<code><mode>3, anylogic): Replace VF with VF_128_256.
9394 (<code><mode>3, fpint_logic): New.
9395 (*<code><mode>3): Extend for AVX512.
9396 (avx512flogicsuff): New.
9397 (avx512f_<logic><mode>): Ditto.
9398 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
9400 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
9401 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
9402 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
9403 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
9404 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
9405 (xop_vpermil2<mode>3): Ditto.
9406 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
9407 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
9408 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
9409 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
9411 2013-09-25 Tom Tromey <tromey@redhat.com>
9413 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
9414 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
9415 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
9416 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
9417 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
9418 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
9419 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
9420 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
9421 (GRAPHITE_HTAB_H): Remove.
9423 2013-09-25 Tom Tromey <tromey@redhat.com>
9425 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
9427 2013-09-25 Tom Tromey <tromey@redhat.com>
9429 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
9431 2013-09-25 Tom Tromey <tromey@redhat.com>
9433 * config/i386/t-i386 (i386.o): Remove.
9434 (i386-c.o): Use COMPILE and POSTCOMPILE.
9436 2013-09-25 Tom Tromey <tromey@redhat.com>
9438 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
9440 2013-09-25 Tom Tromey <tromey@redhat.com>
9442 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
9443 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
9444 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
9445 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
9446 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
9447 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
9448 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
9449 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
9450 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
9451 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
9452 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
9453 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
9454 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
9455 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
9456 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
9457 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
9458 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
9459 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
9460 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
9461 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
9462 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
9463 (tree-ssa-pre.o, tree-ssa-sccvn.o)
9464 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
9465 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
9466 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
9467 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
9468 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
9469 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
9470 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
9471 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
9472 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
9473 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
9474 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
9475 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
9476 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
9477 (graphite-clast-to-gimple.o, graphite-dependences.o)
9478 (graphite-interchange.o, graphite-poly.o)
9479 (graphite-scop-detection.o, graphite-sese-to-poly.o)
9480 (graphite-optimize-isl.o, tree-vect-loop.o)
9481 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
9482 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
9483 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
9484 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
9485 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
9486 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
9487 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
9488 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
9489 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
9490 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
9491 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
9492 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
9493 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
9494 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
9495 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
9496 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
9497 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
9498 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
9499 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
9500 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
9501 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
9502 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
9503 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
9504 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
9505 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
9506 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
9507 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
9508 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
9509 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
9510 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
9511 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
9512 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
9513 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
9514 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
9515 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
9516 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
9517 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
9518 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
9519 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
9520 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
9521 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
9522 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
9523 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
9524 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
9525 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
9526 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
9527 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
9528 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
9529 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
9530 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
9531 (gcov-dump.o): Remove.
9532 (default-c.o): Use COMPILE and POSTCOMPILE.
9533 (CFLAGS-gcc.o): New variable.
9534 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
9536 2013-09-25 Tom Tromey <tromey@redhat.com>
9538 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
9539 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
9540 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
9541 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
9542 (c-family/c-pragma.o, c-family/c-pretty-print.o)
9543 (c-family/c-semantics.o, c-family/c-ada-spec.o)
9544 (c-family/array-notation-common.o, c-family/stub-objc.o)
9545 (c-family/c-ubsan.o): Remove.
9547 2013-09-25 Tom Tromey <tromey@redhat.com>
9549 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
9551 2013-09-25 Tom Tromey <tromey@redhat.com>
9553 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
9554 to add -DENABLE_SHARED_LIBGCC.
9555 (gcc.o): Don't use subshell.
9557 2013-09-25 Tom Tromey <tromey@redhat.com>
9559 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
9560 * configure.ac: Don't invoke AM_PROG_CC_C_O.
9561 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
9562 * configure, config.in: Rebuild.
9564 2013-09-25 Tom Tromey <tromey@redhat.com>
9566 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
9567 (COMPILE, POSTCOMPILE): New variables.
9568 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
9569 (DEPFILES): New variable.
9570 Include ".Po" files.
9571 * configure.ac: Add checks for dependency checking.
9572 * configure, aclocal.m4: Regenerate.
9574 2013-09-25 Tom Tromey <tromey@redhat.com>
9576 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
9577 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
9579 2013-09-25 Tom Tromey <tromey@redhat.com>
9581 * Makefile.in (generated_files): Add options.h,
9582 target-hooks-def.h, insn-opinit.h,
9583 common/common-target-hooks-def.h, pass-instances.def,
9584 c-family/c-target-hooks-def.h.
9586 2013-09-25 Jeff Law <law@redhat.com>
9588 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
9589 than foo[foo.length () - 1] to access last member in a vec.
9590 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
9592 2013-09-25 Richard Biener <rguenther@suse.de>
9595 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
9597 2013-09-25 Jan Hubicka <jh@suse.cz>
9599 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
9602 2013-09-25 Marek Polacek <polacek@redhat.com>
9605 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
9606 when determining the type name.
9608 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
9610 * config/sh/sh.md: Fix formatting.
9612 2013-09-24 Xinliang David Li <davidxl@google.com>
9614 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
9615 max peel iterations parameter.
9616 * param.def: New parameter.
9617 * doc/invoke.texi: Document New parameter.
9619 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
9621 * gimple-pretty-print.c: Various whitespace tweaks.
9622 * tree-core.h: Likewise.
9623 * tree-pretty-print.c: Likewise.
9624 * tree-ssa-alias.c: Likewise.
9625 * tree-ssa-copy.c: Likewise.
9626 * tree-ssanames.c: Likewise.
9627 * tree-ssanames.h: Likewise.
9628 * tree-vrp.c: Likewise.
9630 2013-09-24 Alan Modra <amodra@gmail.com>
9634 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
9635 for output operands that disallow regs. Don't use EXPAND_WRITE on
9638 2013-09-24 Richard Biener <rguenther@suse.de>
9641 * tree.c (reference_alias_ptr_type): Move ...
9642 * alias.c (reference_alias_ptr_type): ... here and implement
9643 in terms of the new reference_alias_ptr_type_1.
9644 (ref_all_alias_ptr_type_p): New helper.
9645 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
9646 use ref_all_alias_ptr_type_p.
9647 (get_deref_alias_set): Add flag_strict_aliasing check here.
9648 (reference_alias_ptr_type_1): New function, split out from ...
9649 (get_alias_set): ... here.
9650 (alias_ptr_types_compatible_p): New function.
9651 * alias.h (reference_alias_ptr_type): Declare.
9652 (alias_ptr_types_compatible_p): Likewise.
9653 * tree.h (reference_alias_ptr_type): Remove.
9654 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
9655 to compare MEM_REF alias types.
9657 2013-09-24 Richard Biener <rguenther@suse.de>
9659 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
9661 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9663 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
9664 reload helper function arrays into a single array reg_addr.
9665 (reload_fpr_gpr): Likewise.
9666 (reload_gpr_vsx): Likewise.
9667 (reload_vsx_gpr): Likewise.
9668 (struct rs6000_reg_addr): Likewise.
9669 (reg_addr): Likewise.
9670 (rs6000_debug_reg_global): Change rs6000_vector_reload,
9671 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
9672 (rs6000_init_hard_regno_mode_ok): Likewise.
9673 (rs6000_secondary_reload_direct_move): Likewise.
9674 (rs6000_secondary_reload): Likewise.
9676 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
9677 constraints: wu, ww, and wy. Repurpose wv constraint added during
9678 power8 changes. Put wg constraint in alphabetical order.
9680 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
9681 for future work to add ISA 2.07 VSX single precision support.
9682 (-mvsx-scalar-double): Change default from -1 to 1, update
9683 documentation comment.
9684 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
9685 (-mupper-regs-df): New debug switch to control whether DF values
9686 can go in the traditional Altivec registers.
9687 (-mupper-regs-sf): New debug switch to control whether SF values
9688 can go in the traditional Altivec registers.
9690 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
9692 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
9693 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
9694 Add new constraints, wu/ww/wy. Repurpose wv constraint.
9695 (rs6000_debug_legitimate_address_p): Print if we are running
9696 before, during, or after reload.
9697 (rs6000_secondary_reload): Add a comment.
9698 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
9700 * config/rs6000/constraints.md (wa constraint): Sort w<x>
9701 constraints. Update documentation string.
9702 (wd constraint): Likewise.
9703 (wf constraint): Likewise.
9704 (wg constraint): Likewise.
9705 (wn constraint): Likewise.
9706 (ws constraint): Likewise.
9707 (wt constraint): Likewise.
9708 (wx constraint): Likewise.
9709 (wz constraint): Likewise.
9710 (wu constraint): New constraint for ISA 2.07 SFmode scalar
9712 (ww constraint): Likewise.
9713 (wy constraint): Likewise.
9714 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
9715 in the previous submissions.
9716 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
9718 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
9720 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
9721 (REG_BR_PROB): Say that the probability is stored in an int_list.
9722 * reg-notes.def: Update commentary to mention INT_LIST.
9723 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
9724 (INT_LIST): New rtx.
9725 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
9726 * rtlanal.c (int_reg_note_p): New function.
9727 (alloc_reg_note): Assert that the note does not have an int argument.
9728 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
9729 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
9730 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
9731 rather than an INSN_LIST. Handle INT_LIST.
9732 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
9733 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
9734 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
9735 Manipulate them as ints rather than rtxes.
9736 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
9737 * regmove.c (copy_src_to_dest): Likewise.
9738 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
9740 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
9741 into the cases that need it.
9742 * config/arm/arm.c (arm_unwind_emit): Likewise.
9744 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
9745 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
9746 * loop-doloop.c (add_test, doloop_modify): Likewise.
9747 * loop-unswitch.c (compare_and_jump_seq): Likewise.
9748 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
9749 * predict.c (combine_predictions_for_insn): Likewise.
9750 * print-rtl.c (print_rtx): Handle INT_LIST.
9751 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
9752 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
9753 * config/arm/arm.c (emit_unlikely_jump): Likewise.
9754 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
9755 (ix86_split_fp_branch, predict_jump): Likewise.
9756 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
9757 * config/sh/sh.c (expand_cbranchsi4): Likewise.
9758 * config/spu/spu.c (ea_load_store_inline): Likewise.
9760 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
9761 value of a REG_BR_PROB note.
9762 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
9763 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
9764 * emit-rtl.c (try_split): Likewise.
9765 * predict.c (br_prob_note_reliable_p): Likewise.
9766 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
9767 * reorg.c (mostly_true_jump): Likewise.
9768 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
9769 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
9770 * config/i386/i386.c (ix86_print_operand): Likewise.
9771 * config/ia64/ia64.c (ia64_print_operand): Likewise.
9772 * config/mmix/mmix.c (mmix_print_operand): Likewise.
9773 * config/rs6000/rs6000.c (output_cbranch): Likewise.
9774 * config/s390/s390.c (s390_expand_tbegin): Likewise.
9775 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
9776 * config/sparc/sparc.c (output_cbranch): Likewise.
9777 * config/spu/spu.c (get_branch_target): Likewise.
9778 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
9779 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
9781 2013-09-23 Jan Hubicka <jh@suse.cz>
9783 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
9785 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
9786 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
9788 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
9789 (gimple_fold_call): Dump inconsistent devirtualizations; add
9790 sanity check for type based devirtualizations.
9791 * ipa-prop.c: Include ipa-utils.h
9792 (ipa_intraprocedural_devirtualization): Add sanity check.
9793 (try_make_edge_direct_virtual_call): Likewise.
9795 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
9797 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
9798 assignment statements.
9800 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
9802 * gimple-pretty-print.c (dump_ssaname_info): New function.
9803 (dump_gimple_phi): Call it.
9804 (pp_gimple_stmt_1): Likewise.
9805 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
9806 (range_info_def): Declare.
9807 * tree-pretty-print.c (pp_double_int): New function.
9808 (dump_generic_node): Call it.
9809 * tree-pretty-print.h (pp_double_int): Declare.
9810 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
9811 * tree-ssanames.h (range_info_def): New structure.
9812 (value_range_type): Move definition here.
9813 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
9815 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
9817 (set_range_info): New function.
9818 (get_range_info): Likewise.
9819 (duplicate_ssa_name_range_info): Likewise.
9820 (duplicate_ssa_name_fn): Check pointer type and call
9821 duplicate_ssa_name_range_info.
9822 * tree-ssa-copy.c (fini_copy_prop): Likewise.
9823 * tree-vrp.c (value_range_type): Remove definition, now in
9825 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
9826 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
9827 (SSA_NAME_RANGE_INFO): New macro.
9829 2013-09-23 Richard Biener <rguenther@suse.de>
9831 PR tree-optimization/58464
9832 * tree-ssa-pre.c (phi_trans_lookup): Remove.
9833 (phi_trans_add): Change to add conditionally on being not present.
9834 (phi_translate_1): Remove recursion detection here.
9835 (phi_translate): Pre-seed the cache with NULL to catch
9836 recursion here in a more generic way.
9837 (bitmap_find_leader): Adjust comment.
9838 (get_representative_for): Dump value-numbers.
9839 (create_expression_by_pieces): Likewise.
9840 (insert_into_preds_of_block): Likewise.
9842 2013-09-23 Christian Bruel <christian.bruel@st.com>
9845 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
9846 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
9848 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
9851 2013-09-20 Renlin Li <renlin.li@arm.com>
9853 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
9855 (aarch64_expand_epilogue): Likewise.
9856 (aarch64_legitimize_reload_address): Likewise.
9858 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
9860 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
9861 NULL_TREE before pushing them onto the vector. Likewise for labels.
9863 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
9865 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
9866 comparison operators when -fno-trapping-math is in effect.
9867 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
9868 comparison operators in TFmode and assert that unsupported operators
9870 (ia64_print_operand): Likewise.
9872 2013-09-21 Jan Hubicka <jh@suse.cz>
9874 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
9875 (sse_typeless_stores): Enable for core
9876 (sse_load0_by_pxor): Likewise.
9877 (four_jump_limit): Disable for core.
9878 (pad_returns): Likewise.
9879 (avoid_vector_decode): Likewise.
9880 (fuse_cmp_and_branch): Enable for cores.
9881 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
9883 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9886 * config/pa/pa.c (pa_option_override): Disable auto increment and
9887 decrement instructions until reload is completed.
9889 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
9890 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
9891 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
9893 2013-09-20 DJ Delorie <dj@redhat.com>
9894 Nick Clifton <nickc@redhat.com>
9896 * config/rl78/rl78.c: Various whitespace and comment tweaks.
9897 (need_to_save): Save bank 0 on interrupts.
9898 (characterize_address): Strip far address wrappers.
9899 (rl78_as_legitimate_address): Likewise.
9900 (transcode_memory_rtx): Likewise.
9901 (rl78_peep_movhi_p): Disable this peephole after devirt.
9902 (rl78_propogate_register_origins): Forget all origins when a
9904 * config/rl78/rl78-virt.md: Various whitespace tweaks.
9905 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
9907 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
9908 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
9909 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
9911 * config/rl78/constraints.md: Various whitespace and paren tweaks.
9913 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9915 * config/pa/pa.md: In "scc" insn patterns, change output template to
9916 handle const0_rtx in reg_or_0_operand operands.
9918 2013-09-20 Martin Husemann <martin@NetBSD.org>
9921 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
9922 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
9924 2013-09-20 Richard Biener <rguenther@suse.de>
9927 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
9928 remembering SSA name version and block index.
9929 (new_scev_info_str): Adjust.
9930 (hash_scev_info): Likewise. Also hash the block index.
9931 (eq_scev_info): Adjust.
9932 (find_var_scev_info): Likewise.
9933 (struct instantiate_cache_entry): Remove.
9934 (struct instantiate_cache_type): Use a htab to map name, block
9936 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9937 (get_instantiated_value_entry): Likewise.
9938 (hash_idx_scev_info, eq_idx_scev_info): New functions.
9939 (instantiate_scev_name): Adjust.
9941 2013-09-20 Jeff Law <law@redhat.com>
9943 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
9945 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
9947 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
9948 Call aarch64_simd_expand_args to update op[argc].
9950 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9952 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
9955 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9957 * gengtype.c (file_rules): Added rule for *.cc files.
9958 (get_output_file_with_visibility): Give fatal message when no
9961 2013-09-20 Renlin Li <renlin.li@arm.com>
9963 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
9964 (aarch64_expand_epilogue): Likewise.
9965 (aarch64_legitimize_reload_address): Likewise.
9967 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
9970 * expr.c (expand_assignment): Remove misalignp code path.
9972 2013-09-20 Marek Polacek <polacek@redhat.com>
9975 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
9976 TYPE_PRECISION. Add asserts.
9978 2013-09-20 Richard Biener <rguenther@suse.de>
9980 PR tree-optimization/58453
9981 * tree-loop-distribution.c (distribute_loop): Apply the cost
9982 model for -ftree-loop-distribute-patterns, too.
9984 2013-09-20 Richard Biener <rguenther@suse.de>
9987 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
9988 make type comparison less strict.
9990 2013-09-20 Alan Modra <amodra@gmail.com>
9992 * configure: Regenerate.
9993 * aclocal.m4: Regenerate.
9995 2013-09-20 Marek Polacek <polacek@redhat.com>
9998 * doc/extend.texi: Document that attribute used is meant to be used
9999 on variables with static storage duration.
10001 2013-09-19 Jakub Jelinek <jakub@redhat.com>
10003 PR tree-optimization/58472
10004 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
10005 simd_lane_access set inv_p = false.
10006 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
10007 the simduid magic VAR_DECL.
10009 2013-09-19 Jan Hubicka <jh@suse.cz>
10011 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
10013 2013-09-17 Jeff Law <law@redhat.com>
10015 * tree-ssa-dom.c (record_temporary_equivalences): New function
10016 split out of dom_opt_dom_walker::after_dom_children.
10017 (dom_opt_dom_walker::thread_across_edge): Move common code
10018 in here from dom_opt_dom_walker::after_dom_children.
10019 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
10021 2013-09-19 Jan Hubicka <jh@suse.cz>
10023 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
10024 (TARGET_GENERIC): Use PROCESOR_GENERIC
10025 (enum processor_type): Unify generic32 and 64.
10026 * i386.md (cpu): Likewise.
10027 * x86-tune.def (use_leave): Enable for generic32.
10028 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
10029 * athlon.md: Change generic64 to generic in all occurences.
10030 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
10031 (ix86_target_macros_internal): Likewise.
10032 * driver-i386.c (host_detect_local_cpu): Likewise.
10033 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
10035 (generic_memcpy, generic_memset, generic_cost): This one.
10036 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
10037 (m_GENERIC32, m_GENERIC64): Remove.
10038 (m_GENERIC): Turn into one flag.
10039 (processor_target): Unify generic tunnings.
10040 (ix86_option_override_internal): Replace generic32/64 by generic.
10041 (ix86_issue_rate): Likewise.
10042 (ix86_adjust_cost): Likewise.
10044 2013-09-19 Jan Hubicka <jh@suse.cz>
10046 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
10047 of speculative flag.
10049 2013-09-19 Jakub Jelinek <jakub@redhat.com>
10051 * omp-low.c (expand_omp_sections): Always pass len - 1 to
10052 GOMP_sections_start, even if !exit_reachable.
10054 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
10056 * lra-constraints.c (need_for_all_save_p): Use macro
10057 HARD_REGNO_CALL_PART_CLOBBERED.
10058 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
10059 set up pseudo conflict hard regs.
10061 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
10064 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
10067 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
10069 PR rtl-optimization/58438
10070 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
10071 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
10072 from previous subpasses.
10074 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
10076 * arm.c (arm_get_frame_offsets): Validate architecture supports
10077 LDRD/STRD before accepting the tuning preference.
10078 (arm_expand_prologue): Likewise.
10079 (arm_expand_epilogue): Likewise.
10081 2013-09-18 Richard Biener <rguenther@suse.de>
10083 PR tree-optimization/58417
10084 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
10085 have chrecs with symbols defined in the loop as operands.
10086 (chrec_fold_multiply): Likewise.
10087 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
10088 parameters before folding binary operations.
10089 (struct instantiate_cache_entry_hasher): Remove.
10090 (struct instantiate_cache_type): Use a pointer-map.
10091 (instantiate_cache_type::instantiate_cache_type): New function.
10092 (instantiate_cache_type::get): Likewise.
10093 (instantiate_cache_type::set): Likewise.
10094 (instantiate_cache_type::~instantiate_cache_type): Adjust.
10095 (get_instantiated_value_entry): Likewise.
10096 (global_cache): New global.
10097 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
10098 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
10099 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
10100 (instantiate_scev_name): Adjust.
10101 (instantiate_scev): Construct global instead of local cache.
10102 (resolve_mixers): Likewise.
10104 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
10105 Paolo Carlini <paolo.carlini@oracle.com>
10108 * doc/implement-cxx.texi: Fix references to the C++ standards.
10110 2013-09-18 Jakub Jelinek <jakub@redhat.com>
10112 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
10113 * tree-vect-data-refs.c (vect_analyze_data_refs): For
10114 simd_lane_access drs, update also DR_ALIGNED_TO.
10116 2013-09-18 Marek Polacek <polacek@redhat.com>
10119 * doc/extend.texi: Document no_sanitize_undefined attribute.
10120 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
10121 no_sanitize_undefined attribute.
10123 2013-09-18 Nick Clifton <nickc@redhat.com>
10125 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
10126 (ASM_DECLARE_FUNCTION_NAME): Define.
10128 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
10130 * compare-elim.c (find_comparison_dom_walker): New class
10131 (find_comparisons_in_bb): Rename to
10132 find_comparison_dom_walker::before_dom_children
10133 (find_comparisons): Adjust
10134 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
10136 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
10137 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
10138 (init_walk_dominator_tree): Remove declaration.
10139 (fini_walk_dominator_tree): Remove declaration.
10140 * fwprop.c (single_def_use_dom_walker): New class
10141 (single_def_use_enter_block): Convert to
10142 single_def_use_dom_walker::before_dom_children.
10143 (single_def_use_leave_block): Convert to
10144 single_def_use_dom_walker::after_dom_children.
10145 (build_single_def_use_links): Adjust.
10146 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
10148 (find_candidates_in_block): Convert to
10149 find_candidates_dom_walker::before_dom_children.
10150 (execute_strength_reduction): Adjust.
10151 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
10152 (sese_dom_walker): New class.
10153 (sese_dom_walker::sese_dom_walker): New constructor.
10154 (sese_dom_walker::~sese_dom_walker): New destructor.
10155 (build_sese_conditions_before): Convert to
10156 sese_dom_walker::before_dom_children.
10157 (build_sese_conditions_after): Convert to
10158 sese_dom_walker::after_dom_children.
10159 (build_poly_scop): Adjust
10160 * tree-into-ssa.c (rewrite_dom_walker): New class
10161 (rewrite_enter_block): Convert to
10162 rewrite_dom_walker::before_dom_children.
10163 (rewrite_leave_block): Convert to
10164 rewrite_dom_walker::after_dom_children.
10165 (rewrite_update_dom_walker): New class.
10166 (rewrite_update_enter_block): Convert to
10167 rewrite_update_dom_walker::before_dom_children.
10168 (rewrite_update_leave_block): Convert to
10169 rewrite_update_dom_walker::after_dom_children.
10170 (rewrite_blocks, rewrite_into_ssa): Adjust.
10171 (mark_def_dom_walker): New class.
10172 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
10173 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
10174 (mark_def_sites_blocks): Convert to
10175 mark_def_dom_walker::before_dom_children.
10176 (mark_def_site_blocks): Remove.
10177 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
10178 (tree_ssa_dominator_optimize): Adjust.
10179 (dom_thread_across_edge): Convert to method
10180 dom_opt_dom_walker::thread_across_edge.
10181 (dom_opt_enter_block): Convert to member function
10182 dom_opt_dom_walker::before_dom_children.
10183 (dom_opt_leave_block): Convert to member function
10184 dom_opt_dom_walker::after_dom_children.
10185 * tree-ssa-dse.c (dse_dom_walker): New class.
10186 (dse_enter_block): Convert to member function
10187 dse_dom_walker::before_dom_children.
10188 (tree_ssa_dse): Adjust.
10189 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
10190 (determine_invariantness_stmt): Convert to method
10191 invariantness_dom_walker::before_dom_children.
10192 (determine_invariantness): Remove
10193 (move_computations_dom_walker): New class.
10194 (move_computations_stmt): Convert to method
10195 move_computations_dom_walker::before_dom_children.
10196 (move_computations, tree_ssa_lim): Adjust.
10197 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
10198 (nt_init_block): Convert to method
10199 notrappping_dom_walker::before_dom_children.
10200 (nt_fini_block): Convert to method
10201 method nontrapping_dom_walker::after_dom_children.
10202 (get_non_trapping): Adjust.
10203 * tree-ssa-pre.c (eliminate_dom_walker): New class.
10204 (eliminate_bb): Convert to method
10205 eliminate_dom_walker::before_dom_children.
10206 (eliminate_leave_block): Convert to method
10207 eliminate_dom_walker::after_dom_children.
10208 (eliminate): Adjust.
10209 * tree-ssa-strlen.c (strlen_dom_walker): New class.
10210 (strlen_enter_block): Convert to method
10211 strlen_dom_walker::before_dom_children.
10212 (strlen_leave_block): Convert to method
10213 method strlen_dom_walker::after_dom_children.
10214 (tree_ssa_strlen): Adjust.
10215 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
10216 (tree_ssa_uncprop): Adjust.
10217 (uncprop_leave_block): Convert to method
10218 uncprop_dom_walker::after_dom_children.
10219 (uncprop_leave_block): Convert to method
10220 uncprop_dom_walker::before_dom_children.
10222 2013-09-18 Bin Cheng <bin.cheng@arm.com>
10224 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
10225 branch in same basic block. Check both src and dest of the move insn.
10227 2013-09-17 Nick Clifton <nickc@redhat.com>
10229 * config/rl78/rl78-real.md (bf): New pattern.
10231 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
10232 (rl78_print_operand): Do not put a # before a %B.
10233 * config/rl78/rl78.opt: Tweak doc strings.
10235 2013-09-17 DJ Delorie <dj@redhat.com>
10237 * config/rl78/constraints.md (Wcv): Allow up to $r31.
10238 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
10239 (rl78_option_override): Likewise, if -mallregs.
10240 (is_virtual_register): Likewise.
10241 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
10242 (REGNO_OK_FOR_BASE_P): Likewise.
10243 * config/rl78/rl78.opt (-mallregs): New.
10245 2013-09-17 Nick Clifton <nickc@redhat.com>
10247 * config/rl78/rl78.c (need_to_save): Change return type to bool.
10248 For interrupt functions: save all call clobbered registers if the
10249 interrupt handler is not a leaf function.
10250 (rl78_expand_prologue): Always recompute the frame information.
10251 For interrupt functions: only select bank 0 if one of the bank 0
10252 registers is going to be psuhed.
10254 2013-09-17 DJ Delorie <dj@redhat.com>
10256 * config/rl78/constraints.md: For each W* constraint, rename to C*
10257 and create a W* constraint that checks for an optional ES: prefix
10259 * config/rl78/rl78.md (UNS_ES_ADDR): New.
10260 (es_addr): New. Used to wrap far addresses.
10261 * config/rl78/rl78-protos.h (rl78_es_addr): New.
10262 (rl78_es_base): New.
10263 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
10264 wrapped far addresses.
10265 (rl78_print_operand_1): Unwrap far addresses before processing.
10266 (rl78_lo16): Wrap far addresses in unspecs.
10267 (rl78_es_addr): New.
10268 (rl78_es_base): New.
10269 (insn_ok_now): Check for not-yet-wrapped far addresses.
10270 (transcode_memory_rtx): Properly re-wrap far addresses.
10272 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
10274 * config/sparc/t-rtems: Add leon3 multilibs.
10276 2013-09-17 Cong Hou <congh@google.com>
10278 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
10279 when checking the dot production pattern. The type of rhs operand
10280 of multiply is now checked correctly.
10282 2013-09-17 Jeff Law <law@redhat.com>
10284 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10285 edge implied equivalences into successor phis.
10286 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
10287 here from tree-ssa-threadedge.c.
10288 (mark_threaded_blocks): When threading through a joiner, if both
10289 successors of the joiner's clone reach the same block, verify the
10290 PHI arguments are equal. If not, cancel the jump threading request.
10291 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
10292 tree-ssa-threadupdate.c
10293 (thread_across_edge): Don't check PHI argument equality when
10294 threading through joiner block here.
10296 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
10298 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
10299 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
10300 (warn_uninit, warn_uninitialized_vars,
10301 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
10302 Move to tree-ssa-uninit.c.
10303 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
10304 (has_undefined_value_p): New. Pass dependant parts of
10305 ssa_undefined_value_p.
10306 (uninit_undefined_value_p): Use has_undefined_value_p.
10307 (warn_uninit, warn_uninitialized_vars,
10308 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
10309 Move from tree-ssa.c.
10310 * tree-ssa.h: Adjust prototypes.
10312 2013-09-17 Jan Hubicka <jh@suse.cz>
10314 PR middle-end/58332
10315 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
10316 * ipa-inline.c (can_inline_edge_p): Do not downgrade
10317 FUNCTION_NOT_OPTIMIZED.
10318 * ipa-inline-analysis.c (compute_inline_parameters): Function
10319 not optimized is not inlinable unless it is alwaysinline.
10320 (inline_analyze_function): Force calls in not optimized
10321 function not inlinable.
10323 2013-09-17 Jan Hubicka <jh@suse.cz>
10325 PR middle-end/58329
10326 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
10328 * ipa.c (function_and_variable_visibility): Likewise.
10329 * ipa-profile.c (ipa_profile): Likewise.
10331 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
10334 * cgraph.c (cgraph_function_body_availability): Check for ifunc
10335 attribute, and don't inline the resolver in this case.
10337 2013-09-17 Teresa Johnson <tejohnson@google.com>
10339 * coverage.c (get_coverage_counts): Add missing newline.
10341 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10343 PR tree-optimization/58088
10344 * fold-const.c (mask_with_trailing_zeros): New function.
10345 (fold_binary_loc): Make sure we don't recurse infinitely
10346 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
10347 Use mask_with_trailing_zeros where appropriate.
10349 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
10351 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
10352 of 'prev' var to get better distance estimation.
10354 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
10356 * tree-inline.h (struct copy_body_data): Add transform_parameter.
10357 * tree-inline.c (is_parameter_of): New predicate.
10358 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
10359 a parameter has been remapped.
10360 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
10361 (optimize_inline_calls): Initialize transform_parameter.
10362 (copy_gimple_seq_and_replace_locals): Likewise.
10363 (tree_function_versioning): Likewise.
10364 (maybe_inline_call_in_expr): Likewise.
10366 2013-09-17 Nick Clifton <nickc@redhat.com>
10368 * config/msp430/msp430-protos.h: Add prototypes for new functions.
10369 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
10370 interrupt handlers.
10371 (is_attr_func): New function.
10372 (msp430_is_interrupt_func): New function.
10373 (is_naked_func): New function.
10374 (is_reentrant_func): New function.
10375 (is_critical_func): New function.
10376 (msp430_start_function): Add annotations for function attributes.
10377 (msp430_attr): New function.
10378 (msp430_attribute_table): New.
10379 (msp430_function_section): New function.
10380 (TARGET_ASM_FUNCTION_SECTION): Define.
10381 (msp430_builtin): New enum.
10382 (msp430_init_builtins): New function.
10383 (msp430_builtin_devl): New function.
10384 (msp430_expand_builtin): New function.
10385 (TARGET_INIT_BUILTINS): Define.
10386 (TARGET_EXPAND_BUILTINS): Define.
10387 (TARGET_BUILTIN_DECL): Define.
10388 (msp430_expand_prologue): Add support for naked, interrupt,
10389 critical and reentrant functions.
10390 (msp430_expand_epilogue): Likewise.
10391 (msp430_print_operand): Handle 'O' character.
10392 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
10394 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
10395 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
10396 (pushm): Use a 'n' rather than an 'i' constraint.
10397 (msp_return): Add generation of the interrupt return instruction.
10398 (disable_interrupts): New pattern.
10399 (enable_interrupts): New pattern.
10400 (push_intr_state): New pattern.
10401 (pop_intr_state): New pattern.
10402 (bic_SR): New pattern.
10403 (bis_SR): New pattern.
10404 * doc/extend.texi: Document MSP430 function attributes and builtin
10407 2013-09-17 Richard Biener <rguenther@suse.de>
10409 PR tree-optimization/58432
10410 * tree-loop-distribution.c (tree_loop_distribution): Also
10411 scan PHIs for outside loop uses and seed a partition from them.
10413 2013-09-17 Bin Cheng <bin.cheng@arm.com>
10415 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
10416 (restructure_reference): Call backtrace_base_for_ref.
10418 2013-09-17 Alan Modra <amodra@gmail.com>
10421 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
10424 2013-09-16 DJ Delorie <dj@redhat.com>
10426 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
10427 vregs location for RL78/G10.
10428 (rl78_expand_prologue): Avoid SEL on G10.
10429 (rl78_expand_epilogue): Likewise.
10430 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
10431 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
10432 __RL78_G10__ when appropriate.
10433 (ASM_SPEC): Pass -mg10 along to the assembler.
10434 * config/rl78/rl78.md (sel_rb): Disable for G10.
10435 * config/rl78/rl78.opt: Add -mg10 option.
10436 * config/rl78/t-rl78: Add -mg10 multilib.
10438 2013-09-16 Xinliang David Li <davidxl@google.com>
10440 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
10441 * omp-low.c (omp_max_vf): Ditto.
10442 (expand_omp_simd): Ditto.
10443 * tree-vectorizer.c (vectorize_loops): Ditto.
10444 (gate_vect_slp): Ditto.
10445 (gate_increase_alignment): Ditto.
10446 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
10447 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
10448 (gate_tree_vectorize): Name change.
10449 (tree_vectorize): Ditto.
10450 (pass_vectorize::gate): Call new function.
10451 (pass_vectorize::execute): Ditto.
10452 * opts.c: O3 default setting change.
10453 (finish_options): Check new flag.
10454 * doc/invoke.texi: Document new flags.
10455 * common.opt: New flags.
10457 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
10459 * doc/tm.texi.in (Cond Exec Macros): Remove node.
10460 (Condition Code): Don't reference it.
10461 * doc/tm.texi: Regenerate.
10463 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
10465 PR middle-end/58418
10466 * lra-constraints.c (undo_optional_reloads): Consider all optional
10467 reload even if it did not get a hard reg.
10469 2013-09-16 Teresa Johnson <tejohnson@google.com>
10471 * dumpfile.c (dump_loc): Remove newline emission.
10472 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
10473 emission to dump_printf_loc calls where missing.
10474 (vect_mark_for_runtime_alias_test): Ditto.
10475 (vect_analyze_data_ref_dependence): Ditto.
10476 (vect_analyze_data_ref_dependences): Ditto.
10477 (vect_slp_analyze_data_ref_dependence): Ditto.
10478 (vect_slp_analyze_data_ref_dependences): Ditto.
10479 (vect_compute_data_ref_alignment): Ditto.
10480 (vect_update_misalignment_for_peel): Ditto.
10481 (vect_verify_datarefs_alignment): Ditto.
10482 (vector_alignment_reachable_p): Ditto.
10483 (vect_get_data_access_cost): Ditto.
10484 (vect_enhance_data_refs_alignment): Ditto.
10485 (vect_find_same_alignment_drs): Ditto.
10486 (vect_analyze_data_refs_alignment): Ditto.
10487 (vect_analyze_group_access): Ditto.
10488 (vect_analyze_data_ref_access): Ditto.
10489 (vect_analyze_data_ref_accesses): Ditto.
10490 (vect_prune_runtime_alias_test_list): Ditto.
10491 (vect_analyze_data_refs): Ditto.
10492 (vect_create_addr_base_for_vector_ref): Ditto.
10493 (vect_create_data_ref_ptr): Ditto.
10494 (vect_grouped_store_supported): Ditto.
10495 (vect_grouped_load_supported): Ditto.
10496 * value-prof.c (check_counter): Ditto.
10497 (check_ic_target): Ditto.
10498 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
10499 (vect_recog_widen_mult_pattern): Ditto.
10500 (vect_recog_widen_sum_pattern): Ditto.
10501 (vect_recog_over_widening_pattern): Ditto.
10502 (vect_recog_widen_shift_pattern): Ditto.
10503 (vect_recog_rotate_pattern): Ditto.
10504 (vect_recog_vector_vector_shift_pattern): Ditto.
10505 (vect_recog_divmod_pattern): Ditto.
10506 (vect_recog_mixed_size_cond_pattern): Ditto.
10507 (vect_recog_bool_pattern): Ditto.
10508 (vect_pattern_recog_1): Ditto.
10509 (vect_pattern_recog): Ditto.
10510 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
10511 (vect_is_simple_iv_evolution): Ditto.
10512 (vect_analyze_scalar_cycles_1): Ditto.
10513 (vect_get_loop_niters): Ditto.
10514 (vect_analyze_loop_1): Ditto.
10515 (vect_analyze_loop_form): Ditto.
10516 (vect_analyze_loop_operations): Ditto.
10517 (vect_analyze_loop_2): Ditto.
10518 (vect_analyze_loop): Ditto.
10519 (report_vect_op): Ditto.
10520 (vect_is_slp_reduction): Ditto.
10521 (vect_is_simple_reduction_1): Ditto.
10522 (vect_get_known_peeling_cost): Ditto.
10523 (vect_estimate_min_profitable_iters): Ditto.
10524 (vect_model_reduction_cost): Ditto.
10525 (vect_model_induction_cost): Ditto.
10526 (get_initial_def_for_induction): Ditto.
10527 (vect_create_epilog_for_reduction): Ditto.
10528 (vectorizable_reduction): Ditto.
10529 (vectorizable_induction): Ditto.
10530 (vectorizable_live_operation): Ditto.
10531 (vect_loop_kill_debug_uses): Ditto.
10532 (vect_transform_loop): Ditto.
10533 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
10534 (vect_stmt_relevant_p): Ditto.
10535 (process_use): Ditto.
10536 (vect_mark_stmts_to_be_vectorized): Ditto.
10537 (vect_model_simple_cost): Ditto.
10538 (vect_model_promotion_demotion_cost): Ditto.
10539 (vect_model_store_cost): Ditto.
10540 (vect_get_store_cost): Ditto.
10541 (vect_model_load_cost): Ditto.
10542 (vect_get_load_cost): Ditto.
10543 (vect_init_vector_1): Ditto.
10544 (vect_get_vec_def_for_operand): Ditto.
10545 (vect_finish_stmt_generation): Ditto.
10546 (vectorizable_call): Ditto.
10547 (vectorizable_conversion): Ditto.
10548 (vectorizable_assignment): Ditto.
10549 (vectorizable_shift): Ditto.
10550 (vectorizable_operation): Ditto.
10551 (vectorizable_store): Ditto.
10552 (vectorizable_load): Ditto.
10553 (vectorizable_condition): Ditto.
10554 (vect_analyze_stmt): Ditto.
10555 (vect_transform_stmt): Ditto.
10556 (vect_is_simple_use): Ditto.
10557 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
10558 (vect_can_advance_ivs_p): Ditto.
10559 (vect_update_ivs_after_vectorizer): Ditto.
10560 (vect_do_peeling_for_loop_bound): Ditto.
10561 (vect_gen_niters_for_prolog_loop): Ditto.
10562 (vect_update_inits_of_drs): Ditto.
10563 (vect_create_cond_for_alias_checks): Ditto.
10564 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
10565 (vect_build_slp_tree_1): Ditto.
10566 (vect_supported_load_permutation_p): Ditto.
10567 (vect_analyze_slp_instance): Ditto.
10568 (vect_analyze_slp): Ditto.
10569 (vect_make_slp_decision): Ditto.
10570 (vect_detect_hybrid_slp): Ditto.
10571 (vect_bb_vectorization_profitable_p): Ditto.
10572 (vect_slp_analyze_bb_1): Ditto.
10573 (vect_update_slp_costs_according_to_vf): Ditto.
10574 (vect_get_mask_element): Ditto.
10575 (vect_transform_slp_perm_load): Ditto.
10576 (vect_schedule_slp_instance): Ditto.
10577 (vect_schedule_slp): Ditto.
10578 (vect_slp_transform_bb): Ditto.
10579 * profile.c (read_profile_edge_counts): Ditto.
10580 (compute_branch_probabilities): Ditto.
10581 * coverage.c (get_coverage_counts): Ditto.
10583 2013-09-16 Diego Novillo <dnovillo@google.com>
10585 * tree-core.h: Add missing comment lines from refactoring of tree.h.
10587 2013-09-16 Jan Hubicka <jh@suse.cz>
10589 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
10590 abstract functions; for static functions check the presence of body.
10592 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10594 * config/aarch64/aarch64-simd-builtins.def (fma): New.
10595 * config/aarch64/aarch64-simd.md
10596 (aarch64_mla_elt<mode>): New.
10597 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10598 (aarch64_mls_elt<mode>): Likewise.
10599 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10600 (aarch64_fma4_elt<mode>): Likewise.
10601 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10602 (aarch64_fma4_elt_to_128v2df): Likewise.
10603 (aarch64_fma4_elt_to_64df): Likewise.
10604 (fnma<mode>4): Likewise.
10605 (aarch64_fnma4_elt<mode>): Likewise.
10606 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10607 (aarch64_fnma4_elt_to_128v2df): Likewise.
10608 (aarch64_fnma4_elt_to_64df): Likewise.
10609 * config/aarch64/iterators.md (VDQSF): New.
10610 * config/aarch64/arm_neon.h
10611 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
10612 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
10614 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10616 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
10617 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
10618 (aarch64_mul3_elt_to_128df): Likewise.
10619 (aarch64_mul3_elt_to_64v2df): Likewise.
10620 * config/aarch64/iterators.md (VEL): Also handle DFmode.
10622 (VMUL_CHANGE_NLANES) Likewise.
10625 * config/aarch64/arm_neon.h
10626 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
10628 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10630 * config/aarch64/arm_neon.h
10631 (vcvtx_high_f32_f64): Fix parameters.
10633 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10634 Uros Bizjak <ubizjak@gmail.com>
10636 * config/alpha.c: Include tree-ssanames.h.
10638 2013-09-16 Richard Biener <rguenther@suse.de>
10640 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
10641 (dot_rdg_1): Handle control_dd.
10642 (create_edge_for_control_dependence): New function.
10643 (create_rdg_edges): Add control dependences if asked for.
10644 (build_rdg): Likewise.
10645 (generate_loops_for_partition): If there are not necessary
10646 control stmts remove all their dependencies.
10647 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
10648 (distribute_loop): Pass on control dependences.
10649 (tree_loop_distribution): Compute control dependences and remove
10650 restriction on number of loop nodes.
10652 2013-09-16 Jakub Jelinek <jakub@redhat.com>
10654 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
10655 for internal calls.
10657 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
10659 * cse.c (try_const_anchors): Punt on CC modes.
10661 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10663 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
10665 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
10668 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
10669 calls to word_mode.
10671 2013-09-14 Iain Sandoe <iain@codesourcery.com>
10674 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
10676 (darwin_objc1_section): Likewise.
10677 (darwin_file_end): Emit Image Info section when required.
10679 2013-09-14 Jan Hubicka <jh@suse.cz>
10681 * tree-into-ssa.c (gate_into_ssa): New.
10682 (pass_data_build_ssa): Use it.
10683 * cgraph.h (expand_thunk): Update prototype.
10684 * cgraphunit.c (analyze_function): Expand thunks early.
10685 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
10686 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
10687 set lowered flag; do not add new function.
10688 (assemble_thunks_and_aliases): Update.
10689 * tree-ssa.c (gate_init_datastructures): New gate.
10690 (pass_data_init_datastructures): Use it.
10692 2013-09-14 Iain Sandoe <iain@codesourcery.com>
10695 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
10696 xmm register set described in the psABI.
10698 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
10700 * dwarf2out.c (should_emit_struct_debug): Add check
10701 for type_decl variable is not NULL.
10703 2013-09-13 Jacek Caban <jacek@codeweavers.com>
10705 * config.gcc: Use new winnt-c.c target hooks
10706 * config/t-winnt: New file
10707 * config/winnt-c.c: New file
10708 * doc/tm.texi.in: Document new hook
10709 * doc/tm.texi: Regenerated
10711 2013-09-13 Jan Hubicka <jh@suse.cz>
10713 PR middle-end/58094
10714 * ipa-inline.c (check_callers): New function.
10715 (check_caller_edge): Remove.
10716 (want_inline_function_to_all_callers_p): Also permit alises that are
10718 (inline_to_all_callers): Terminate the walk when devirtualization
10719 introduce new calls.
10721 2013-09-13 Jan Hubicka <jh@suse.cz>
10723 * ipa-inline-analysis.c (struct growth_data): Add node.
10724 (do_estimate_growth_1): Fix detection of recursion.
10726 2013-09-13 Jakub Jelinek <jakub@redhat.com>
10728 PR tree-optimization/58392
10729 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
10730 to avoid shadowing of outer loop variable. If
10731 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
10732 replace_by_duplicate_decl simduid of loops that have it set and
10733 set dest_cfun->has_simduid_loops and/or
10734 dest_cfun->has_force_vect_loops.
10735 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
10736 instead of maybe_lookup_decl.
10737 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
10738 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
10739 call. Copy over force_vect and copy and remap simduid. Set
10740 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
10741 (copy_cfg_body): Remove blocks_to_copy argument. Use
10742 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
10743 caller. Don't set cfun->has_simduid_loops and/or
10744 cfun->has_force_vect_loops here.
10745 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
10747 (expand_call_inline, tree_function_versioning): Adjust copy_body
10750 2013-09-13 Martin Jambor <mjambor@suse.cz>
10753 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
10754 the assert if the edge was a speculative one.
10756 2013-09-13 Richard Biener <rguenther@suse.de>
10758 * tree-data-ref.h (known_dependences_p): Move here ...
10759 * tree-loop-distribution.c (known_dependences_p): ... from here.
10760 (dump_rdg_component, debug_rdg_component): Remove.
10761 (dump_rdg): Adjust.
10762 (generate_loops_for_partition): Use gimple_uid instead of
10763 relying on matching stmt visit order.
10764 (rdg_build_partitions): Take starting stmt vector.
10765 (ldist_gen): Merge into ...
10766 (distribute_loop): ... this function. Do not compute starting
10768 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
10770 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10772 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
10773 Set type attribute correctly. Set predicable_short_it attribute.
10774 (cmpsi_shiftsi): Remove %? from output template.
10776 2013-09-13 Richard Biener <rguenther@suse.de>
10778 * tree-loop-distribution.c (struct rdg_component,
10779 rdg_defs_used_in_other_loops_p, free_rdg_components,
10780 rdg_build_components): Remove.
10781 (stmts_from_loop): Do not record virtual PHIs.
10782 (generate_loops_for_partition): Skip virtual PHIs.
10783 (build_rdg_partition_for_component): Rename to ...
10784 (build_rdg_partition_for_vertex): ... this and adjust.
10785 (rdg_build_partitions): Take a vector of starting vertices
10786 instead of components. Remove unnecessary leftover handling.
10787 (ldist_gen): Do not build components or record other stores.
10788 (distribute_loop): Do not distribute loops containing stmts
10791 2013-09-13 Christian Bruel <christian.bruel@st.com>
10794 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
10796 2013-09-13 Kai Tietz <ktietz@redhat.com>
10798 * config.gcc: Separate cases for mingw and cygwin targets,
10799 and add 64-bit cygwin target case.
10801 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
10802 dll-export inline-functions.
10803 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
10805 2013-09-13 Jeff Law <law@redhat.com>
10807 PR middle-end/58387
10809 2013-09-06 Jeff Law <law@redhat.com>
10811 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10812 edge implied equivalences into successor phis.
10814 2013-09-12 DJ Delorie <dj@redhat.com>
10816 * config/rl78/rl78-virt.md: Change from | to \; for asm line
10819 2013-09-12 Brooks Moses <bmoses@google.com>
10822 * Makefile.in: Do not install driver binaries in $(target)/bin.
10824 2013-09-12 DJ Delorie <dj@redhat.com>
10826 * config/rl78/rl78.opt (mrelax): New.
10827 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
10828 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
10830 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
10832 (rl78_expand_epilogue): Likewise.
10834 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
10836 PR middle-end/58335
10837 * lra-eliminations.c (remove_reg_equal_offset_note): New.
10838 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
10839 pointer elimination with using remove_reg_equal_offset_note.
10841 2013-09-12 DJ Delorie <dj@redhat.com>
10843 * config/msp430/: New port.
10844 * config.gcc (msp430): Added.
10845 * doc/invoke.texi: Document MSP430 options.
10846 * doc/install.texi: Document msp430-elf
10847 * doc/md.texi: Document msp430-elf
10848 * doc/contrib.texi: Document msp430-elf
10850 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
10853 2013-09-12 Martin Jambor <mjambor@suse.cz>
10856 * ipa-prop.c (remove_described_reference): Give up if the edge in the
10857 reference descriptor is NULL.
10858 (ipa_edge_removal_hook): If owning a reference descriptor, set its
10861 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10863 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
10864 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
10865 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
10866 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
10867 make_temp_ssa_name): move to tree-ssanames.h
10868 * tree-ssa-alias.h: Move prototype.
10869 * tree-ssa.h: Include tree-ssanames.h.
10870 * tree-ssanames.c (FREE_SSANAMES): Move to here.
10871 * tree-ssanames.h: New. Move items from tree-flow*.h
10872 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
10874 2013-09-12 Richard Biener <rguenther@suse.de>
10876 PR tree-optimization/58404
10877 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
10878 propagate non-invariant addresses into dereferences wrapped
10879 in component references.
10881 2013-09-12 Richard Biener <rguenther@suse.de>
10883 PR tree-optimization/58402
10884 * passes.def: Move pass_late_warn_uninitialized later.
10886 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10888 * tree-ssa.h: New. Move content from tree-flow.h and
10889 tree-flow-inline.h.
10890 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
10891 Move prototypes belonging to tree-ssa.c.
10892 * tree-flow-inline.h (redirect_edge_var_map_def,
10893 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
10895 * gimple.h: Adjust prototypes.
10896 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
10898 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
10899 * tree.h: Move prototype to tree-ssa.h.
10900 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
10901 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
10902 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
10903 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
10904 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
10905 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
10906 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
10907 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
10908 graphite-blocking.c, graphite-clast-to-gimple.c,
10909 graphite-dependences.c, graphite-interchange.c,
10910 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
10911 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
10912 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
10913 ipa-reference.c, ipa-split.c, ipa-utils.c,
10914 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
10915 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
10916 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
10917 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
10918 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
10919 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
10920 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
10921 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
10922 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
10923 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
10924 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
10925 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
10926 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
10927 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
10928 value-prof.c, var-tracking.c,
10929 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
10931 2013-09-12 Richard Biener <rguenther@suse.de>
10933 PR tree-optimization/58396
10934 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
10935 (build_rdg): Take a loop-nest parameter, fix memleaks.
10936 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
10938 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
10940 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
10943 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
10945 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
10946 The FMA instruction names should have a 'v' prefix.
10948 2013-09-12 Richard Biener <rguenther@suse.de>
10950 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
10951 (dot_rdg): Use popen instead of system in optional code.
10952 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
10953 (already_processed_vertex_p): Adjust.
10954 (has_anti_or_output_dependence, predecessor_has_mem_write,
10955 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
10956 rdg_flag_uses): Remove.
10957 (rdg_flag_vertex): Simplify.
10958 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
10960 (build_rdg_partition_for_component): Process the first vertex
10961 of a component only.
10962 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
10964 2013-09-12 Alan Modra <amodra@gmail.com>
10966 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
10968 2013-09-11 DJ Delorie <dj@redhat.com>
10969 Nick Clifton <nickc@redhat.com>
10971 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
10972 (rl78_stack_based_mem): New.
10973 * config/rl78/constraints.md (Iv08): New.
10982 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
10984 (movsi): Change from expand to insn-and-split.
10985 (ashrsi3): Clobber AX.
10989 * config/rl78/rl78.md (CC_REG): Fix.
10990 (addsi3): Allow memory and immediate operands.
10991 (addsi3_internal): Split into...
10992 (addsi3_internal_virt): ...new, and ...
10993 (addsi3_internal_real): ...new.
10995 (subsi3_internal_virt): New.
10996 (subsi3_internal_real): New.
10997 (mulsi3): Add memory operand.
10998 (mulsi3_rl78): Likewise.
10999 (mulsi3_g13): Likewise.
11000 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
11001 (cbranchqi4_real): Add more constraint options.
11002 (cbranchhi4_real): Expand pattern.
11003 (cbranchhi4_real_signed): New.
11004 (cbranchhi4_real_inverted): New.
11005 (cbranchsi4_real_lt): New.
11006 (cbranchsi4_real_ge): New.
11007 (cbranchsi4_real_signed): New.
11008 (cbranchsi4_real): New.
11010 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
11012 (lshrsi3_virt): Likewise.
11013 (ashlsi3_virt): Likewise.
11014 (cbranchqi4_virt_signed): New.
11015 (cbranchhi4_virt_signed): New.
11016 (cbranchsi4_virt): New.
11017 * config/rl78/rl78.c: Whitespace fixes throughout.
11018 (move_elim_pass): New.
11019 (pass_data_rl78_move_elim): New.
11020 (pass_rl78_move_elim): New.
11021 (make_pass_rl78_move_elim): New.
11022 (rl78_devirt_info): Run devirt earlier.
11023 (rl78_move_elim_info): New.
11024 (rl78_asm_file_start): Register it.
11025 (rl78_split_movsi): New.
11026 (rl78_as_legitimate_address): Allow virtual base registers when
11028 (rl78_addr_space_convert): Remove spurious debug stuff.
11029 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
11030 (rl78_print_operand): More cases for not printing '#'.
11031 (rl78_expand_compare): Remove most of the logic.
11032 (content_memory): New.
11033 (clear_content_memory): New.
11034 (get_content_index): New.
11035 (get_content_name): New.
11036 (display_content_memory): New.
11037 (update_content): New.
11038 (record_content): New.
11039 (already_contains): New.
11040 (insn_ok_now): Re-recog insns with virtual registers.
11041 (add_postponed_content_update): New.
11042 (process_postponed_content_update): New.
11043 (gen_and_emit_move): New.
11044 (transcode_memory_rtx): Record new location content.
11045 Use gen_and_emit_move.
11046 (force_into_acc): New.
11047 (move_to_acc): Use gen_and_emit_move.
11048 (move_from_acc): Likewise.
11049 (move_acc_to_reg): Likewise.
11050 (move_to_x): Likewise.
11051 (move_to_hl): Likewise.
11052 (move_to_de): Likewise.
11053 (rl78_alloc_physical_registers_op1): Record location content.
11054 (has_constraint): New.
11055 (rl78_alloc_physical_registers_op2): Record location content.
11056 Optimize use of HL.
11057 (rl78_alloc_physical_registers_ro1): Likewise.
11058 (rl78_alloc_physical_registers_cmp): Likewise.
11059 (rl78_alloc_physical_registers_umul): Likewise.
11060 (rl78_alloc_address_registers_macax): New.
11061 (rl78_alloc_physical_registers): Initialize and set location
11062 content memory as needed.
11063 (rl78_reorg): Make sure split2 is called.
11064 (rl78_rtx_costs): New.
11066 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
11068 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
11069 for (not (neg ...)) and (neg (not ...)) cases.
11071 2013-09-11 Richard Biener <rguenther@suse.de>
11073 PR middle-end/58377
11074 * passes.def: Split critical edges before late uninit warning passes.
11075 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
11077 2013-09-11 Jakub Jelinek <jakub@redhat.com>
11079 PR tree-optimization/58385
11080 * fold-const.c (build_range_check): If both low and high are NULL,
11081 use omit_one_operand_loc to preserve exp side-effects.
11083 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11085 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
11087 2013-09-11 Richard Biener <rguenther@suse.de>
11089 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
11090 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
11091 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
11092 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
11093 stmts_from_loop, known_dependences_p, build_empty_rdg,
11094 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
11095 * tree-loop-distribution.c: ... here.
11096 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
11097 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
11098 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
11099 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
11100 * tree-loop-distribution.c: ... here.
11101 * tree-loop-distribution.c: Include gimple-pretty-print.h.
11102 (struct partition_s): Add loops member.
11103 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
11104 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
11105 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
11107 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
11108 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11109 Sergey Lega <sergey.s.lega@intel.com>
11110 Anna Tikhonova <anna.tikhonova@intel.com>
11111 Ilya Tocar <ilya.tocar@intel.com>
11112 Andrey Turetskiy <andrey.turetskiy@intel.com>
11113 Ilya Verbin <ilya.verbin@intel.com>
11114 Kirill Yukhin <kirill.yukhin@intel.com>
11115 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11117 * config/i386/constraints.md (k): New.
11119 * config/i386/i386.c (const regclass_map): Add new mask registers.
11120 (dbx_register_map): Ditto.
11121 (dbx64_register_map): Ditto.
11122 (svr4_dbx_register_map): Ditto.
11123 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
11125 (ix86_preferred_reload_class): Disable constants for mask registers.
11126 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
11127 (ix86_hard_regno_mode_ok): Support new mask registers.
11128 (x86_order_regs_for_local_alloc): Ditto.
11129 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
11130 (FIXED_REGISTERS): Add new mask registers.
11131 (CALL_USED_REGISTERS): Ditto.
11132 (REG_ALLOC_ORDER): Ditto.
11133 (VALID_MASK_REG_MODE): New.
11134 (FIRST_MASK_REG): Ditto.
11135 (LAST_MASK_REG): Ditto.
11136 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
11137 (MAYBE_MASK_CLASS_P): New.
11138 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
11139 (REG_CLASS_CONTENTS): Ditto.
11140 (MASK_REGNO_P): New.
11141 (ANY_MASK_REG_P): Ditto.
11142 (HI_REGISTER_NAMES): Add new mask registers.
11143 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
11144 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
11146 (attribute "type"): Add mskmov, msklog.
11147 (attribute "length_immediate"): Support them.
11148 (attribute "memory"): Ditto.
11149 (attribute "prefix_0f"): Ditto.
11150 (*movhi_internal): Support new mask registers.
11151 (*movqi_internal): Ditto.
11152 (define_split): Split out clobber pattern is a logic
11153 insn on mask registers.
11154 (*k<logic><mode>): New.
11155 (*andhi_1): Extend to support mask regs.
11156 (*andqi_1): Extend to support mask regs.
11157 (kandn<mode>): New.
11158 (define_split): Split and-not to and and not if operands
11160 (*<code><mode>_1): Separate HI mode to new pattern...
11161 (*<code>hi_1): This.
11162 (*<code>qi_1): Extend to support mask regs.
11163 (kxnor<mode>): New.
11164 (kortestzhi): Ditto.
11165 (kortestchi): Ditto.
11167 (*one_cmpl<mode>2_1): Remove HImode and handle it...
11168 (*one_cmplhi2_1): ...Here, now with mask registers support.
11169 (*one_cmplqi2_1): Support new mask registers.
11170 (HI/QImode arithmetics splitter): Don't split if mask registers
11172 (HI/QImode not splitter): Ditto.
11173 * config/i386/predicated.md (mask_reg_operand): New.
11174 (general_reg_operand): Ditto.
11176 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
11178 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
11179 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
11181 2013-09-10 Jeff Law <law@redhat.com>
11183 PR tree-optimization/58380
11184 * tree-ssa-threadupdate.c (thread_block): Recognize another case
11185 of threading through a buried loop header.
11187 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
11188 return value for single successor case.
11190 2013-09-10 Jan Hubicka <jh@suse.cz>
11192 * ipa-devirt.c (ipa_devirt): Enable with LTO.
11194 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
11197 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
11198 support conditional execution.
11199 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
11201 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
11203 * lra.c (lra): Clear lra_optional_reload_pseudos before every
11205 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
11206 Check destination too to check move insn.
11207 (undo_optional_reloads): Add check that the original peudo did not
11208 changed its allocation and the optional reload was inherited on last
11209 inheritance pass. Break loop after deciding to keep optional reload.
11210 (lra_undo_inheritance): Add check that inherited pseudo still in
11213 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
11215 * config/aarch64/aarch64.md (generic_sched): New.
11216 * config/aarch64/aarch64-generic.md (load): Make conditional
11217 on generic_sched attribute.
11218 (nonload): Likewise.
11220 2013-09-10 Jan Hubicka <jh@suse.cz>
11222 * lto-cgraph.c: Include ipa-utils.h.
11223 (compute_ltrans_boundary): Also add possible targets into the boundary.
11225 2013-09-10 Jan Hubicka <jh@suse.cz>
11227 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
11228 VAR_DECL of vtable rather than full expression.
11230 2013-09-10 Jan Hubicka <jh@suse.cz>
11231 Paolo Carlini <paolo.carlini@oracle.com>
11233 * cgraphunit.c (analyze_functions): Save input_location, set it
11234 to UNKNOWN_LOCATION and restore it at the end.
11236 2013-09-10 Martin Jambor <mjambor@suse.cz>
11238 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
11239 represented by a thunk.
11241 2013-09-10 Jeff Law <law@redhat.com>
11243 PR tree-optimization/58343
11244 * tree-ssa-threadupdate.c (thread_block): Identify and disable
11245 jump threading requests through loop headers buried in the middle
11246 of a jump threading path.
11248 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
11249 in return value/type.
11251 2013-09-10 Jakub Jelinek <jakub@redhat.com>
11253 PR rtl-optimization/58365
11254 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
11255 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
11258 2013-09-10 Richard Biener <rguenther@suse.de>
11260 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
11261 * tree-data-ref.c (create_rdg_vertices): Collect all data
11262 references, signal failure to the caller, use data-ref API.
11263 (build_rdg): Compute data references only once. Maintain lifetime
11264 of data references and data dependences from within RDG.
11265 (free_rdg): Free dependence relations.
11266 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
11267 inventing extra dependences.
11268 (distribute_loop): Update for RDG API changes.
11270 2013-09-10 Kai Tietz <ktietz@redhat.com>
11272 * doc/invoke.texi (fms-extensions): Document changed
11273 behavior for ms-abi targets.
11274 * config/i386/i386.c (ix86_option_override_internal):
11275 Set default value of option -fms-extension for ms-abi targets.
11277 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
11279 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
11281 2013-09-10 Alan Modra <amodra@gmail.com>
11284 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
11286 2013-09-10 Alan Modra <amodra@gmail.com>
11288 * config/rs6000/predicates.md (add_cint_operand): New.
11289 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
11290 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
11291 using add_cint_operand.
11292 (largetoc_high_plus_aix): Likewise.
11294 2013-09-09 Jakub Jelinek <jakub@redhat.com>
11296 PR tree-optimization/58364
11297 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
11298 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
11299 the current range can't be an unconditional true or false.
11301 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
11303 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
11305 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
11307 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
11309 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
11312 * doc/invoke.texi (-Wdelete-incomplete): Document it.
11314 2013-09-09 Ian Bolton <ian.bolton@arm.com>
11316 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
11317 NO_REGS for immediate that can't be moved directly into FP_REGS.
11319 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11321 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
11322 comparison with negated operand.
11323 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
11326 2013-09-09 Richard Biener <rguenther@suse.de>
11328 PR middle-end/58326
11329 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
11330 of a subloop record all its block as affecting loop-closed SSA form.
11332 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11334 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
11335 of an rtx/bitpos pair.
11336 (store_fixed_bit_field): Update accordingly.
11338 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11340 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
11342 * builtins.c (expand_errno_check): Likewise.
11343 * dwarf2cfi.c (init_return_column_size): Likewise.
11344 * except.c (sjlj_mark_call_sites): Likewise.
11345 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
11346 * lra-constraints.c (emit_inc): Likewise.
11347 * ree.c (combine_set_extension): Likewise.
11348 * regmove.c (fixup_match_2): Likewise.
11349 * reload1.c (inc_for_reload): Likewise.
11351 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11353 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
11354 (force_to_mode, simplify_shift_const_1, simplify_comparison):
11355 Use gen_int_mode with the mode of the associated simplify_* call.
11356 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
11357 * expmed.c (expand_shift_1): Likewise.
11358 * function.c (instantiate_virtual_regs_in_insn): Likewise.
11359 * loop-iv.c (iv_number_of_iterations): Likewise.
11360 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
11361 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
11363 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11365 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
11366 of the associated expand_* call.
11367 (asan_emit_stack_protection): Likewise.
11368 * builtins.c (round_trampoline_addr): Likewise.
11369 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
11370 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
11371 (emit_store_flag): Likewise.
11372 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
11373 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
11375 * function.c (instantiate_virtual_regs_in_insn): Likewise.
11376 * ifcvt.c (noce_try_store_flag_constants): Likewise.
11377 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
11378 * modulo-sched.c (generate_prolog_epilog): Likewise.
11379 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
11380 (expand_ctz, expand_ffs, expand_unop): Likewise.
11382 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11384 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
11385 of the associated gen_rtx_* call.
11386 * caller-save.c (init_caller_save): Likewise.
11387 * combine.c (find_split_point, make_extraction): Likewise.
11388 (make_compound_operation): Likewise.
11389 * dwarf2out.c (mem_loc_descriptor): Likewise.
11390 * explow.c (plus_constant, probe_stack_range): Likewise.
11391 * expmed.c (expand_mult_const): Likewise.
11392 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
11393 * reload1.c (init_reload): Likewise.
11394 * valtrack.c (cleanup_auto_inc_dec): Likewise.
11395 * var-tracking.c (adjust_mems): Likewise.
11396 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
11397 rather than gen_rtx_fmt_ee.
11399 2013-09-09 Jan Hubicka <jh@suse.cz>
11401 PR middle-end/58294
11402 * value-prof.c (gimple_ic): Copy also abnormal edges.
11404 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11406 * asan.c (asan_shadow_cst): Use gen_int_mode.
11408 2013-09-08 Jan Hubicka <jh@suse.cz>
11410 * ipa-profile.c: Add toplevel comment.
11411 (ipa_propagate_frequency_1): Be more conservative when profile is read.
11412 (contains_hot_call_p): New function.
11413 (ipa_propagate_frequency): Set frequencies based on counts when
11415 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
11416 profile; do not tamper with profile after inlining if it is read.
11418 2013-09-08 Jan Hubicka <jh@suse.cz>
11420 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
11423 2013-09-08 Jan Hubicka <jh@suse.cz>
11425 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
11426 summary generation.
11428 2013-09-08 Jeff Law <law@redhat.com>
11431 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
11434 2013-09-08 Andi Kleen <ak@linux.intel.com>
11436 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
11438 2013-09-08 Jan Hubicka <jh@suse.cz>
11440 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
11442 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
11444 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
11445 for non-debug insns.
11446 * lra.c (new_insn_reg): Take the containing insn as a parameter.
11447 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
11448 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
11451 2013-09-08 Jan Hubicka <jh@suse.cz>
11453 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
11454 targets and devirtualize to BUILT_IN_UNREACHABLE.
11455 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
11456 * ipa.c (walk_polymorphic_call_targets): New function.
11457 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
11458 functions; use the new timevar.
11459 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
11460 was removed from the program.
11461 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
11462 not consider it in the walk when its vtable is dead.
11463 (possible_polymorphic_call_targets_1): Pass anonymous flag to
11465 (devirt_variable_node_removal_hook): New function.
11466 (possible_polymorphic_call_targets): Also register
11467 devirt_variable_node_removal_hook.
11468 (ipa_devirt): Do not do non-speculative devirtualization.
11469 (gate_ipa_devirt): One execute if devirtualizing speculatively.
11471 2013-09-08 Jan Hubicka <jh@suse.cz>
11473 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
11474 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
11475 varpool_remove_variable_insertion_hook): Declare.
11476 * varpool.c (varpool_node_hook_list): New structure.
11477 (first_varpool_node_removal_hook,
11478 first_varpool_variable_insertion_hook): New variables.
11479 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
11480 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
11481 varpool_remove_variable_insertion_hook,
11482 varpool_call_variable_insertion_hooks): New functions.
11483 (varpool_remove_node): Use it.
11485 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
11488 * diagnostic.c (diagnostic_build_prefix): When s.file is
11489 "<built-in>" don't output line and column numbers.
11491 2013-09-06 Jan Hubicka <jh@suse.cz>
11493 * cgraphunit.c (expand_thunk): Get body before touching arguments.
11494 * lto-streamer-out.c: Stream thunks, too.
11495 * lto-streamer-in.c (input_function): Pop cfun here
11496 (lto_read_body): Instead of here.
11498 2013-09-06 Caroline Tice <cmtice@google.com>
11500 * doc/install.texi: Add documentation for the --enable-vtable-verify
11501 and the --disable-libvtv configure options.
11503 2013-09-06 Jeff Law <law@redhat.com>
11505 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
11506 edge implied equivalences into successor phis.
11508 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
11510 * resource.c (mark_referenced_resources): Handle COND_EXEC.
11512 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
11514 * resource.c (mark_target_live_regs): Compute resources taking
11515 into account if a call is predicated or not.
11517 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
11519 * toplev.c (output_stack_usage): Be prepared for suffixes created by
11520 the compiler in the function names.
11522 2013-09-06 Jan Hubicka <jh@suse.cz>
11524 PR middle-end/58094
11525 * ipa-inline.c (has_caller_p): New function.
11526 (want_inline_function_to_all_callers_p): Use it.
11527 (sum_callers, inline_to_all_callers): Break out from ...
11528 (ipa_inline): ... here.
11530 2013-09-06 Jan Hubicka <jh@suse.cz>
11532 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
11533 only when AVX is enabled.
11535 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11537 * config/aarch64/aarch64.md
11538 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
11539 is fpsimd_<load/store>2.
11540 (load_pair<mode>): Likewise.
11541 (store_pair<mode>): Likewise.
11543 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11545 * config/arm/types.md (type): Add "mrs" type.
11546 * config/aarch64/aarch64.md
11547 (aarch64_load_tp_hard): Make type "mrs".
11548 * config/arm/arm.md
11549 (load_tp_hard): Make type "mrs".
11550 * config/arm/cortex-a15.md: Update with new attributes.
11551 * config/arm/cortex-a5.md: Update with new attributes.
11552 * config/arm/cortex-a53.md: Update with new attributes.
11553 * config/arm/cortex-a7.md: Update with new attributes.
11554 * config/arm/cortex-a8.md: Update with new attributes.
11555 * config/arm/cortex-a9.md: Update with new attributes.
11556 * config/arm/cortex-m4.md: Update with new attributes.
11557 * config/arm/cortex-r4.md: Update with new attributes.
11558 * config/arm/fa526.md: Update with new attributes.
11559 * config/arm/fa606te.md: Update with new attributes.
11560 * config/arm/fa626te.md: Update with new attributes.
11561 * config/arm/fa726te.md: Update with new attributes.
11563 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11565 * config/aarch64/aarch64.md
11566 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
11567 (*movtf_aarch64): Likewise.
11568 * config/arm/arm.md
11569 (thumb1_movdi_insn): Use "multiple" for type where more than one
11570 instruction is used for a move.
11571 (*arm32_movhf): Likewise.
11572 (*thumb_movdf_insn): Likewise.
11574 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11576 * config/arm/types.md (type): Rename fcpys to fmov.
11577 * config/arm/vfp.md
11578 (*arm_movsi_vfp): Rename type fcpys as fmov.
11579 (*thumb2_movsi_vfp): Likewise
11580 (*movhf_vfp_neon): Likewise
11581 (*movhf_vfp): Likewise
11582 (*movsf_vfp): Likewise
11583 (*thumb2_movsf_vfp): Likewise
11584 (*movsfcc_vfp): Likewise
11585 (*thumb2_movsfcc_vfp): Likewise
11586 * config/aarch64/aarch64-simd.md
11587 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
11588 * config/aarch64/aarch64.md
11589 (*movsi_aarch64): Replace type mov_reg with fmovs.
11590 (*movdi_aarch64): Likewise
11591 (*movsf_aarch64): Likewise
11592 (*movdf_aarch64): Likewise
11594 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
11595 * config/arm/iwmmxt.md
11596 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
11597 * config/arm/arm1020e.md: Update with new attributes.
11598 * config/arm/cortex-a15-neon.md: Update with new attributes.
11599 * config/arm/cortex-a5.md: Update with new attributes.
11600 * config/arm/cortex-a53.md: Update with new attributes.
11601 * config/arm/cortex-a7.md: Update with new attributes.
11602 * config/arm/cortex-a8-neon.md: Update with new attributes.
11603 * config/arm/cortex-a9.md: Update with new attributes.
11604 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11605 * config/arm/cortex-r4f.md: Update with new attributes.
11606 * config/arm/marvell-pj4.md: Update with new attributes.
11607 * config/arm/vfp11.md: Update with new attributes.
11609 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11611 * config/aarch64/aarch64.md
11612 (*madd<mode>): Fix type attribute.
11613 (*maddsi_uxtw): Likewise.
11614 (*msub<mode>): Likewise.
11615 (*msubsi_uxtw): Likewise.
11616 (<su_optab>maddsidi4): Likewise.
11617 (<su_optab>msubsidi4): Likewise.
11619 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11621 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
11622 * config/arm/arm.md (core_cycles): Remove fdiv.
11623 * config/arm/vfp.md:
11624 (*sqrtsf2_vfp): Update for attribute changes.
11625 (*sqrtdf2_vfp): Likewise.
11626 * config/aarch64/aarch64.md:
11627 (sqrt<mode>2): Update for attribute changes.
11628 * config/arm/arm1020e.md: Update with new attributes.
11629 * config/arm/cortex-a15-neon.md: Update with new attributes.
11630 * config/arm/cortex-a5.md: Update with new attributes.
11631 * config/arm/cortex-a53.md: Update with new attributes.
11632 * config/arm/cortex-a7.md: Update with new attributes.
11633 * config/arm/cortex-a8-neon.md: Update with new attributes.
11634 * config/arm/cortex-a9.md: Update with new attributes.
11635 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11636 * config/arm/cortex-r4f.md: Update with new attributes.
11637 * config/arm/marvell-pj4.md: Update with new attributes.
11638 * config/arm/vfp11.md: Update with new attributes.
11640 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11642 * config/arm/types.md
11643 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
11644 * config/aarch64/aarch64.md
11645 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
11647 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
11648 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
11649 (float<GPI:mode><GPF:mode>2): Likewise.
11650 * config/arm/vfp.md
11651 (*truncsisf2_vfp): Update with new attributes.
11652 (*truncsidf2_vfp): Likewise.
11653 (fixuns_truncsfsi2): Likewise.
11654 (fixuns_truncdfsi2): Likewise.
11655 (*floatsisf2_vfp): Likewise.
11656 (*floatsidf2_vfp): Likewise.
11657 (floatunssisf2): Likewise.
11658 (floatunssidf2): Likewise.
11659 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
11660 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
11661 * config/arm/arm1020e.md: Update with new attributes.
11662 * config/arm/cortex-a15-neon.md: Update with new attributes.
11663 * config/arm/cortex-a5.md: Update with new attributes.
11664 * config/arm/cortex-a53.md: Update with new attributes.
11665 * config/arm/cortex-a7.md: Update with new attributes.
11666 * config/arm/cortex-a8-neon.md: Update with new attributes.
11667 * config/arm/cortex-a9.md: Update with new attributes.
11668 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11669 * config/arm/cortex-r4f.md: Update with new attributes.
11670 * config/arm/marvell-pj4.md: Update with new attributes.
11671 * config/arm/vfp11.md: Update with new attributes.
11673 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11675 * config/aarch64/arm_neon.h
11676 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
11677 (vqtbx<1,2,3,4><q>_s8): Likewise.
11679 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11681 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
11682 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
11683 (add<mode>3): Add type attribute.
11684 (add<mode>3): Likewise.
11685 (usadd<mode>3): Likewise.
11686 (ssadd<mode>3): Likewise.
11687 (sub<mode>3): Likewise.
11688 (sub<mode>3): Likewise.
11689 (ussub<mode>3): Likewise.
11690 (sssub<mode>3): Likewise.
11691 (ssmulsa3): Likewise.
11692 (usmulusa3): Likewise.
11693 (arm_usatsihi): Likewise.
11694 * config/arm/vfp.md
11695 (*movdi_vfp): Add types for all instructions.
11696 (*movdi_vfp_cortexa8): Likewise.
11697 (*movhf_vfp_neon): Likewise.
11698 (*movhf_vfp): Likewise.
11699 (*movdf_vfp): Likewise.
11700 (*thumb2_movdf_vfp): Likewise.
11701 (*thumb2_movdfcc_vfp): Likewise.
11702 * config/arm/arm.md: Add type attribute to all insn patterns.
11703 (*thumb1_adddi3): Add type attribute.
11704 (*arm_adddi3): Likewise.
11705 (*adddi_sesidi_di): Likewise.
11706 (*adddi_zesidi_di): Likewise.
11707 (*thumb1_addsi3): Likewise.
11708 (addsi3_compare0): Likewise.
11709 (*addsi3_compare0_scratch): Likewise.
11710 (*compare_negsi_si): Likewise.
11711 (cmpsi2_addneg): Likewise.
11712 (*addsi3_carryin_<optab>): Likewise.
11713 (*addsi3_carryin_alt2_<optab>): Likewise.
11714 (*addsi3_carryin_clobercc_<optab>): Likewise.
11715 (*subsi3_carryin): Likewise.
11716 (*subsi3_carryin_const): Likewise.
11717 (*subsi3_carryin_compare): Likewise.
11718 (*subsi3_carryin_compare_const): Likewise.
11719 (*arm_subdi3): Likewise.
11720 (*thumb_subdi3): Likewise.
11721 (*subdi_di_zesidi): Likewise.
11722 (*subdi_di_sesidi): Likewise.
11723 (*subdi_zesidi_di): Likewise.
11724 (*subdi_sesidi_di): Likewise.
11725 (*subdi_zesidi_ze): Likewise.
11726 (thumb1_subsi3_insn): Likewise.
11727 (*arm_subsi3_insn): Likewise.
11728 (*anddi3_insn): Likewise.
11729 (*anddi_zesidi_di): Likewise.
11730 (*anddi_sesdi_di): Likewise.
11731 (*ne_zeroextracts): Likewise.
11732 (*ne_zeroextracts): Likewise.
11733 (*ite_ne_zeroextr): Likewise.
11734 (*ite_ne_zeroextr): Likewise.
11735 (*anddi_notdi_di): Likewise.
11736 (*anddi_notzesidi): Likewise.
11737 (*anddi_notsesidi): Likewise.
11738 (andsi_notsi_si): Likewise.
11739 (thumb1_bicsi3): Likewise.
11740 (*iordi3_insn): Likewise.
11741 (*iordi_zesidi_di): Likewise.
11742 (*iordi_sesidi_di): Likewise.
11743 (*thumb1_iorsi3_insn): Likewise.
11744 (*xordi3_insn): Likewise.
11745 (*xordi_zesidi_di): Likewise.
11746 (*xordi_sesidi_di): Likewise.
11747 (*arm_xorsi3): Likewise.
11748 (*andsi_iorsi3_no): Likewise.
11749 (*smax_0): Likewise.
11750 (*smax_m1): Likewise.
11751 (*arm_smax_insn): Likewise.
11752 (*smin_0): Likewise.
11753 (*arm_smin_insn): Likewise.
11754 (*arm_umaxsi3): Likewise.
11755 (*arm_uminsi3): Likewise.
11756 (*minmax_arithsi): Likewise.
11757 (*minmax_arithsi_): Likewise.
11758 (*satsi_<SAT:code>): Likewise.
11759 (arm_ashldi3_1bit): Likewise.
11760 (arm_ashrdi3_1bit): Likewise.
11761 (arm_lshrdi3_1bit): Likewise.
11762 (*arm_negdi2): Likewise.
11763 (*thumb1_negdi2): Likewise.
11764 (*arm_negsi2): Likewise.
11765 (*thumb1_negsi2): Likewise.
11766 (*negdi_extendsid): Likewise.
11767 (*negdi_zero_extend): Likewise.
11768 (*arm_abssi2): Likewise.
11769 (*thumb1_abssi2): Likewise.
11770 (*arm_neg_abssi2): Likewise.
11771 (*thumb1_neg_abss): Likewise.
11772 (one_cmpldi2): Likewise.
11773 (extend<mode>di2): Likewise.
11774 (*compareqi_eq0): Likewise.
11775 (*arm_extendhisi2addsi): Likewise.
11776 (*arm_movdi): Likewise.
11777 (*thumb1_movdi_insn): Likewise.
11778 (*arm_movt): Likewise.
11779 (*thumb1_movsi_insn): Likewise.
11780 (pic_add_dot_plus_four): Likewise.
11781 (pic_add_dot_plus_eight): Likewise.
11782 (tls_load_dot_plus_eight): Likewise.
11783 (*thumb1_movhi_insn): Likewise.
11784 (*thumb1_movsf_insn): Likewise.
11785 (*movdf_soft_insn): Likewise.
11786 (*thumb_movdf_insn): Likewise.
11787 (cbranchsi4_insn): Likewise.
11788 (cbranchsi4_scratch): Likewise.
11789 (*negated_cbranchsi4): Likewise.
11790 (*tbit_cbranch): Likewise.
11791 (*tlobits_cbranch): Likewise.
11792 (*tstsi3_cbranch): Likewise.
11793 (*cbranchne_decr1): Likewise.
11794 (*addsi3_cbranch): Likewise.
11795 (*addsi3_cbranch_scratch): Likewise.
11796 (*arm_cmpdi_insn): Likewise.
11797 (*arm_cmpdi_unsig): Likewise.
11798 (*arm_cmpdi_zero): Likewise.
11799 (*thumb_cmpdi_zero): Likewise.
11800 (*deleted_compare): Likewise.
11801 (*mov_scc): Likewise.
11802 (*mov_negscc): Likewise.
11803 (*mov_notscc): Likewise.
11804 (*cstoresi_eq0_thumb1_insn): Likewise.
11805 (cstoresi_nltu_thumb1): Likewise.
11806 (cstoresi_ltu_thu): Likewise.
11807 (thumb1_addsi3_addgeu): Likewise.
11808 (*arm_jump): Likewise.
11809 (*thumb_jump): Likewise.
11810 (*check_arch2): Likewise.
11811 (arm_casesi_internal): Likewise.
11812 (thumb1_casesi_dispatch): Likewise.
11813 (*arm_indirect_jump): Likewise.
11814 (*thumb1_indirect_jump): Likewise.
11816 (*and_scc): Likewise.
11817 (*ior_scc): Likewise.
11818 (*compare_scc): Likewise.
11819 (*cond_move): Likewise.
11820 (*cond_arith): Likewise.
11821 (*cond_sub): Likewise.
11822 (*cmp_ite0): Likewise.
11823 (*cmp_ite1): Likewise.
11824 (*cmp_and): Likewise.
11825 (*cmp_ior): Likewise.
11826 (*ior_scc_scc): Likewise.
11827 (*ior_scc_scc_cmp): Likewise.
11828 (*and_scc_scc): Likewise.
11829 (*and_scc_scc_cmp): Likewise.
11830 (*and_scc_scc_nod): Likewise.
11831 (*negscc): Likewise.
11832 (movcond_addsi): Likewise.
11833 (movcond): Likewise.
11834 (*ifcompare_plus_move): Likewise.
11835 (*if_plus_move): Likewise.
11836 (*ifcompare_move_plus): Likewise.
11837 (*if_move_plus): Likewise.
11838 (*ifcompare_arith_arith): Likewise.
11839 (*if_arith_arith): Likewise.
11840 (*ifcompare_arith_move): Likewise.
11841 (*if_arith_move): Likewise.
11842 (*ifcompare_move_arith): Likewise.
11843 (*if_move_arith): Likewise.
11844 (*ifcompare_move_not): Likewise.
11845 (*if_move_not): Likewise.
11846 (*ifcompare_not_move): Likewise.
11847 (*if_not_move): Likewise.
11848 (*ifcompare_shift_move): Likewise.
11849 (*if_shift_move): Likewise.
11850 (*ifcompare_move_shift): Likewise.
11851 (*if_move_shift): Likewise.
11852 (*ifcompare_shift_shift): Likewise.
11853 (*ifcompare_not_arith): Likewise.
11854 (*ifcompare_arith_not): Likewise.
11855 (*if_arith_not): Likewise.
11856 (*ifcompare_neg_move): Likewise.
11857 (*if_neg_move): Likewise.
11858 (*ifcompare_move_neg): Likewise.
11859 (*if_move_neg): Likewise.
11860 (prologue_thumb1_interwork): Likewise.
11861 (*cond_move_not): Likewise.
11862 (*sign_extract_onebit): Likewise.
11863 (*not_signextract_onebit): Likewise.
11864 (stack_tie): Likewise.
11865 (align_4): Likewise.
11866 (align_8): Likewise.
11867 (consttable_end): Likewise.
11868 (consttable_1): Likewise.
11869 (consttable_2): Likewise.
11870 (consttable_4): Likewise.
11871 (consttable_8): Likewise.
11872 (consttable_16): Likewise.
11873 (*thumb1_tablejump): Likewise.
11874 (prefetch): Likewise.
11875 (force_register_use): Likewise.
11876 (thumb_eh_return): Likewise.
11877 (load_tp_hard): Likewise.
11878 (load_tp_soft): Likewise.
11879 (tlscall): Likewise.
11880 (*arm_movtas_ze): Likewise.
11881 (*arm_rev): Likewise.
11882 (*arm_revsh): Likewise.
11883 (*arm_rev16): Likewise.
11884 * config/arm/thumb2.md
11885 (*thumb2_smaxsi3): Likewise.
11886 (*thumb2_sminsi3): Likewise.
11887 (*thumb32_umaxsi3): Likewise.
11888 (*thumb2_uminsi3): Likewise.
11889 (*thumb2_negdi2): Likewise.
11890 (*thumb2_abssi2): Likewise.
11891 (*thumb2_neg_abss): Likewise.
11892 (*thumb2_movsi_insn): Likewise.
11893 (tls_load_dot_plus_four): Likewise.
11894 (*thumb2_movhi_insn): Likewise.
11895 (*thumb2_mov_scc): Likewise.
11896 (*thumb2_mov_negs): Likewise.
11897 (*thumb2_mov_negs): Likewise.
11898 (*thumb2_mov_nots): Likewise.
11899 (*thumb2_mov_nots): Likewise.
11900 (*thumb2_movsicc_): Likewise.
11901 (*thumb2_movsfcc_soft_insn): Likewise.
11902 (*thumb2_indirect_jump): Likewise.
11903 (*thumb2_and_scc): Likewise.
11904 (*thumb2_ior_scc): Likewise.
11905 (*thumb2_ior_scc_strict_it): Likewise.
11906 (*thumb2_cond_move): Likewise.
11907 (*thumb2_cond_arith): Likewise.
11908 (*thumb2_cond_ari): Likewise.
11909 (*thumb2_cond_sub): Likewise.
11910 (*thumb2_negscc): Likewise.
11911 (*thumb2_movcond): Likewise.
11912 (thumb2_casesi_internal): Likewise.
11913 (thumb2_casesi_internal_pic): Likewise.
11914 (*thumb2_alusi3_short): Likewise.
11915 (*thumb2_mov<mode>_shortim): Likewise.
11916 (*thumb2_addsi_short): Likewise.
11917 (*thumb2_subsi_short): Likewise.
11918 (thumb2_addsi3_compare0): Likewise.
11919 (*thumb2_cbz): Likewise.
11920 (*thumb2_cbnz): Likewise.
11921 (*thumb2_one_cmplsi2_short): Likewise.
11922 (*thumb2_negsi2_short): Likewise.
11923 (*orsi_notsi_si): Likewise.
11924 * config/arm/arm1020e.md: Update with new attributes.
11925 * config/arm/arm1026ejs.md: Update with new attributes.
11926 * config/arm/arm1136jfs.md: Update with new attributes.
11927 * config/arm/arm926ejs.md: Update with new attributes.
11928 * config/arm/cortex-a15.md: Update with new attributes.
11929 * config/arm/cortex-a5.md: Update with new attributes.
11930 * config/arm/cortex-a53.md: Update with new attributes.
11931 * config/arm/cortex-a7.md: Update with new attributes.
11932 * config/arm/cortex-a8.md: Update with new attributes.
11933 * config/arm/cortex-a9.md: Update with new attributes.
11934 * config/arm/cortex-m4.md: Update with new attributes.
11935 * config/arm/cortex-r4.md: Update with new attributes.
11936 * config/arm/fa526.md: Update with new attributes.
11937 * config/arm/fa606te.md: Update with new attributes.
11938 * config/arm/fa626te.md: Update with new attributes.
11939 * config/arm/fa726te.md: Update with new attributes.
11941 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11943 * config/aarch64/aarch64-simd.md
11944 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
11945 <vwx> iterator to ensure correct register choice.
11946 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
11947 (aarch64_sqdmull_n<mode>): Likewise.
11948 (aarch64_sqdmull2_n<mode>_internal): Likewise.
11949 * config/aarch64/arm_neon.h
11950 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
11951 (vml<as><q>_n_<su>16): Likewise.
11952 (vml<as>l_high_lane<q>_<su>16): Likewise.
11953 (vml<as>l_high_n_<su>16): Likewise.
11954 (vml<as>l_lane<q>_<su>16): Likewise.
11955 (vml<as>l_n_<su>16): Likewise.
11956 (vmul<q>_lane<q>_<su>16): Likewise.
11957 (vmul<q>_n_<su>16): Likewise.
11958 (vmull_lane<q>_<su>16): Likewise.
11959 (vmull_n_<su>16): Likewise.
11960 (vmull_high_lane<q>_<su>16): Likewise.
11961 (vmull_high_n_<su>16): Likewise.
11962 (vqrdmulh<q>_n_s16): Likewise.
11964 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
11966 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
11967 have the correct lane parameter.
11969 2013-09-06 Richard Biener <rguenther@suse.de>
11971 * cfganal.c (control_dependences::~control_dependences):
11972 Properly free all of the vector.
11974 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
11977 * config/i386/i386.c (ix86_conditional_register_usage):
11978 Proper initialize extended SSE registers.
11980 2013-09-06 Jan Hubicka <jh@suse.cz>
11982 PR tree-optimization/58311
11983 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
11985 2013-09-06 Jan Hubicka <jh@suse.cz>
11987 * Makefile.in (tree-sra.o): Update dependencies.
11988 * tree-sra.c: Include ipa-utils.h
11989 (scan_function): Use recursive_call_p.
11990 (has_caller_p): New function.
11991 (cgraph_for_node_and_aliases): Count also callers of aliases.
11993 2013-09-06 Jan Hubicka <jh@suse.cz>
11995 PR middle-end/58094
11996 * cgraph.h (symtab_semantically_equivalent_p): Declare.
11997 * tree-tailcall.c: Include ipa-utils.h.
11998 (find_tail_calls): Use it.
11999 * ipa-pure-const.c (check_call): Likewise.
12000 * ipa-utils.c (recursive_call_p): New function.
12001 * ipa-utils.h (recursive_call_p): Dclare.
12002 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
12003 (symtab_semantically_equivalent_p): New function.
12004 * Makefile.in (tree-tailcall.o): Update dependencies.
12006 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
12008 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
12009 non-inlinable part.
12011 2013-09-06 Richard Biener <rguenther@suse.de>
12013 * lto-streamer.h (lto_global_var_decls): Remove.
12014 * Makefile.in (OBJS): Remove lto-symtab.o.
12015 (lto-symtab.o): Remove.
12016 (GTFILES): Remove lto-symtab.c
12017 * lto-symtab.c: Move to lto/
12019 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12021 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
12022 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
12023 (UNSPEC_FPINT_RINT): New constant definitions.
12024 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
12025 definition with 2 attributes.
12026 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
12027 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
12030 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12032 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
12033 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
12034 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
12035 attribute to "z196_alone".
12036 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
12039 2013-09-06 Richard Biener <rguenther@suse.de>
12041 * basic-block.h (class control_dependences): New.
12042 * tree-ssa-dce.c (control_dependence_map): Remove.
12044 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
12045 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
12046 find_pdom, find_control_dependence, find_all_control_dependences):
12048 (mark_control_dependent_edges_necessary,
12049 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
12050 tree_dce_done, perform_tree_ssa_dce): Adjust.
12051 * cfganal.c (set_control_dependence_map_bit,
12052 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
12053 find_all_control_dependences): Move from tree-ssa-dce.c and
12054 implement as methods of control_dependences class.
12055 (control_dependences::control_dependences): New.
12056 (control_dependences::~control_dependences): Likewise.
12057 (control_dependences::get_edges_dependent_on): Likewise.
12058 (control_dependences::get_edge): Likewise.
12060 2013-09-04 Jan Hubicka <jh@suse.cz>
12062 * tree.c (types_same_for_odr): Drop overactive check.
12063 * ipa-devirt.c (hash_type_name): Likewise.
12065 2013-09-04 Jan Hubicka <jh@suse.cz>
12067 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
12068 (analyze_functions): ... here.
12070 2013-09-04 Jan Hubicka <jh@suse.cz>
12072 PR middle-end/58201
12073 * cgraphunit.c (analyze_functions): Clear AUX fields
12074 after processing; initialize assembler name has.
12076 2013-09-05 Jeff Law <law@redhat.com>
12078 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
12079 from thread_around_empty_block. Record threading path into PATH.
12080 Recurse if threading through the initial block is successful.
12081 (thread_across_edge): Corresponding changes to slightly simplify.
12083 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
12085 * config/aarch64/aarch64.md
12086 (type): Remove frecpe, frecps, frecpx.
12087 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
12088 fix to be a TARGET_SIMD instruction.
12089 (aarch64_frecps): Remove.
12090 * config/aarch64/aarch64-simd.md
12091 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
12092 (aarch64_frecps<mode>): Handle all float/vector of float modes.
12094 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
12095 Sofiane Naci <sofiane.naci@arm.com>
12097 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
12098 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
12099 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
12100 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
12101 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
12102 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
12103 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
12104 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
12105 into "clz, "rbit". Rename "shift" to "shift_imm".
12106 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
12107 changes. Update for attribute changes all occurrences of arlo_* and
12109 * config/arm/arm-fixed.md: Update for attribute changes
12110 all occurrences of arlo_* types.
12111 * config/arm/thumb2.md: Update for attribute changes all occurrences
12113 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
12114 (cortexa7_older_only): Likewise.
12115 (cortexa7_younger): Likewise.
12116 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12117 (1020alu_shift_op): Likewise.
12118 (1020alu_shift_reg_op): Likewise.
12119 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12120 (alu_shift_op): Likewise.
12121 (alu_shift_reg_op): Likewise.
12122 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12123 (11_alu_shift_op): Likewise.
12124 (11_alu_shift_reg_op): Likewise.
12125 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12126 (9_alu_shift_reg_op): Likewise.
12127 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
12129 (cortex_a15_alu_shift): Likewise.
12130 (cortex_a15_alu_shift_reg): Likewise.
12131 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
12133 (cortex_a5_alu_shift): Likewise.
12134 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
12136 (cortex_a53_alu_shift): Likewise.
12137 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
12139 (cortex_a7_alu_reg): Likewise.
12140 (cortex_a7_alu_shift): Likewise.
12141 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
12143 (cortex_a8_alu_shift): Likewise.
12144 (cortex_a8_alu_shift_reg): Likewise.
12145 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12146 (cortex_a9_dp_shift): Likewise.
12147 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
12149 * config/arm/cortex-r4.md
12150 (cortex_r4_alu): Update for attribute changes.
12151 (cortex_r4_mov): Likewise.
12152 (cortex_r4_alu_shift_reg): Likewise.
12153 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12154 (526_alu_shift_op): Likewise.
12155 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12156 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12157 (626te_alu_shift_op): Likewise.
12158 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
12159 (726te_alu_shift_op): Likewise.
12160 (726te_alu_shift_reg_op): Likewise.
12161 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12162 (mp626_alu_shift_op): Likewise.
12163 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
12164 (pj4_alu_conds): Likewise.
12165 (pj4_shift): Likewise.
12166 (pj4_shift_conds): Likewise.
12167 (pj4_alu_shift): Likewise.
12168 (pj4_alu_shift_conds): Likewise.
12169 * config/aarch64/aarch64.md: Update for attribute change
12170 all occurrences of arlo_* and shift* types.
12172 2013-09-05 Mike Stump <mikestump@comcast.net>
12174 * tree.h: Move documentation for tree_function_decl to tree-core.h
12175 with the declaration.
12177 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
12180 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
12181 looking for widest mode.
12183 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
12185 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
12187 2013-09-05 Richard Biener <rguenther@suse.de>
12189 PR tree-optimization/58137
12190 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
12191 Do not create vectors of pointers.
12192 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
12193 types for the components of the vector initializer.
12194 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
12195 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
12197 2013-09-05 Martin Jambor <mjambor@suse.cz>
12199 * ipa-prop.c (remove_described_reference): Accept missing references,
12200 return false if that hppens, otherwise return true.
12201 (cgraph_node_for_jfunc): New function.
12202 (try_decrement_rdesc_refcount): Likewise.
12203 (try_make_edge_direct_simple_call): Use them.
12204 (ipa_edge_removal_hook): Remove references from rdescs.
12205 (ipa_edge_duplication_hook): Clone rdescs and their references
12206 when the new edge has the same caller as the old one.
12207 * cgraph.c (cgraph_resolve_speculation): Remove speculative
12208 reference before removing any edges.
12210 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
12212 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
12214 * thumb2.md (thumb2_storewb_parisi): New pattern.
12216 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
12218 * config/aarch64/aarch64-option-extensions.def: Add
12219 AARCH64_OPT_EXTENSION of 'crc'.
12220 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
12221 (AARCH64_ISA_CRC): Ditto.
12222 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
12223 description of the CRC extension.
12225 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
12227 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
12228 * config/rs6000/linux.h: Ditto.
12229 * alpha/linux.h: Ditto.
12230 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
12231 no_c99_libc_has_function.
12232 * config/c6x/uclinux-elf.h: Ditto.
12233 * config/lm32/uclinux-elf.h: Ditto.
12234 * config/m68k/uclinux.h: Ditto.
12235 * config/moxie/uclinux.h: Ditto.
12236 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
12237 (crisv32-*-linux*, cris-*-linux*): Ditto.
12238 * config/bfin/bfin.c: Include "tm_p.h".
12240 2013-09-05 Richard Biener <rguenther@suse.de>
12242 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
12243 check for a definition without a basic-block.
12245 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
12246 Sofiane Naci <sofiane.naci@arm.com>
12248 * config/aarch64/aarch64.md
12249 (*movti_aarch64): Rename r_2_f and f_2_r.
12250 (*movsf_aarch64): Likewise.
12251 (*movdf_aarch64): Likewise.
12252 (*movtf_aarch64): Likewise.
12253 (aarch64_movdi_<mode>low): Likewise.
12254 (aarch64_movdi_<mode>high): Likewise.
12255 (aarch64_mov<mode>high_di): Likewise.
12256 (aarch64_mov<mode>low_di): Likewise.
12257 (aarch64_movtilow_tilow): Likewise.
12258 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
12259 values to config/arm/types.md
12260 (attribute "conds"): Update for attribute change.
12261 (anddi3_insn): Likewise.
12262 (iordi3_insn): Likewise.
12263 (xordi3_insn): Likewise.
12264 (one_cmpldi2): Likewise.
12265 * config/arm/types.md (type): Add Neon types.
12266 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
12267 use "type" attribute.
12268 (movmisalign<mode>_neon_store): Likewise.
12269 (movmisalign<mode>_neon_load): Likewise.
12270 (vec_set<mode>_internal): Likewise.
12271 (vec_setv2di_internal): Likewise.
12272 (vec_extract<mode>): Likewise.
12273 (vec_extractv2di): Likewise.
12274 (add<mode>3_neon): Likewise.
12275 (adddi3_neon): Likewise.
12276 (sub<mode>3_neon): Likewise.
12277 (subdi3_neon): Likewise.
12278 (mul<mode>3_neon): Likewise.
12279 (mul<mode>3add<mode>_neon): Likewise.
12280 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
12281 (fma<VCVTF:mode>4)): Likewise.
12282 (fma<VCVTF:mode>4_intrinsic): Likewise.
12283 (fmsub<VCVTF:mode>4)): Likewise.
12284 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
12285 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
12286 (ior<mode>3): Likewise.
12287 (and<mode>3): Likewise.
12288 (anddi3_neon): Likewise.
12289 (orn<mode>3_neon): Likewise.
12290 (orndi3_neon): Likewise.
12291 (bic<mode>3_neon): Likewise.
12292 (bicdi3_neon): Likewise.
12293 (xor<mode>3): Likewise.
12294 (one_cmpl<mode>2): Likewise.
12295 (abs<mode>2): Likewise.
12296 (neg<mode>2): Likewise.
12297 (umin<mode>3_neon): Likewise.
12298 (umax<mode>3_neon): Likewise.
12299 (smin<mode>3_neon): Likewise.
12300 (smax<mode>3_neon): Likewise.
12301 (vashl<mode>3): Likewise.
12302 (vashr<mode>3_imm): Likewise.
12303 (vlshr<mode>3_imm): Likewise.
12304 (ashl<mode>3_signed): Likewise.
12305 (ashl<mode>3_unsigned): Likewise.
12306 (neon_load_count): Likewise.
12307 (ashldi3_neon_noclobber): Likewise.
12308 (signed_shift_di3_neon): Likewise.
12309 (unsigned_shift_di3_neon): Likewise.
12310 (ashrdi3_neon_imm_noclobber): Likewise.
12311 (lshrdi3_neon_imm_noclobber): Likewise.
12312 (widen_ssum<mode>3): Likewise.
12313 (widen_usum<mode>3): Likewise.
12314 (quad_halves_<code>v4si): Likewise.
12315 (quad_halves_<code>v4sf): Likewise.
12316 (quad_halves_<code>v8hi): Likewise.
12317 (quad_halves_<code>v16qi): Likewise.
12318 (reduc_splus_v2di): Likewise.
12319 (neon_vpadd_internal<mode>): Likewise.
12320 (neon_vpsmin<mode>): Likewise.
12321 (neon_vpsmax<mode>): Likewise.
12322 (neon_vpumin<mode>): Likewise.
12323 (neon_vpumax<mode>): Likewise.
12324 (ss_add<mode>_neon): Likewise.
12325 (us_add<mode>_neon): Likewise.
12326 (ss_sub<mode>_neon): Likewise.
12327 (us_sub<mode>_neon): Likewise.
12328 (neon_vadd<mode>_unspec): Likewise.
12329 (neon_vaddl<mode>): Likewise.
12330 (neon_vaddw<mode>): Likewise.
12331 (neon_vhadd<mode>): Likewise.
12332 (neon_vqadd<mode>): Likewise.
12333 (neon_vaddhn<mode>): Likewise.
12334 (neon_vmul<mode>): Likewise.
12335 (neon_vmla<mode>): Likewise.
12336 (neon_vmlal<mode>): Likewise.
12337 (neon_vmls<mode>): Likewise.
12338 (neon_vmlsl<mode>): Likewise.
12339 (neon_vqdmulh<mode>): Likewise.
12340 (neon_vqdmlal<mode>): Likewise.
12341 (neon_vqdmlsl<mode>): Likewise.
12342 (neon_vmull<mode>): Likewise.
12343 (neon_vqdmull<mode>): Likewise.
12344 (neon_vsub<mode>_unspec): Likewise.
12345 (neon_vsubl<mode>): Likewise.
12346 (neon_vsubw<mode>): Likewise.
12347 (neon_vqsub<mode>): Likewise.
12348 (neon_vhsub<mode>): Likewise.
12349 (neon_vsubhn<mode>): Likewise.
12350 (neon_vceq<mode>): Likewise.
12351 (neon_vcge<mode>): Likewise.
12352 (neon_vcgeu<mode>): Likewise.
12353 (neon_vcgt<mode>): Likewise.
12354 (neon_vcgtu<mode>): Likewise.
12355 (neon_vcle<mode>): Likewise.
12356 (neon_vclt<mode>): Likewise.
12357 (neon_vcage<mode>): Likewise.
12358 (neon_vcagt<mode>): Likewise.
12359 (neon_vtst<mode>): Likewise.
12360 (neon_vabd<mode>): Likewise.
12361 (neon_vabdl<mode>): Likewise.
12362 (neon_vaba<mode>): Likewise.
12363 (neon_vabal<mode>): Likewise.
12364 (neon_vmax<mode>): Likewise.
12365 (neon_vmin<mode>): Likewise.
12366 (neon_vpaddl<mode>): Likewise.
12367 (neon_vpadal<mode>): Likewise.
12368 (neon_vpmax<mode>): Likewise.
12369 (neon_vpmin<mode>): Likewise.
12370 (neon_vrecps<mode>): Likewise.
12371 (neon_vrsqrts<mode>): Likewise.
12372 (neon_vqabs<mode>): Likewise.
12373 (neon_vqneg<mode>): Likewise.
12374 (neon_vcls<mode>): Likewise.
12375 (clz<mode>2): Likewise.
12376 (popcount<mode>2): Likewise.
12377 (neon_vrecpe): Likewise.
12378 (neon_vrsqrte): Likewise.
12379 (neon_vget_lane<mode>_sext_internal): Likewise.
12380 (neon_vget_lane<mode>_zext_internal): Likewise.
12381 (neon_vdup_n<mode>): Likewise.
12382 (neon_vdup_nv2di): Likewise.
12383 (neon_vdpu_lane<mode>_internal): Likewise.
12384 (neon_vswp<mode>): Likewise.
12385 (float<mode><V_cvtto>2): Likewise.
12386 (floatuns<mode><V_cvtto>2): Likewise.
12387 (fix_trunc<mode><V_cvtto>)2): Likewise
12388 (fixuns_trunc<mode><V_cvtto)2): Likewise.
12389 (neon_vcvt<mode>): Likewise.
12390 (neon_vcvtv4sfv4hf): Likewise.
12391 (neon_vcvtv4hfv4sf): Likewise.
12392 (neon_vcvt_n<mode>): Likewise.
12393 (neon_vmovn<mode>): Likewise.
12394 (neon_vqmovn<mode>): Likewise.
12395 (neon_vqmovun<mode>): Likewise.
12396 (neon_vmovl<mode>): Likewise.
12397 (neon_vmul_lane<mode>): Likewise.
12398 (neon_vmull_lane<mode>): Likewise.
12399 (neon_vqdmull_lane<mode>): Likewise.
12400 (neon_vqdmulh_lane<mode>): Likewise.
12401 (neon_vmla_lane<mode>): Likewise.
12402 (neon_vmlal_lane<mode>): Likewise.
12403 (neon_vqdmlal_lane<mode>): Likewise.
12404 (neon_vmls_lane<mode>): Likewise.
12405 (neon_vmlsl_lane<mode>): Likewise.
12406 (neon_vqdmlsl_lane<mode>): Likewise.
12407 (neon_vext<mode>): Likewise.
12408 (neon_vrev64<mode>): Likewise.
12409 (neon_vrev32<mode>): Likewise.
12410 (neon_vrev16<mode>): Likewise.
12411 (neon_vbsl<mode>_internal): Likewise.
12412 (neon_vshl<mode>): Likewise.
12413 (neon_vqshl<mode>): Likewise.
12414 (neon_vshr_n<mode>): Likewise.
12415 (neon_vshrn_n<mode>): Likewise.
12416 (neon_vqshrn_n<mode>): Likewise.
12417 (neon_vqshrun_n<mode>): Likewise.
12418 (neon_vshl_n<mode>): Likewise.
12419 (neon_vqshl_n<mode>): Likewise.
12420 (neon_vqshlu_n<mode>): Likewise.
12421 (neon_vshll_n<mode>): Likewise.
12422 (neon_vsra_n<mode>): Likewise.
12423 (neon_vsri_n<mode>): Likewise.
12424 (neon_vsli_n<mode>): Likewise.
12425 (neon_vtbl1v8qi): Likewise.
12426 (neon_vtbl2v8qi): Likewise.
12427 (neon_vtbl3v8qi): Likewise.
12428 (neon_vtbl4v8qi): Likewise.
12429 (neon_vtbx1v8qi): Likewise.
12430 (neon_vtbx2v8qi): Likewise.
12431 (neon_vtbx3v8qi): Likewise.
12432 (neon_vtbx4v8qi): Likewise.
12433 (neon_vtrn<mode>_internal): Likewise.
12434 (neon_vzip<mode>_internal): Likewise.
12435 (neon_vuzp<mode>_internal): Likewise.
12436 (neon_vld1<mode>): Likewise.
12437 (neon_vld1_lane<mode>): Likewise.
12438 (neon_vld1_dup<mode>): Likewise.
12439 (neon_vld1_dupv2di): Likewise.
12440 (neon_vst1<mode>): Likewise.
12441 (neon_vst1_lane<mode>): Likewise.
12442 (neon_vld2<mode>): Likewise.
12443 (neon_vld2_lane<mode>): Likewise.
12444 (neon_vld2_dup<mode>): Likewise.
12445 (neon_vst2<mode>): Likewise.
12446 (neon_vst2_lane<mode>): Likewise.
12447 (neon_vld3<mode>): Likewise.
12448 (neon_vld3qa<mode>): Likewise.
12449 (neon_vld3qb<mode>): Likewise.
12450 (neon_vld3_lane<mode>): Likewise.
12451 (neon_vld3_dup<mode>): Likewise.
12452 (neon_vst3<mode>): Likewise.
12453 (neon_vst3qa<mode>): Likewise.
12454 (neon_vst3qb<mode>): Likewise.
12455 (neon_vst3_lane<mode>): Likewise.
12456 (neon_vld4<mode>): Likewise.
12457 (neon_vld4qa<mode>): Likewise.
12458 (neon_vld4qb<mode>): Likewise.
12459 (neon_vld4_lane<mode>): Likewise.
12460 (neon_vld4_dup<mode>): Likewise.
12461 (neon_vst4<mode>): Likewise.
12462 (neon_vst4qa<mode>): Likewise.
12463 (neon_vst4qb<mode>): Likewise.
12464 (neon_vst4_lane<mode>): Likewise.
12465 (neon_vec_unpack<US>_lo_<mode>): Likewise.
12466 (neon_vec_unpack<US>_hi_<mode>): Likewise.
12467 (neon_vec_<US>mult_lo_<mode>): Likewise.
12468 (neon_vec_<US>mult_hi_<mode>): Likewise.
12469 (neon_vec_<US>shiftl_<mode>): Likewise.
12470 (neon_unpack<US>_<mode>): Likewise.
12471 (neon_vec_<US>mult_<mode>): Likewise.
12472 (vec_pack_trunc_<mode>): Likewise.
12473 (neon_vec_pack_trunk_<mode>): Likewise.
12474 (neon_vabd<mode>_2): Likewise.
12475 (neon_vabd<mode>_3): Likewise.
12476 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12477 (thumb2_movsi_vfp): Likewise.
12478 (movdi_vfp): Likewise.
12479 (movdi_vfp_cortexa8): Likewise.
12480 (movhf_vfp_neon): Likewise.
12481 (movhf_vfp): Likewiwse.
12482 (movsf_vfp): Likewiwse.
12483 (thumb2_movsf_vfp): Likewiwse.
12484 (movdf_vfp): Likewise.
12485 (thumb2_movdf_vfp): Likewise.
12486 (movsfcc_vfp): Likewise.
12487 (thumb2_movsfcc_vfp): Likewise.
12488 (movdfcc_vfp): Likewise.
12489 (thumb2_movdfcc_vfp): Likewise.
12490 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
12491 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
12492 (v10_v2c): Likewise.
12493 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
12495 (cortex_a15_neon_int_2): Likewise.
12496 (cortex_a15_neon_int_3): Likewise.
12497 (cortex_a15_neon_int_4): Likewise.
12498 (cortex_a15_neon_int_5): Likewise.
12499 (cortex_a15_neon_vqneg_vqabs): Likewise.
12500 (cortex_a15_neon_vmov): Likewise.
12501 (cortex_a15_neon_vaba): Likewise.
12502 (cortex_a15_neon_vaba_qqq): Likewise.
12503 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12504 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12505 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
12506 scalar_64_32_long_scalar): Likewise.
12507 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12508 (cortex_a15_neon_mla_qqq_8_16): Likewise.
12509 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12510 lotype_qdd_64_32_long): Likewise.
12511 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12512 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12513 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
12514 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12515 (cortex_a15_neon_shift_1): Likewise.
12516 (cortex_a15_neon_shift_2): Likewise.
12517 (cortex_a15_neon_shift_3): Likewise.
12518 (cortex_a15_neon_vshl_ddd): Likewise.
12519 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12520 (cortex_a15_neon_vsra_vrsra): Likewise.
12521 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
12522 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
12523 (cortex_a15_neon_fp_vmul_ddd): Likewise.
12524 (cortex_a15_neon_fp_vmul_qqd): Likewise.
12525 (cortex_a15_neon_fp_vmla_ddd): Likewise.
12526 (cortex_a15_neon_fp_vmla_qqq): Likewise.
12527 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
12528 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
12529 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12530 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12531 (cortex_a15_neon_bp_simple): Likewise.
12532 (cortex_a15_neon_bp_2cycle): Likewise.
12533 (cortex_a15_neon_bp_3cycle): Likewise.
12534 (cortex_a15_neon_vld1_1_2_regs): Likewise.
12535 (cortex_a15_neon_vld1_3_4_regs): Likewise.
12536 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12537 (cortex_a15_neon_vld2_4_regs): Likewise.
12538 (cortex_a15_neon_vld3_vld4): Likewise.
12539 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12540 (cortex_a15_neon_vst1_3_4_regs): Likewise.
12541 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
12542 (cortex_a15_neon_vst3_vst4): Likewise.
12543 (cortex_a15_neon_vld1_vld2_lane): Likewise.
12544 (cortex_a15_neon_vld3_vld4_lane" 10
12545 (cortex_a15_neon_vst1_vst2_lane): Likewise.
12546 (cortex_a15_neon_vst3_vst4_lane): Likewise.
12547 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
12548 (cortex_a15_neon_ldm_2): Likewise.0
12549 (cortex_a15_neon_stm_2): Likewise.
12550 (cortex_a15_neon_mcr): Likewise.
12551 (cortex_a15_neon_mcr_2_mcrr): Likewise.
12552 (cortex_a15_neon_mrc): Likewise.
12553 (cortex_a15_neon_mrrc): Likewise.
12554 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12556 (cortex_a15_alu_shift): Likewise.
12557 (cortex_a15_alu_shift_reg): Likewise.
12558 (cortex_a15_mult32): Likewise.
12559 (cortex_a15_mult64): Likewise.
12560 (cortex_a15_block): Likewise.
12561 (cortex_a15_branch): Likewise.
12562 (cortex_a15_load1): Likewise.
12563 (cortex_a15_load3): Likewise.
12564 (cortex_a15_store1): Likewise.
12565 (cortex_a15_store3): Likewise.
12566 (cortex_a15_call): Likewise.
12567 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
12568 (cortex_a5_f2r): Likewise.
12569 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
12571 (cortex_a53_f2r): Likewise.
12572 * config/arm/cortex-a7.md
12573 (cortex_a7_branch): Update for attribute change.
12574 (cortex_a7_call): Likewise.
12575 (cortex_a7_alu_imm): Likewise.
12576 (cortex_a7_alu_reg): Likewise.
12577 (cortex_a7_alu_shift): Likewise.
12578 (cortex_a7_mul): Likewise.
12579 (cortex_a7_load1): Likewise.
12580 (cortex_a7_store1): Likewise.
12581 (cortex_a7_load2): Likewise.
12582 (cortex_a7_store2): Likewise.
12583 (cortex_a7_load3): Likewise.
12584 (cortex_a7_store3): Likewise.
12585 (cortex_a7_load4): Likewise.
12586 (cortex_a7_store4): Likewise.
12587 (cortex_a7_fpalu): Likewise.
12588 (cortex_a7_fconst): Likewise.
12589 (cortex_a7_fpmuls): Likewise.
12590 (cortex_a7_neon_mul): Likewise.
12591 (cortex_a7_fpmacs): Likewise.
12592 (cortex_a7_neon_mla: Likewise.
12593 (cortex_a7_fpmuld: Likewise.
12594 (cortex_a7_fpmacd: Likewise.
12595 (cortex_a7_fpfmad: Likewise.
12596 (cortex_a7_fdivs: Likewise.
12597 (cortex_a7_fdivd: Likewise.
12598 (cortex_a7_r2f: Likewise.
12599 (cortex_a7_f2r: Likewise.
12600 (cortex_a7_f_flags: Likewise.
12601 (cortex_a7_f_loads: Likewise.
12602 (cortex_a7_f_loadd: Likewise.
12603 (cortex_a7_f_stores: Likewise.
12604 (cortex_a7_f_stored: Likewise.
12605 (cortex_a7_neon): Likewise.
12606 * config/arm/cortex-a8-neon.md
12607 (cortex_a8_neon_mrc): Update for attribute change.
12608 (cortex_a8_neon_mrrc): Likewise.
12609 (cortex_a8_neon_int_1): Likewise.
12610 (cortex_a8_neon_int_2): Likewise.
12611 (cortex_a8_neon_int_3): Likewise.
12612 (cortex_a8_neon_int_4): Likewise.
12613 (cortex_a8_neon_int_5): Likewise.
12614 (cortex_a8_neon_vqneg_vqabs): Likewise.
12615 (cortex_a8_neon_vmov): Likewise.
12616 (cortex_a8_neon_vaba): Likewise.
12617 (cortex_a8_neon_vaba_qqq): Likewise.
12618 (cortex_a8_neon_vsma): Likewise.
12619 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12620 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12621 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
12622 long_scalar): Likewise.
12623 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12624 (cortex_a8_neon_mla_qqq_8_16): Likewise.
12625 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12626 long_scalar_qdd_64_32_long): Likewise.
12627 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12628 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12629 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
12630 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12631 (cortex_a8_neon_shift_1): Likewise.
12632 (cortex_a8_neon_shift_2): Likewise.
12633 (cortex_a8_neon_shift_3): Likewise.
12634 (cortex_a8_neon_vshl_ddd): Likewise.
12635 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12636 (cortex_a8_neon_vsra_vrsra): Likewise.
12637 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
12638 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
12639 (cortex_a8_neon_fp_vsum): Likewise.
12640 (cortex_a8_neon_fp_vmul_ddd): Likewise.
12641 (cortex_a8_neon_fp_vmul_qqd): Likewise.
12642 (cortex_a8_neon_fp_vmla_ddd): Likewise.
12643 (cortex_a8_neon_fp_vmla_qqq): Likewise.
12644 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
12645 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
12646 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12647 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12648 (cortex_a8_neon_bp_simple): Likewise.
12649 (cortex_a8_neon_bp_2cycle): Likewise.
12650 (cortex_a8_neon_bp_3cycle): Likewise.
12651 (cortex_a8_neon_ldr): Likewise.
12652 (cortex_a8_neon_str): Likewise.
12653 (cortex_a8_neon_vld1_1_2_regs): Likewise.
12654 (cortex_a8_neon_vld1_3_4_regs): Likewise.
12655 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12656 (cortex_a8_neon_vld2_4_regs): Likewise.
12657 (cortex_a8_neon_vld3_vld4): Likewise.
12658 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12659 (cortex_a8_neon_vst1_3_4_regs): Likewise.
12660 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
12661 (cortex_a8_neon_vst3_vst4): Likewise.
12662 (cortex_a8_neon_vld1_vld2_lane): Likewise.
12663 (cortex_a8_neon_vld3_vld4_lane): Likewise.
12664 (cortex_a8_neon_vst1_vst2_lane): Likewise.
12665 (cortex_a8_neon_vst3_vst4_lane): Likewise.
12666 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
12667 (cortex_a8_neon_mcr): Likewise.
12668 (cortex_a8_neon_mcr_2_mcrr): Likewise.
12669 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12670 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
12672 (ca9_neon_mrrc): Likewise.
12673 (cortex_a9_neon_int_1): Likewise.
12674 (cortex_a9_neon_int_2): Likewise.
12675 (cortex_a9_neon_int_3): Likewise.
12676 (cortex_a9_neon_int_4): Likewise.
12677 (cortex_a9_neon_int_5): Likewise.
12678 (cortex_a9_neon_vqneg_vqabs): Likewise.
12679 (cortex_a9_neon_vmov): Likewise.
12680 (cortex_a9_neon_vaba): Likewise.
12681 (cortex_a9_neon_vaba_qqq): Likewise.
12682 (cortex_a9_neon_vsma): Likewise.
12683 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12684 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12685 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
12686 long_scalar): Likewise.
12687 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12688 (cortex_a9_neon_mla_qqq_8_16): Likewise.
12689 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12690 long_scalar_qdd_64_32_long): Likewise.
12691 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12692 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12693 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
12694 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12695 (cortex_a9_neon_shift_1): Likewise.
12696 (cortex_a9_neon_shift_2): Likewise.
12697 (cortex_a9_neon_shift_3): Likewise.
12698 (cortex_a9_neon_vshl_ddd): Likewise.
12699 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12700 (cortex_a9_neon_vsra_vrsra): Likewise.
12701 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
12702 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
12703 (cortex_a9_neon_fp_vsum): Likewise.
12704 (cortex_a9_neon_fp_vmul_ddd): Likewise.
12705 (cortex_a9_neon_fp_vmul_qqd): Likewise.
12706 (cortex_a9_neon_fp_vmla_ddd): Likewise.
12707 (cortex_a9_neon_fp_vmla_qqq): Likewise.
12708 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
12709 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
12710 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12711 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12712 (cortex_a9_neon_bp_simple): Likewise.
12713 (cortex_a9_neon_bp_2cycle): Likewise.
12714 (cortex_a9_neon_bp_3cycle): Likewise.
12715 (cortex_a9_neon_ldr): Likewise.
12716 (cortex_a9_neon_str): Likewise.
12717 (cortex_a9_neon_vld1_1_2_regs): Likewise.
12718 (cortex_a9_neon_vld1_3_4_regs): Likewise.
12719 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12720 (cortex_a9_neon_vld2_4_regs): Likewise.
12721 (cortex_a9_neon_vld3_vld4): Likewise.
12722 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12723 (cortex_a9_neon_vst1_3_4_regs): Likewise.
12724 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
12725 (cortex_a9_neon_vst3_vst4): Likewise.
12726 (cortex_a9_neon_vld1_vld2_lane): Likewise.
12727 (cortex_a9_neon_vld3_vld4_lane): Likewise.
12728 (cortex_a9_neon_vst1_vst2_lane): Likewise.
12729 (cortex_a9_neon_vst3_vst4_lane): Likewise.
12730 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
12731 (cortex_a9_neon_mcr): Likewise.
12732 (cortex_a9_neon_mcr_2_mcrr): Likewise.
12733 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12734 (cortex_a9_fps): Likewise.
12735 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
12737 (cortex_m4_fmuls): Likewise.
12738 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
12740 (cortex_r4_mrc): Likewise.
12741 * config/arm/iterators.md: Update comment referring to neon_type.
12742 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
12743 (iwmmxt_movsi_insn): Likewise.
12744 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
12746 (pj4_core_to_vfp): Likewise.
12747 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
12749 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
12750 (vfp_fstore): Likewise.
12751 * doc/md.texi: Change references to neon_type to refer to type.
12753 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12755 * tree.h (DECL_BUILT_IN): Fix typo in comment.
12757 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
12759 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
12760 lglobl if not weak.
12762 2013-09-04 Easwaran Raman <eraman@google.com>
12764 PR middle-end/57370
12765 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
12766 (build_and_add_sum): Use it.
12767 (appears_later_in_bb): Simplify code.
12769 2013-09-04 Teresa Johnson <tejohnson@google.com>
12771 * dumpfile.c (dump_finish): Don't close stderr/stdout.
12773 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
12775 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
12777 2013-09-04 Jan Hubicka <jh@suse.cz>
12779 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
12780 * ipa-devirt.c: Include diganostic.h
12781 (odr_type_d): Add types and types_set.
12782 (hash_type_name): Work for types with vtables during LTO.
12783 (odr_hasher::remove): Fix comment; destroy types_set.
12784 (add_type_duplicate): New function,
12785 (get_odr_type): Use it.
12786 (dump_type_inheritance_graph): Dump type duplicates.
12787 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
12789 * tree.c (types_same_for_odr): Give exact answers on types with
12792 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12794 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
12795 explaining their differences.
12797 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
12799 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
12801 2013-09-03 Jeff Law <law@redhat.com>
12803 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
12804 when not threading through a joiner block. Pass joiner/no joiner
12805 state to register_jump_thread.
12806 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
12807 state from argument rather than implying on path length.
12808 Dump the entire jump thread path into debugging dump.
12809 * tree-flow.h (register_jump_thread): Update prototype.
12811 2013-08-29 Xinliang David Li <davidxl@google.com>
12813 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12814 Remove a trivial gcc_assert.
12816 2013-08-29 Xinliang David Li <davidxl@google.com>
12818 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
12819 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
12820 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12821 Delay base decl alignment adjustment.
12822 * tree-vectorizer.c (vect_destroy_datarefs): New function.
12823 * tree-vectorizer.h: New data structure.
12824 (set_dr_misalignment): New function.
12825 (dr_misalignment): Ditto.
12826 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
12827 (vectorizable_load): Ditto.
12828 (ensure_base_align): New function.
12829 (vectorize_loops): Add dbg_cnt support.
12830 (execute_vect_slp): Ditto.
12831 * dbgcnt.def: New debug counter.
12832 * Makefile: New dependency.
12834 2013-09-03 Meador Inge <meadori@codesourcery.com>
12838 2013-08-30 Meador Inge <meadori@codesourcery.com>
12840 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12842 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
12844 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
12845 function descriptor.
12847 2013-09-03 Richard Biener <rguenther@suse.de>
12849 * tree-affine.c (add_elt_to_tree): Fix association issue,
12850 avoid useless converts and make sure to always return a
12851 properly typed result.
12853 2013-09-03 Richard Biener <rguenther@suse.de>
12855 PR middle-end/57656
12856 * fold-const.c (negate_expr_p): Fix division case.
12857 (negate_expr): Likewise.
12859 2013-09-03 Richard Biener <rguenther@suse.de>
12862 * tree-streamer-out.c: Include tm.h.
12863 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
12865 2013-09-03 Jan Hubicka <jh@suse.cz>
12867 * tree-profile.c (tree_profiling): Cleanup CFG when done.
12869 2013-09-03 Alan Modra <amodra@gmail.com>
12871 * config.gcc (powerpc*-*-linux*): Add support for little-endian
12872 multilibs to big-endian target and vice versa.
12873 * config/rs6000/t-linux64: Use := assignment on all vars.
12874 (MULTILIB_EXTRA_OPTS): Remove fPIC.
12875 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
12876 * config/rs6000/t-linux64le: New file.
12877 * config/rs6000/t-linux64bele: New file.
12878 * config/rs6000/t-linux64lebe: New file.
12880 2013-09-02 Jan Hubicka <jh@suse.cz>
12882 * ipa-inline-transform.c (inline_transform): Do not
12883 optimize_inline_calls when not optimizing.
12885 2013-09-02 Jan Hubicka <jh@suse.cz>
12887 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
12888 duplicated nodes for assembler names.
12889 * symtab.c (symtab_unregister_node): Do not attempt to unlink
12890 hard registers from assembler name hash.
12892 2013-09-02 Jan Hubicka <jh@suse.cz>
12894 * ipa-split.c (execute_split_functions): Split externally visible
12895 functions called once.
12897 2013-09-02 Martin Jambor <mjambor@suse.cz>
12900 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
12901 linked list. When finding the correct duplicate, also consider also
12902 the caller in additon to its inlined_to node.
12904 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
12906 * config/aarch64/aarch64-simd-builtins.def
12907 (dup_lane_scalar): Remove.
12908 * config/aarch64/aarch64-simd.md
12909 (aarch64_simd_dup): Add 'w->w' alternative.
12910 (aarch64_dup_lane<mode>): Allow for VALL.
12911 (aarch64_dup_lane_scalar<mode>): Remove.
12912 (aarch64_dup_lane_<vswap_width_name><mode>): New.
12913 (aarch64_get_lane_signed<mode>): Add w->w altenative.
12914 (aarch64_get_lane_unsigned<mode>): Likewise.
12915 (aarch64_get_lane<mode>): Likewise.
12916 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
12917 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
12918 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
12919 (VCON): Change container of V2SF.
12920 (vswap_width_name): Likewise.
12921 * config/aarch64/arm_neon.h
12922 (__aarch64_vdup_lane_any): New.
12923 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12924 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
12925 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12927 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
12929 PR middle-end/56382
12930 * expr.c (emit_move_complex): Do not move complex FP values as parts if
12931 the source or the destination is a single hard register.
12933 2013-09-02 Richard Biener <rguenther@suse.de>
12935 PR middle-end/57511
12936 * tree-scalar-evolution.c (instantiate_scev_name): Allow
12939 2013-09-02 Richard Biener <rguenther@suse.de>
12941 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
12942 arithmetic to sizetype.
12944 2013-09-02 Bin Cheng <bin.cheng@arm.com>
12946 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
12947 Find auto-increment use both before and after candidate.
12949 2013-09-02 Marek Polacek <polacek@redhat.com>
12951 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
12953 2013-09-01 Jan Hubicka <jh@suse.cz>
12955 * Makefile.in: Add ipa-profile.o
12956 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
12957 * cgraph.c (struct cgraph_propagate_frequency_data,
12958 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
12959 ipa-profile.c; replace cgraph_ by ipa_ prefix.
12960 * cgraph.h (cgraph_propagate_frequency): Remove.
12961 * ipa-inline-analysis.c: Include ipa-utils.h;
12962 drop duplicated cfgloop.h.
12963 (inline_update_callee_summaries): Update.
12964 * ipa-profile.c: New file.
12965 * ipa-utils.h (ipa_propagate_frequency): Declare.
12966 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
12967 data-streamer.h, value-prof.h.
12968 (symtab_remove_unreachable_nodes): Update profile.
12969 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
12970 account_time_size, cmp_counts, dump_histogram,
12971 ipa_profile_generate_summary, ipa_profile_write_summary,
12972 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
12973 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
12974 Move to ipa-profile.c.
12976 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
12978 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
12980 2013-09-01 Jan Hubicka <jh@suse.cz>
12982 * common.opt (fdevirtualize-speculatively): New function.
12983 * invoke.texi (fdevirtualize-speculatively): Document.
12984 * ipa-devirt.c: Include ipa-inline.h
12985 (likely_target_p): New function.
12986 (ipa_devirt): New function.
12987 (gate_ipa_devirt): New function.
12988 (pass_data_ipa_devirt): New static var.
12989 (pass_ipa_devirt): Likewise.
12990 (make_pass_ipa_devirt): New function.
12991 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
12992 (common_handle_option): Disable devirtualization when
12993 value range profiling is available.
12994 * passes.def (pass_ipa_devirt): Add.
12995 * timever.def (TV_IPA_DEVIRT): New timevar.
12996 * tree-pass.h (make_pass_ipa_devirt):
12998 2013-09-01 Iain Sandoe <iain@codesourcery.com>
13000 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
13001 include sanitize(undefined).
13003 2013-08-31 Diego Novillo <dnovillo@google.com>
13005 * Makefile.in (TREE_CORE_H): Define.
13007 (GTFILES): Add tree-core.h.
13008 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
13010 * tree-core.h: New file.
13011 Move all data structures, enum, typedefs, global
13012 declarations and constants from ...
13013 * tree.h: ... here.
13015 2013-08-31 Jan Hubicka <jh@suse.cz>
13017 * bulitins.c (expand_builtin): Do not early exit for gcov
13018 instrumented functions.
13020 2013-08-31 Marek Polacek <polacek@redhat.com>
13022 * ubsan.c: Include tm_p.h.
13024 2013-08-31 Jan Hubicka <jh@suse.cz>
13026 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
13029 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
13030 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
13031 * tree-cfg.c (verify_gimple_label): ... here.
13032 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
13033 (ipa_merge_profiles): New function.
13034 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
13035 (lto_input_function_body): Likewise.
13036 * ipa-utils.h (ipa_merge_profiles): Declare.
13037 * lto-streamer.h (lto_input_function_body): Update prototype.
13038 (emit_label_in_global_context_p): Remove.
13039 * lto-symtab.c: Include ipa-utils.h
13040 (lto_cgraph_replace_node): Use ipa_merge_profiles.
13042 2013-08-31 Jan Hubicka <jh@suse.cz>
13044 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
13046 2013-08-31 Jan Hubicka <jh@suse.cz>
13048 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
13050 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
13052 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
13053 "cmp" RTX before signed_comparison_operator check to account
13054 for "code" changes.
13056 2013-08-30 Jan Hubicka <jh@suse.cz>
13058 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
13059 (detect_type_change_1): Rename to ...
13060 (detect_type_change): ... this one; early return on non-polymorphic
13062 (detect_type_change_ssa): Add comp_type parameter; update
13063 use of detect_type_change.
13064 (compute_complex_assign_jump_func): Add param_type parameter;
13065 update use of detect_type_change_ssa.
13066 (compute_complex_ancestor_jump_func): Likewise.
13067 (ipa_get_callee_param_type): New function.
13068 (ipa_compute_jump_functions_for_edge): Compute parameter type;
13069 update calls to the jump function computation functions.
13071 2013-08-30 Teresa Johnson <tejohnson@google.com>
13072 Steven Bosscher <steven@gcc.gnu.org>
13074 * cfgrtl.c (fixup_new_cold_bb): New routine.
13075 (commit_edge_insertions): Invoke fixup_partitions.
13076 (find_partition_fixes): New routine.
13077 (fixup_partitions): Ditto.
13078 (verify_hot_cold_block_grouping): Update comments.
13079 (rtl_verify_edges): Invoke find_partition_fixes.
13080 (rtl_verify_bb_pointers): Update comments.
13081 (rtl_verify_bb_layout): Ditto.
13082 * basic-block.h (probably_never_executed_edge_p): Declare.
13083 (fixup_partitions): Ditto.
13084 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
13085 * bb-reorder.c (sanitize_hot_paths): New function.
13086 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
13087 sanitize_hot_paths.
13088 * predict.c (probably_never_executed_edge_p): New routine.
13089 * cfg.c (check_bb_profile): Add partition insanity warnings.
13091 2013-08-30 Meador Inge <meadori@codesourcery.com>
13093 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
13095 2013-08-30 Marek Polacek <polacek@redhat.com>
13097 * Makefile.in (ubsan.o): Add.
13098 (c-family/c-ubsan.o): Add.
13099 (builtins.o): Add ubsan.h dependency.
13100 * ubsan.h: New file.
13101 * ubsan.c: New file.
13102 * common.opt: Add -fsanitize=undefined option.
13103 (flag_sanitize): Add variable.
13104 (fsanitize=): Add option. Add Driver.
13105 (fsanitize=thread): Remove option.
13106 (fsanitize=address): Likewise.
13107 (static-libubsan): New option.
13108 * doc/invoke.texi: Document the new flag and -static-libubsan.
13109 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
13110 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
13111 * builtin-attrs.def (ATTR_COLD): Define.
13112 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
13113 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
13114 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
13115 * flag-types.h (sanitize_code): New enum.
13116 * opts.c (common_handle_option): Parse command line arguments
13117 of -fsanitize=. Add -fsanitize=unreachable option.
13118 * varasm.c (get_variable_section): Adjust.
13119 (assemble_noswitch_variable): Likewise.
13120 (assemble_variable): Likewise.
13121 (output_constant_def_contents): Likewise.
13122 (categorize_decl_for_section): Likewise.
13123 (place_block_symbol): Likewise.
13124 (output_object_block): Likewise.
13125 * builtins.def: Likewise.
13126 * toplev.c (compile_file): Likewise.
13127 (process_options): Likewise.
13128 * cppbuiltin.c: Likewise.
13129 * tsan.c (tsan_pass): Likewise.
13130 (tsan_gate): Likewise.
13131 (tsan_gate_O0): Likewise.
13132 * cfgexpand.c (partition_stack_vars): Likewise.
13133 (expand_stack_vars): Likewise.
13134 (defer_stack_allocation): Likewise.
13135 (expand_used_vars): Likewise.
13136 * cfgcleanup.c (old_insns_match_p): Likewise.
13137 * asan.c (asan_finish_file): Likewise.
13138 (asan_instrument): Likewise.
13139 (gate_asan): Likewise.
13140 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
13141 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
13142 (asan_global_struct): Use pointer_sized_int_node instead
13143 calling build_nonstandard_integer_type.
13144 (initialize_sanitizer_builtins): Likewise.
13145 (asan_finish_file): Likewise.
13146 * gcc.c: Document %{%:function(args):X}.
13147 (static_spec_functions): Add sanitize.
13148 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
13149 store funcval != NULL there.
13150 (do_spec_1): Adjust handle_spec_function caller.
13151 (handle_braces): Allow %:function(args) as condition.
13152 (sanitize_spec_function): New function.
13153 (ADD_STATIC_LIBUBSAN_LIBS): Define.
13154 (LIBUBSAN_SPEC): Likewise.
13155 (LIBUBSAN_EARLY_SPEC): Likewise.
13156 (SANITIZER_SPEC): Handle libubsan.
13157 (SANITIZER_EARLY_SPEC): Likewise.
13158 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
13159 instead of fsanitize=address.
13160 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
13161 instead of fsanitize=address*.
13162 * builtins.c: Include ubsan.h.
13163 (fold_builtin_0): Instrument __builtin_unreachable.
13164 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
13165 instead of flag_asan.
13166 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
13167 (pointer_sized_int_node): Define.
13168 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
13170 2013-08-30 Mike Stump <mikestump@comcast.net>
13172 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
13175 2013-08-29 Jan Hubicka <jh@suse.cz>
13177 * cgraph.c (cgraph_function_body_availability): Handle weakref
13179 * passes.def: Remove pass_fixup_cfg.
13180 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
13181 track when we need to remove functions.
13182 (gate_ipa_inline): Execute inlining always; add comment why.
13183 (pass_data_ipa_inline): Remove TODO_remove_functions.
13184 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
13185 do not produce summaries.
13186 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
13187 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
13188 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
13189 constant pool and vtable.
13191 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
13193 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
13194 New arm_neon.h's internal macros to specify 64-bit constants.
13195 Avoid using stdint.h's macros.
13197 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
13199 * recog.c (verify_changes): Verify that changes[i].old is non-zero
13200 before applying REG_P.
13202 2013-08-30 Jakub Jelinek <jakub@redhat.com>
13204 PR tree-optimization/58277
13205 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
13206 after seeing too many stmts with vdef in between dombb and current
13207 bb, invalidate everything.
13209 2013-08-30 Richard Biener <rguenther@suse.de>
13211 * fold-const.c (fold_single_bit_test): Fix overflow test.
13213 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
13215 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
13216 and which can live in a register, always retrieve the value on entry.
13217 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
13218 passed by invisible reference specially.
13219 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
13220 (vt_add_function_parameter): Correctly deal with a parameter passed by
13221 invisible reference.
13223 2013-08-30 Jan Hubicka <jh@suse.cz>
13225 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
13227 2013-08-30 Richard Biener <rguenther@suse.de>
13229 PR tree-optimization/58228
13230 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
13231 allow invariant loads in nested loop vectorization.
13233 2013-08-30 Richard Biener <rguenther@suse.de>
13235 PR tree-optimization/58223
13236 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
13237 (has_anti_or_output_dependence): ... this and adjust to also
13238 look for output dependences.
13239 (mark_nodes_having_upstream_mem_writes): Adjust.
13240 (rdg_flag_uses): Likewise.
13242 2013-08-30 Richard Biener <rguenther@suse.de>
13244 PR tree-optimization/58010
13245 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
13246 assert that we have a loop-closed PHI.
13248 2013-08-29 Jan Hubicka <jh@suse.cz>
13250 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
13251 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
13252 * lto-section-in.c (lto_free_function_in_decl_state): New function.
13253 (lto_free_function_in_decl_state_for_node): New function.
13255 2013-08-29 Xinliang David Li <davidxl@google.com>
13257 * loop-unroll.c (report_unroll_peel): Minor message change.
13258 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
13259 Emit alignment peeling message with default -fopt-info.
13260 (vect_loop_versioning): Emit loop version info message.
13261 * tree-vectorizer.c (vectorize_loops): Minor message change.
13262 (execute_vect_slp): Ditto.
13264 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
13266 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
13267 of the clone from the DECL_NAME of the original function.
13269 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
13271 * passes.c (register_pass): Add overload.
13272 * tree-pass.h (register_pass): Forward declare it. Add comment.
13274 2013-08-29 Jan Hubicka <jh@suse.cz>
13276 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
13277 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
13278 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
13279 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
13280 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
13281 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
13282 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
13283 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
13285 2013-08-29 Teresa Johnson <tejohnson@google.com>
13287 * dumpfile.c (dump_loc): Output column number.
13288 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
13289 * doc/invoke.texi: Document optall -fopt-info flag.
13290 * profile.c (read_profile_edge_counts): Use new dump framework.
13291 (compute_branch_probabilities): Ditto.
13292 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
13293 when pass not in any opt group.
13294 * pass_manager.h (pass_manager::get_pass_profile): New method.
13295 * value-prof.c (check_counter): Use new dump framework.
13296 (check_ic_target): Ditto.
13297 * coverage.c (get_coverage_counts): Ditto.
13298 (coverage_init): Setup new dump framework.
13300 2013-08-29 Richard Biener <rguenther@suse.de>
13302 PR tree-optimization/58246
13303 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
13304 handle the dominance check inside a basic-block.
13306 2013-08-29 Richard Biener <rguenther@suse.de>
13308 PR middle-end/57287
13309 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
13310 of default defs that appear in abnormal PHI nodes.
13312 2013-08-29 Richard Biener <rguenther@suse.de>
13314 PR tree-optimization/57685
13315 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
13316 single-use operands to avoid exponential complexity.
13318 2013-08-28 Dehao Chen <dehao@google.com>
13320 * ipa-inline.c (edge_badness): Fix integer underflow.
13322 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
13324 * gtm-builtins.def (_ITM_free): Declare leaf.
13326 2013-08-28 Jakub Jelinek <jakub@redhat.com>
13329 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
13330 (*tls_local_dynamic_base_64_largepic): Likewise.
13331 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
13332 Remove predicate from call operand.
13333 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
13334 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
13336 2013-08-28 Jeff Law <law@redhat.com>
13338 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
13339 checks for the number of predecessors and successors allowed.
13340 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
13341 which require copying a joiner block if there is a request which
13342 is a subpath that requires no joiner block copying.
13344 2013-08-28 Jan Hubicka <jh@suse.cz>
13346 * lto-streamer-out.c (DFS_write_tree_body): Drop
13347 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
13348 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
13349 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
13350 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
13351 Do not read DECL_ERROR_ISSUED.
13352 (unpack_ts_decl_with_vis_value_fields): Do not read
13354 (lto_input_ts_binfo_tree_pointers): Do not read
13355 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
13356 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
13357 write DECL_ERROR_ISSUED..
13358 (pack_ts_decl_with_vis_value_fields): Do not write
13360 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
13361 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
13362 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
13363 * tree.h (tree_decl_common): Update comment.
13364 (DECL_ERROR_ISSUED): Remove.
13366 2013-08-28 Jakub Jelinek <jakub@redhat.com>
13368 PR middle-end/58257
13369 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
13371 2013-08-28 Jan Hubicka <jh@suse.cz>
13373 * builtins.def (free): Declare leaf.
13375 2013-08-27 David Malcolm <dmalcolm@redhat.com>
13377 * gdbhooks.py: New.
13378 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
13379 * configure: Regenerate.
13381 2013-08-27 Martin Jambor <mjambor@suse.cz>
13383 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
13384 (ipa_ancestor_jf_data): Likewise.
13385 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
13386 (ipa_get_jf_pass_through_type_preserved): New function.
13387 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
13388 (ipa_get_jf_ancestor_type_preserved): New function.
13389 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
13390 (ipa_get_jf_ancestor_result): Likewise.
13391 (propagate_vals_accross_pass_through): Use
13392 ipa_get_jf_pass_through_result to do all the value mappings.
13393 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
13394 type_preserved flag.
13395 (ipa_set_jf_cst_copy): New function.
13396 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
13397 (ipa_set_jf_arith_pass_through): Likewise.
13398 (ipa_set_ancestor_jf): Likewise.
13399 (compute_complex_assign_jump_func): Set type_preserved instead of
13401 (ipa_compute_jump_functions_for_edge): Likewise.
13402 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
13403 (update_jump_functions_after_inlining): Update type_preserved.
13404 Explicitely create jump functions when combining one with pass_through.
13405 (ipa_write_jump_function): Stream the type_preserved flags.
13406 (ipa_read_jump_function): Likewise.
13408 2013-08-27 Jakub Jelinek <jakub@redhat.com>
13409 Aldy Hernandez <aldyh@redhat.com>
13411 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
13412 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
13413 * function.h (struct function): Add has_force_vect_loops and
13415 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
13416 * gimple.c (gimple_build_omp_critical): Add KIND argument and
13418 * gimple.def: Update CLAUSES comments.
13419 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
13420 (gimple_build_omp_for): Add argument to prototype.
13421 (gimple_omp_for_kind): New.
13422 (gimple_omp_for_set_kind): New.
13423 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
13424 GOVD_DATA_SHARE_CLASS.
13425 (enum omp_region_type): Add ORT_SIMD.
13426 (gimple_add_tmp_var): Handle ORT_SIMD.
13427 (gimplify_var_or_parm_decl): Same.
13428 (is_gimple_stmt): Same.
13429 (omp_firstprivatize_variable): Same.
13430 (omp_add_variable): Only use splay_tree_insert if lookup failed.
13431 (omp_notice_variable): Handle ORT_SIMD.
13432 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
13433 (omp_check_private): Handle ORT_SIMD.
13434 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
13435 OMP_CLAUSE_SAFELEN.
13436 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
13437 Handle OMP_CLAUSE_LASTPRIVATE.
13438 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
13439 OMP_CLAUSE_SAFELEN.
13440 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
13441 (gimplify_expr): Handle OMP_SIMD.
13442 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
13443 (expand_GOMP_SIMD_VF): New.
13444 (expand_GOMP_SIMD_LAST_LANE): New.
13445 * internal-fn.def (GOMP_SIMD_LANE): New.
13446 (GOMP_SIMD_VF): New.
13447 (GOMP_SIMD_LAST_LANE): New.
13448 * omp-low.c: Include target.h.
13449 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
13450 OMP_CLAUSE_SAFELEN.
13451 (check_omp_nesting_restrictions): Same.
13453 (lower_rec_simd_input_clauses): New.
13454 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
13456 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
13457 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
13458 (expand_omp_build_assign): New.
13459 (expand_omp_for_init_counts): New.
13460 (expand_omp_for_init_vars): New.
13461 (extract_omp_for_update_vars): New.
13462 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
13463 and rewrite accordingly.
13464 (expand_omp_simd): New.
13465 (expand_omp_for): Use expand_omp_simd.
13466 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
13467 (lower_omp_for): Do not lower the body.
13468 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
13469 in their own loops.
13470 * tree-flow.h (find_omp_clause): Remove prototype.
13471 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
13472 forcing vectorization of the loop, or if flag_tree_vectorize.
13473 (gate_tree_if_conversion): Similarly.
13474 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
13475 gimple_build_omp_for.
13476 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
13477 * tree-parloops (create_parallel_loop): Pass kind argument to
13478 gimple_build_omp_for.
13479 * tree-pretty-print.c (dump_omp_clause): Add cases for
13480 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
13481 OMP_CLAUSE__SIMDUID_.
13482 (dump_generic_node): Handle OMP_SIMD.
13483 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
13484 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
13485 unroll OMP_SIMD loops here.
13486 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
13487 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
13489 (vect_analyze_data_refs): Handle simd loops.
13490 * tree-vect-loop.c (vectorizable_live_operation): Handle
13492 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
13493 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
13494 (vectorizable_load): Same.
13495 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
13496 (struct simduid_to_vf): New.
13497 (simduid_to_vf::hash): New.
13498 (simduid_to-vf::equal): New.
13499 (struct simd_array_to_simduid): New.
13500 (simd_array_to_simduid::hash): New.
13501 (simd_array_to_simduid::equal): New.
13502 (adjust_simduid_builtins): New.
13503 (struct note_simd_array_uses_struct): New.
13504 (note_simd_array_uses_cb): New.
13505 (note_simd_array_uses): New.
13506 (vectorize_loops): Handle simd hints and adjust simd builtins
13508 * tree-vectorizer.h (struct _stmt_vec_info): Add
13509 simd_lane_access_p field.
13510 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
13511 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
13512 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
13513 (omp_clause_code_name): Same.
13514 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
13515 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
13516 * tree.def (OMP_SIMD): New entry.
13517 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
13518 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
13519 (OMP_CLAUSE_DECL): Adjust range for new clauses.
13520 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
13521 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
13522 (OMP_CLAUSE_LINEAR_STEP): New.
13523 (OMP_CLAUSE_SAFELEN_EXPR): New.
13524 (OMP_CLAUSE__SIMDUID__DECL): New.
13525 (find_omp_clause): New prototype.
13527 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
13529 * config/i386/driver-i386.c (host_detect_local_cpu): Update
13530 Haswell processor detection.
13532 2013-08-27 Christian Widmer <shadow@umbrox.de>
13535 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
13536 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
13537 AVX2 capable processors.
13539 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
13541 * config/aarch64/arm_neon.h: Replace all inline asm implementations
13542 of vget_low_* with implementations in terms of other intrinsics.
13544 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
13546 PR middle-end/57219
13547 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
13548 values to -1, 0 and 1.
13550 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
13552 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
13553 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
13554 (<optab><mode>3_insn): Remove.
13555 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
13556 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
13557 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
13558 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
13559 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
13560 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
13561 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
13562 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
13563 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
13564 (ror<mode>3_insn): Likewise.
13565 * config/aarch64/predicates.md (aarch64_simd_register): New.
13567 2013-08-27 Richard Biener <rguenther@suse.de>
13569 PR tree-optimization/57521
13570 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
13571 one edge is non-critical.
13572 (find_phi_replacement_condition): Make sure to use a non-critical
13573 edge. Cleanup and remove old bug workarounds.
13574 (bb_postdominates_preds): Remove.
13575 (if_convertible_loop_p_1): Do not compute post-dominators.
13576 (combine_blocks): Do not free post-dominators.
13577 (main_tree_if_conversion): Likewise.
13578 (pass_data_if_conversion): Add TODO_verify_ssa.
13580 2013-08-27 DJ Delorie <dj@redhat.com>
13582 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
13584 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
13586 * function.c (assign_parm_find_data_types): Set passed_mode and
13587 nominal_mode to the TYPE_MODE of nominal_type for the built
13588 pointer type in case of the struct-pass-by-reference.
13590 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13592 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
13593 (UINT16_TYPE): Change default to "unsigned int".
13595 * config/avr/avr.opt (mfract-convert-truncate): New option.
13596 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
13597 is set, round negative fractional integers according to n1169
13598 when converting to integer types.
13600 2013-08-26 Jan Hubicka <jh@suse.cz>
13602 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
13603 methods can not be called indirectly when their address is not taken.
13605 2013-08-26 Jan Hubicka <jh@suse.cz>
13607 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
13610 2013-08-26 Jan Hubicka <jh@suse.cz>
13612 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
13615 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13617 * reload.c (find_valid_class): Allow classes that do not include
13618 FIRST_PSEUDO_REGISTER - 1.
13620 2013-08-26 Jan Hubicka <jh@suse.cz>
13622 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
13623 fix edge count/frequency when speculation failed; fix type check
13624 for the direct call.
13626 2013-08-26 Jan Hubicka <jh@suse.cz>
13628 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
13630 2013-08-26 Jan Hubicka <jh@suse.cz>
13632 * ipa-inline-transform.c (inline_transform): Be ready for basic block
13633 to be changed by edge redirection.
13635 2013-08-26 Jan Hubicka <jh@suse.cz>
13637 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
13638 formating; add sanity check.
13639 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
13640 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
13641 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
13642 (resolve_noninline_speculation): Update callee keys, too.
13644 2013-08-26 Jan Hubicka <jh@suse.cz>
13646 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
13647 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
13649 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13651 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
13654 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
13656 * config/i386/i386.c (ix86_debug_options): Remove prototype.
13657 (x86_64_elf_select_section): Ditto.
13658 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
13660 (ix86_pass_by_reference): Ditto.
13661 (output_set_got): Ditto.
13662 (ix86_unary_operator_ok): Ditto.
13663 (ix86_expand_builtin): Ditto.
13665 2013-08-23 Jan Hubicka <jh@suse.cz>
13667 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
13669 2013-08-23 Jan Hubicka <jh@suse.cz>
13671 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
13672 (tree_decl_with_vis): Add FINAL field.
13674 2013-08-23 Jeff Law <law@redhat.com>
13676 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
13677 the debugging dump when the expression is fully redundant.
13679 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
13681 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
13682 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
13683 * pretty-print.c (pp_formatted_text_data): Likewise.
13684 (pp_write_text_to_stream): Likewise.
13685 (pp_write_text_as_dot_label_to_stream): Likewise.
13686 (pp_append_r): Likewise.
13687 (pp_format): Likewise.
13688 (pp_flush): Likewise.
13689 (pp_clear_output_area): Likewise.
13690 (pp_append_text): Likewise.
13691 (pp_formatted_text): Likewise.
13692 (pp_remaining_character_count_for_line): Likewise.
13693 (pp_newline): Likewise.
13694 (pp_character): Likewise.
13695 (output_buffer::~output_buffer): Define.
13696 (pretty_printer::~pretty_printer): Destruct output buffer.
13697 * pretty-print.h (output_buffer::~output_buffer): Declare.
13698 (pretty_printer::~pretty_printer): Declare virtual.
13700 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
13703 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
13704 HOST_WIDE_INT_M1U): New macros.
13705 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
13706 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
13707 unsigned -1 for lshift.
13708 * cse.c (cse_insn): Likewise.
13709 * double-int.c (rshift_double, lshift_double): Likewise.
13710 * builtins.c (fold_builtin_bitop): Likewise.
13711 * combine.c (force_to_mode): Likewise.
13712 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
13713 * simplify-rtx.c (simplify_const_unary_operation,
13714 simplify_const_binary_operation): Likewise.
13715 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
13716 check_va_list_escapes): Likewise.
13717 * rtlanal.c (nonzero_bits1): Likewise.
13718 * expmed.c (expand_smod_pow2): Likewise.
13719 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
13721 2013-08-23 Jan Hubicka <jh@suse.cz>
13723 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
13724 as having address taken.
13726 2013-08-23 Jan Hubicka <jh@suse.cz>
13728 * ipa-utils.h (method_class_type): Declare.
13729 * ipa-devirt.c (method_class_type): Export.
13731 * cgraphunit.c (analyze_functions): Do basic devirtualization;
13732 do not walk base classes of anonymous types.
13734 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
13736 PR rtl-optimization/58220
13737 PR regression/58221
13738 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
13739 handle SEQUENCE insns properly.
13741 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
13743 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
13745 (pp_newline_and_indent): Likewise.
13746 (pp_separate_with): Likewise.
13747 * pretty-print.c (pp_newline_and_flush): Define.
13748 (pp_newline_and_indent): Likewise.
13749 (pp_separate_with): Likewise.
13751 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13754 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
13755 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
13757 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
13759 * config/i386/predicates.md (ext_sse_reg_operand): New.
13760 * config/i386/i386.md (*movti_internal): Use
13761 predicate to determine if EVEX is needed.
13762 (*movsi_internal): Ditto.
13763 (*movdf_internal): Ditto.
13764 (*movsf_internal): Ditto.
13765 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
13767 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13769 PR tree-optimization/58209
13770 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
13771 (find_tail_calls): Give up for pointer result types if m is non-NULL.
13772 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
13773 emit POINTER_PLUS_EXPR.
13774 (create_tailcall_accumulator): For pointer result type accumulate in
13777 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
13779 * configure.ac: Add backslashes missing from the last change.
13780 * configure: Regenerate.
13782 2013-08-22 Jan Hubicka <jh@susue.cz>
13784 * ipa.c (function_and_variable_visibility): First remember function
13785 was global and then make it local.
13787 2013-08-22 Julian Brown <julian@codesourcery.com>
13789 * configure.ac: Add aarch64 to list of arches which use "nop" in
13791 * configure: Regenerate.
13793 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13795 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
13796 gnu_libc_has_function.
13797 * config/s390/tpf.h: Likewise.
13799 2013-08-22 Jan Hubicka <jh@susue.cz>
13801 * timevar.c (validate_phases): Add cast.
13803 2013-08-22 Jan Hubicka <jh@susue.cz>
13805 * timevar.c (validate_phases): Use size_t for memory.
13806 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
13808 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
13810 * pretty-print.h (output_buffer::output_buffer): Declare.
13811 (pretty_printer::pretty_printer): Likewise.
13812 (pp_construct): Remove.
13813 * pretty-print.c (output_buffer::output_buffer): Define.
13814 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
13815 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
13816 (print_gimple_expr): Likewise.
13817 (print_gimple_seq): Likewise.
13818 (gimple_dump_bb): Likewise.
13819 * sched-vis.c (dump_value_slim): Likewise.
13820 (dump_insn_slim): Likewise.
13821 (dump_rtl_slim): Likewise.
13822 (str_pattern_slim): Likewise.
13823 * tree-mudflap.c (mf_varname_tree): Likewise.
13824 * graph.c (print_graph_cfg): Likewise.
13825 (start_graph_dump): Likewise.
13826 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
13828 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
13829 pretty printer initialization.
13830 * coretypes.h (diagnostic_context): Remove superflous type alias
13832 (pretty_printer): Likewise. Declare directly as a class.
13833 (pretty_print_info): Remove declaration as class.
13834 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
13835 and pp_clear_output_area.
13836 (asan_add_global): Likewise.
13838 2013-08-22 Jan Hubicka <jh@suse.cz>
13840 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
13841 * ipa-utils.h (update_type_inheritance_graph): Declare.
13842 (possible_polymorphic_call_target_p): Declare.
13843 (possible_polymorphic_call_target_p): New.
13844 * ipa-devirt.c: Update toplevel comments.
13845 (cached_polymorphic_call_targets): Move up.
13846 (odr_type_d): Move ID down.
13847 (polymorphic_type_binfo_p): Update comment.
13848 (odr_hasher::remove): Likewise;
13849 (get_odr_type): Set anonymous_namespace.
13850 (dump_odr_type): Dump it.
13851 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
13852 (maybe_record_node): Record node in cached_polymorphic_call_targets.
13853 (record_binfo): Add comment.
13854 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
13856 (devirt_node_removal_hook): Do not iCE when cache is freed.
13857 (possible_polymorphic_call_target_p): New predicate.
13858 (update_type_inheritance_graph): New function.
13860 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
13861 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
13862 Sergey Lega <sergey.s.lega@intel.com>
13863 Anna Tikhonova <anna.tikhonova@intel.com>
13864 Ilya Tocar <ilya.tocar@intel.com>
13865 Andrey Turetskiy <andrey.turetskiy@intel.com>
13866 Ilya Verbin <ilya.verbin@intel.com>
13867 Kirill Yukhin <kirill.yukhin@intel.com>
13868 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13870 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
13871 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
13872 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
13873 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
13874 (OPTION_MASK_ISA_AVX2_UNSET): Update.
13875 (OPTION_MASK_ISA_AVX512F_UNSET): New.
13876 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
13877 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
13878 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
13879 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
13880 OPT_mavx512pf, OPT_mavx512er cases.
13881 * config/i386/constraints.md (v): New constraint.
13882 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
13883 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
13884 (bit_AVX512CD): New.
13885 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13886 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
13887 * config/i386/i386-c.c (ix86_target_macros_internal):
13888 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
13890 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
13891 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
13892 * config/i386/i386.c (regclass_map, dbx_register_map)
13893 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
13894 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
13895 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
13896 -mavx512pf options.
13897 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
13898 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
13899 -mavx512pf options. Fix formatting.
13900 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
13901 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
13902 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
13903 -mavx512cd, -mavx512pf options.
13904 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
13905 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
13906 (ix86_preferred_output_reload_class): Replace SSE_REGS with
13908 (ix86_hard_regno_mode_ok): Support 512-bit registers.
13909 (ix86_set_reg_reg_cost): Ditto.
13910 (x86_order_regs_for_local_alloc): Ditto.
13911 (MAX_VECT_LEN): Extend to 64-byte.
13912 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
13913 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
13914 (TARGET_AVX512ER, TARGET_AVX512CD): New.
13915 (BIGGEST_ALIGNMENT): Extend to 512-bits.
13916 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
13917 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
13918 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
13919 (SSE_REG_MODE_P): Support new modes.
13920 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
13921 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
13922 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
13923 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
13924 (REG_CLASS_CONTENTS): Add new registers.
13925 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
13926 (EXT_REX_SSE_REGNO_P): New.
13927 (HI_REGISTER_NAMES): Add new registers.
13928 * config/i386/i386.md: Define constants for new registers.
13929 (mode): Add new 512-bit modes.
13930 (prefix): Support evex prefix.
13931 (isa): Support avx512f, noavx512f, fma_avx512f.
13932 (ssemodesuffix): Add new 512-bit modes.
13934 (*movxi_internal_avx512f): Ditto.
13935 (*movdi_internal): Replace constraint "x" with the new constraint "v".
13937 (*movsi_internal): Likewise.
13938 (*movdf_internal): Likewise.
13939 (*movsf_internal): Likewise.
13940 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
13941 (<code><mode>3): Likewise.
13942 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
13944 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
13945 with the new constraint "v".
13946 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
13948 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
13949 with the new constraint "v".
13950 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
13951 (<sse2>_storedqu<avxsizesuffix>): Likewise.
13952 (*<plusminus_insn><mode>3): Likewise.
13953 (<sse>_vm<plusminus_insn><mode>3): Likewise.
13954 (*mul<mode>3): Likewise.
13955 (<sse>_vmmul<mode>3): Likewise.
13956 (<sse>_div<mode>3): Likewise.
13957 (<sse>_vmdiv<mode>3): Likewise.
13958 (<sse>_sqrt<mode>2): Likewise.
13959 (<sse>_vmsqrt<mode>2): Likewise.
13960 (*<code><mode>3_finite): Likewise.
13961 (*<code><mode>3) <smaxmin>: Likewise.
13962 (<sse>_vm<code><mode>3): Likewise.
13963 (*<code><mode>3) <any_logic>: Likewise.
13964 (*fma_fmadd_<mode>): Likewise.
13965 (*fma_fmsub_<mode>): Likewise.
13966 (*fma_fnmadd_<mode>): Likewise.
13967 (*fma_fnmsub_<mode>): Likewise.
13968 (*fma_fmaddsub_<mode>): Likewise.
13969 (*fma_fmsubadd_<mode>): Likewise.
13970 (*fmai_fmadd_<mode>): Likewise.
13971 (*fmai_fmsub_<mode>): Likewise.
13972 (*fmai_fnmadd_<mode>): Likewise.
13973 (*fmai_fnmsub_<mode>): Likewise.
13974 (sse_cvtsi2ss): Likewise.
13975 (sse_cvtsi2ssq): Likewise.
13976 (sse_cvtss2si): Likewise.
13977 (sse_cvtss2si_2): Likewise.
13978 (sse_cvtss2siq): Likewise.
13979 (sse_cvtss2siq_2): Likewise.
13980 (sse_cvttss2si): Likewise.
13981 (sse_cvtss2siq_2): Likewise.
13982 (float<sseintvecmodelower><mode>2): Likewise.
13983 (sse2_cvtsd2si_2): Likewise.
13984 (sse2_cvtsd2siq_2): Likewise.
13985 (*<plusminus_insn><mode>3): Likewise.
13986 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
13987 (*<sse4_1_avx2>_mul<mode>3): Likewise.
13988 (ashr<mode>3): Likewise.
13989 (<shift_insn><mode>3): Likewise.
13990 (avx2_<code><mode>3): Likewise.
13991 (*avx2_<code><mode>3): Likewise.
13992 (*andnot<mode>3): Likewise.
13993 (*<code><mode>3) <any_logic>: Likewise.
13994 (abs<mode>2): Likewise.
13995 (avx2_permvar<mode>): Likewise.
13996 (avx2_perm<mode>_1): Likewise.
13997 (*avx_vpermilp<mode>): Likewise.
13998 (avx_vpermilvar<mode>3): Likewise.
13999 (avx2_ashrv<mode>): Likewise.
14000 (avx2_<shift_insn>v<mode>): Likewise.
14001 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
14003 * doc/rtl.texi: Document XImode.
14005 2013-08-21 Jeff Law <law@redhat.com>
14007 * tree-flow.h (register_jump_thread): Pass vector of edges
14008 instead of each important edge.
14009 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
14010 thread path into a vector and pass that to register_jump_thread.
14011 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
14012 passed in edge vector to the current 3-edge form.
14015 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
14017 * dce.c (fini_dce): Call df_analyze again just in case
14018 delete_unmarked_insns removed anything.
14020 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
14022 * reload.h (struct reg_equivs): Rename to ..
14023 (struct reg_equivs_s): .. this.
14025 2013-08-20 Martin Liska <marxin.liska@gmail.com>
14027 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
14029 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14031 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
14033 2013-08-21 Jeff Law <law@redhat.com>
14035 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
14036 simplify assignments too. If the RHS collapses to a singleton
14037 range, then return the value for the range.
14039 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
14041 * config/i386/sse.md (V16): Rename to...
14043 (mov<mode>): Update iterator name.
14044 (*mov<mode>_internal): Ditto.
14045 (push<mode>1): Ditto.
14046 (movmisalign<mode>): Ditto.
14048 2013-08-20 Jan Hubicka <jh@suse.cz>
14051 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
14052 entry for direct edges.
14053 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
14055 2013-08-20 David Malcolm <dmalcolm@redhat.com>
14057 Revert my last two changes, r201865 and r201864:
14060 2013-08-20 David Malcolm <dmalcolm@redhat.com>
14062 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
14063 instances can own GC refs.
14065 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
14066 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
14067 (gcc::context::gt_pch_nx): Likewise.
14068 (gcc::context::gt_pch_nx): Likewise.
14069 * ggc.h (gt_ggc_mx <T>): New.
14070 (gt_pch_nx_with_op <T>): New.
14071 (gt_pch_nx <T>): New.
14072 * passes.c (opt_pass::gt_ggc_mx): New.
14073 (opt_pass::gt_pch_nx): New.
14074 (opt_pass::gt_pch_nx_with_op): New.
14075 (pass_manager::gt_ggc_mx): New.
14076 (pass_manager::gt_pch_nx): New.
14077 (pass_manager::gt_pch_nx_with_op): New.
14078 (pass_manager::operator new): Use
14079 ggc_internal_cleared_alloc_stat rather than xcalloc.
14080 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
14081 (pass_manager::gt_ggc_mx): New.
14082 (pass_manager::gt_pch_nx): New.
14083 (pass_manager::gt_pch_nx_with_op): New.
14084 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
14085 (opt_pass::operator new): New.
14086 (opt_pass::gt_ggc_mx): New.
14087 (opt_pass::gt_pch_nx): New.
14088 (opt_pass::gt_pch_nx_with_op): New.
14091 2013-08-20 David Malcolm <dmalcolm@redhat.com>
14093 * Makefile.in (GTFILES): Add context.h.
14094 * context.c (gcc::context::operator new): New.
14095 (gcc::context::gt_ggc_mx): New.
14096 (gcc::context::gt_pch_nx): New.
14097 (gcc::context::gt_pch_nx): New.
14098 * context.h (gcc::context): Add GTY((user)) marking.
14099 (gcc::context::operator new): New.
14100 (gcc::context::gt_ggc_mx): New.
14101 (gcc::context::gt_pch_nx): New.
14102 (gcc::context::gt_pch_nx): New.
14103 (g): Add GTY marking.
14104 (gt_ggc_mx (gcc::context *)): New.
14105 (gt_pch_nx (gcc::context *)): New.
14106 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
14107 void *cookie)): New.
14108 * gengtype.c (open_base_files) <ifiles>: Add context.h.
14110 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
14112 * dce.c (fini_dce): Call df_analyze again just in case
14113 delete_unmarked_insns removed anything.
14115 2013-08-20 Teresa Johnson <tejohnson@google.com>
14117 PR rtl-optimizations/57451
14118 * final.c (reemit_insn_block_notes): Prevent lexical blocks
14119 from crossing split section boundaries.
14121 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
14123 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
14124 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
14125 with MULTLIB_DEFAULTS.
14127 2013-08-20 Nick Clifton <nickc@redhat.com>
14129 * target.def (narrow_volatile_bitfield): Note that the default
14130 value is false, not !TARGET_STRICT_ALIGN.
14131 * doc/tm.texi: Regenerate.
14133 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
14135 Fix LIB_SPEC for systems without libpthread.
14137 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
14138 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
14140 * config/i386/linux-common.h: Likewise.
14141 * config/mips/linux-common.h: Likewise.
14143 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
14145 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
14148 2013-08-20 David Malcolm <dmalcolm@redhat.com>
14150 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
14151 instances can own GC refs.
14153 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
14154 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
14155 (gcc::context::gt_pch_nx): Likewise.
14156 (gcc::context::gt_pch_nx): Likewise.
14157 * ggc.h (gt_ggc_mx <T>): New.
14158 (gt_pch_nx_with_op <T>): New.
14159 (gt_pch_nx <T>): New.
14160 * passes.c (opt_pass::gt_ggc_mx): New.
14161 (opt_pass::gt_pch_nx): New.
14162 (opt_pass::gt_pch_nx_with_op): New.
14163 (pass_manager::gt_ggc_mx): New.
14164 (pass_manager::gt_pch_nx): New.
14165 (pass_manager::gt_pch_nx_with_op): New.
14166 (pass_manager::operator new): Use
14167 ggc_internal_cleared_alloc_stat rather than xcalloc.
14168 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
14169 (pass_manager::gt_ggc_mx): New.
14170 (pass_manager::gt_pch_nx): New.
14171 (pass_manager::gt_pch_nx_with_op): New.
14172 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
14173 (opt_pass::operator new): New.
14174 (opt_pass::gt_ggc_mx): New.
14175 (opt_pass::gt_pch_nx): New.
14176 (opt_pass::gt_pch_nx_with_op): New.
14178 2013-08-20 David Malcolm <dmalcolm@redhat.com>
14180 * Makefile.in (GTFILES): Add context.h.
14181 * context.c (gcc::context::operator new): New.
14182 (gcc::context::gt_ggc_mx): New.
14183 (gcc::context::gt_pch_nx): New.
14184 (gcc::context::gt_pch_nx): New.
14185 * context.h (gcc::context): Add GTY((user)) marking.
14186 (gcc::context::operator new): New.
14187 (gcc::context::gt_ggc_mx): New.
14188 (gcc::context::gt_pch_nx): New.
14189 (gcc::context::gt_pch_nx): New.
14190 (g): Add GTY marking.
14191 (gt_ggc_mx (gcc::context *)): New.
14192 (gt_pch_nx (gcc::context *)): New.
14193 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
14194 void *cookie)): New.
14195 * gengtype.c (open_base_files) <ifiles>: Add context.h.
14197 2013-08-20 Alan Modra <amodra@gmail.com>
14200 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
14201 (rs6000_emit_epilogue): Likewise.
14203 2013-08-19 Dehao Chen <dehao@google.com>
14205 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
14207 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
14208 Jakub Jelinek <jakub@redhat.com>
14210 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
14211 (BUILT_IN_FABSD64): Likewise.
14212 (BUILT_IN_FABSD128): Likewise.
14213 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
14214 (fold_builtin_1): Likewise.
14215 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
14216 destination and source operands.
14217 (*abstd2_fpr): Likewise.
14218 (*nabstd2_fpr): Likewise.
14220 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
14222 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
14225 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
14227 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
14230 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
14232 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
14233 * builtins.c (default_libc_has_function): New.
14234 (gnu_libc_has_function): Ditto.
14235 (no_c99_libc_has_function): Ditto.
14236 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
14237 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
14238 (fold_builtin_sincos): Likewise.
14239 (fold_builtin_cexp): Likewise.
14240 * builtins.def (DEF_C94_BUILTIN): Likewise.
14241 (DEF_C99_BUILTIN): Likewise.
14242 (DEF_C99_C90RES_BUILTIN): Likewise.
14243 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
14244 definitions to using this define.
14245 * config/darwin-protos.h (darwin_libc_has_function): New.
14246 * config/darwin.c (darwin_libc_has_function): Ditto.
14247 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
14248 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
14249 * config/darwin.h: Ditto.
14250 * config/elfos.h: Ditto.
14251 * config/freebsd.h: Ditto.
14252 * config/i386/cygming.h: Ditto.
14253 * config/i386/djgpp.h: Ditto.
14254 * config/i386/i386-interix.h: Ditto.
14255 * config/microblaze/microblaze.h: Ditto.
14256 * config/mmix/mmix.h: Ditto.
14257 * config/gnu-user.h: Ditto.
14258 * config/ia64/hpux.h: Ditto.
14259 * config/pa/pa-hpux.h: Ditto.
14260 * config/pdp11/pdp11.h: Ditto.
14261 * config/picochip/picochip.h: Ditto.
14262 * config/linux.h: Ditto.
14263 * config/netbsd.h: Ditto.
14264 * config/openbsd.h: Ditto.
14265 * config/rs6000/aix43.h: Ditto.
14266 * config/rs6000/aix51.h: Ditto.
14267 * config/rs6000/aix52.h: Ditto.
14268 * config/rs6000/aix53.h: Ditto.
14269 * config/rs6000/aix61.h: Ditto.
14270 * config/rs6000/darwin.h: Ditto.
14271 * config/rs6000/linux.h: Ditto.
14272 * config/rs6000/linux64.h: Ditto.
14273 * config/s390/tpf.h: Ditto.
14274 * config/sol2-10.h: Ditto.
14275 * config/sol2.h: Ditto.
14276 * config/vms/vms.h: Ditto.
14277 * config/vxworks.h: Ditto.
14278 * config/linux-android.c (linux_android_libc_has_function):
14279 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
14280 * config/linux-protos.h (linux_android_libc_has_function):
14282 * config/i386/i386.c (ix86_libc_has_function): New.
14283 * config/i386/i386-protos.h
14284 (ix86_libc_has_function): New declaration.
14285 * config/i386/i386.md
14286 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
14287 ("isinf<mode>2): Likewise.
14288 * convert.c (convert_to_integer): Using new target hook
14289 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
14290 TARGET_C99_FUNCTIONS.
14291 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
14292 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
14293 * coretypes.h (function_class): New enum for different
14294 classes of functions.
14295 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
14296 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
14297 (TARGET_HAS_SINCOS): Likewise.
14298 (TARGET_LIBC_HAS_FUNCTION): New.
14299 * doc/tm.texi: Regenerated.
14300 * targhooks.h (default_libc_has_function): New declaration.
14301 (no_c99_libc_has_function): Ditto.
14302 (gnu_libc_has_function): Ditto.
14303 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
14304 and TARGET_HAS_SINCOS.
14306 2013-08-18 Jan Hubicka <jh@suse.cz>
14308 * Makeifle-in (ipa-devirt.o): New.
14309 (GTFILES): Add ipa-utils.h and ipa-devirt.c
14310 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
14311 (analyze_functions): Look into possible targets of polymorphic call.
14312 * dumpfile.c (dump_files): Add type-inheritance dump.
14313 * dumpfile.h (TDI_inheritance): New.
14314 * ipa-devirt.c: New file.
14315 * ipa-utils.h (odr_type_d): Forward declare.
14316 (odr_type): New type.
14317 (build_type_inheritance_graph): Declare.
14318 (possible_polymorphic_call_targets): Declare and introduce inline
14319 variant when only edge is pased.
14320 (dump_possible_polymorphic_call_targets): Likewise.
14321 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
14322 * tree.c (type_in_anonymous_namespace_p): Break out from ...
14323 (types_same_for_odr): ... here.
14324 * tree.h (type_in_anonymous_namespace_p): Declare.
14326 2013-08-18 Jakub Jelinek <jakub@redhat.com>
14328 PR tree-optimization/58006
14329 * tree-parloops.c (take_address_of): Don't ICE if get_name
14331 (eliminate_local_variables_stmt): Remove clobber stmts.
14333 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
14335 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
14336 error message is issued for an alias to undefined symbol.
14338 2013-08-18 Jan Hubicka <jh@suse.cz>
14340 * cgraph.c (cgraph_create_indirect_edge): Discover
14341 polymorphic calls and record basic info into indirect_info.
14342 * gimple-fold.c (gimple_fold_call): When doing BINFO based
14343 devirtualization, ignore objc function calls.
14344 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
14345 call with no parm index info.
14346 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
14347 * tree.c (virtual_method_call_p): New function.
14348 * tree.h (virtual_method_call_p): Declare.
14350 2013-08-16 Jan Hubicka <jh@suse.cz>
14352 PR middle-end/58179
14353 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
14355 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
14357 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
14360 2013-08-16 David Malcolm <dmalcolm@redhat.com>
14362 * gengtype.c (type_for_name): Add special-case support for
14363 locating types within the "gcc::" namespace.
14364 (open_base_files): Emit a "using namespace gcc" directive.
14366 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
14369 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
14370 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
14372 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
14373 array instead of each individual operand as a separate argument.
14374 (emit_fusion_gpr_load): Likewise.
14375 (expand_fusion_gpr_load): Add new function declaration.
14377 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
14378 signature to have the operands passed as an array, instead of as
14379 separate arguments. Allow ZERO_EXTEND to be in the memory
14380 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
14381 depend on the register live/dead flags when peepholes are run.
14382 (expand_fusion_gpr_load): New function to be called from the
14383 peephole2 pass, to change the register that addis sets to be the
14385 (emit_fusion_gpr_load): Change the calling signature to have the
14386 operands passed as an array, instead of as separate arguments.
14387 Allow ZERO_EXTEND to be in the memory address, and also
14388 SIGN_EXTEND if -mpower8-fusion-sign.
14390 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
14391 unspec enumeration.
14392 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
14393 adjust the register addis loads up in the peephole2 pass. Do not
14394 depend on the register live/dead state when the peephole pass is done.
14396 2013-08-16 David Malcolm <dmalcolm@redhat.com>
14398 * gengtype.c (create_user_defined_type): Ensure that the kind
14399 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
14400 declaration is seen before the GTY((user)) marking.
14402 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
14405 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
14407 2013-08-16 Jan Hubicka <jh@suse.cz>
14409 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
14410 arugment expected_type.
14411 (gimple_fold_call): Use it.
14412 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
14413 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
14414 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
14415 (try_make_edge_direct_virtual_call): Likewise.
14416 * tree.c (obj_type_ref_class): New.
14417 * tree.h (obj_type_ref_class): Use it.
14419 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
14421 * sched-vis.c (rtl_slim_pp_initialized): Remove.
14422 (rtl_slim_pp): Likewise.
14423 (init_rtl_slim_pretty_print): Likewise.
14424 (dump_value_slim): Don't call it. Use local pretty printer.
14425 (dump_insn_slim): Likewise.
14426 (dump_rtl_slim): Likewise.
14427 (str_pattern_slim): Likewise.
14428 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
14431 2013-08-16 Jakub Jelinek <jakub@redhat.com>
14433 PR tree-optimization/58164
14434 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
14435 walk gimple_goto_dest of GIMPLE_GOTO.
14437 PR tree-optimization/58165
14438 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
14439 bi_call must be the last stmt in a bb, don't split_block, instead
14440 use fallthru edge from it and give up if there is none.
14441 Release conds vector when returning early.
14443 2013-08-14 Xinliang David Li <davidxl@google.com>
14445 * config/i386/i386.c (ix86_option_override_internal):
14446 Remove unused variable and field.
14448 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14451 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
14452 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
14453 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
14454 and Linux, correct BLKmode alignment when 128-bit alignment is
14455 required and compatibility flag is not set.
14456 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
14457 for zero-size arguments when compatibility flag is not set.
14459 2013-08-14 Jakub Jelinek <jakub@redhat.com>
14461 PR tree-optimization/58145
14462 * tree-sra.c (build_ref_for_offset): If prev_base has
14463 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
14465 2013-08-14 Xinliang David Li <davidxl@google.com>
14467 * config/i386/i386.c (ix86_option_override_internal):
14468 Fix uninitialized variable error.
14470 2013-08-14 Xinliang David Li <davidxl@google.com>
14472 * config/i386/i386.opt: Define two new options.
14473 * config/i386/x86-tune.def: Add arch selector field in macros.
14474 * config/i386/i386.h: Adjust macro definition.
14475 * config/i386/i386.c (ix86_option_override_internal):
14477 (parse_mtune_ctrl_str): New function.
14478 (set_ix86_tune_features): New function.
14479 (ix86_function_specific_restore): Call the new helper function.
14481 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
14483 PR rtl-optimization/57662
14484 * sel-sched.c (code_motion_process_successors): When the current insn
14485 is removed after the recursive traversal, break from the loop.
14486 Add comments and debug printouts.
14488 2013-08-14 Jakub Jelinek <jakub@redhat.com>
14489 Alexandre Oliva <aoliva@redhat.com>
14492 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
14493 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
14494 there also UNSPEC_PLTOFF.
14496 2013-08-14 Marek Polacek <polacek@redhat.com>
14498 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
14499 NUM_CONDITIONS bit positions.
14501 2013-08-13 Cary Coutant <ccoutant@google.com>
14503 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
14504 (attr_checksum): Hash vector contents instead of pointer.
14505 (attr_checksum_ordered): Likewise.
14507 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
14509 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
14510 when Pmode != word_mode. Add length_address attribute.
14511 (sse3_monitor_<mode>): Merge from sse3_monitor and
14512 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
14513 Pmode != word_mode. Update insn length attribute.
14514 * config/i386/i386.c (ix86_option_override_internal): Update
14515 ix86_gen_monitor selection for merged sse3_monitor insn.
14517 2013-08-13 Julian Brown <julian@codesourcery.com>
14519 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
14520 perform invalid legitimization on greater-than-word-size modes for
14521 TARGET_E500_DOUBLE.
14523 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
14525 * ira.c (setup_class_translate_array): Use aclass instead of cl
14526 for classes not fully covered by allocno classes.
14528 2013-08-13 Jakub Jelinek <jakub@redhat.com>
14530 PR tree-optimization/57661
14531 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
14532 * tree-inline.c (tree_function_versioning): Initialize it.
14533 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
14534 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
14535 that is not being copied.
14538 * asan.c (instrument_strlen_call): Fix typo in comment.
14539 Use char * type even for the lhs of POINTER_PLUS_EXPR.
14541 2013-08-13 Steve Ellcey <sellcey@mips.com>
14543 * config/mips/mips.md (prefetch): Use lw instead of ld on
14544 loongson in 32bit mode.
14546 2013-08-13 Nick Clifton <nickc@redhat.com>
14548 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
14550 2013-08-13 Jan Hubicka <jh@suse.cz>
14552 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
14553 introduced edge; fix typo in sanity check.
14554 (cgraph_resolve_speculation): Export; improve diagnostic.
14555 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
14556 speculation at type mismatch.
14557 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
14558 (cgraph_resolve_speculation): Declare.
14559 (symtab_can_be_discarded): New function.
14560 * value-prof.c (gimple_ic_transform): Remove actual transform code.
14561 * ipa-inline-transform.c (speculation_removed): New global var.
14562 (clone_inlined_nodes): See if speculation can be removed.
14563 (inline_call): If speculations was removed, we growths may not match.
14564 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
14565 (speculation_useful_p): New function.
14566 (resolve_noninline_speculation): New function.
14567 (inline_small_functions): Resolve useless speculations.
14568 * ipa-inline.h (speculation_useful_p): Declare
14569 * ipa.c (can_replace_by_local_alias): Simplify.
14570 (ipa_profile): Produce speculative calls in non-lto, too;
14571 add simple cost model; produce local aliases.
14573 2013-08-13 David Malcolm <dmalcolm@redhat.com>
14575 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
14578 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
14580 * config/i386/i386.c (ix86_function_versions): Use error + inform.
14582 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
14584 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
14585 iterator instead of X87MODEF.
14587 2013-08-12 Perez Read <netfirewall@gmail.com>
14590 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
14591 operand 0 for intel asm alternative.
14592 (*movabs<mode>_2): Ditto for operand 1.
14594 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
14596 * config/aarch64/arm_none.h
14597 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
14599 2013-08-12 Nick Clifton <nickc@redhat.com>
14601 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
14603 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
14605 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
14606 expand for QI/HImode operand to produce more effictive code for
14607 unsigned char(short) --> float(double) conversion.
14609 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
14611 * doc/invoke.texi: Mention that -ftls-model does not force the final
14614 2013-08-12 Marek Polacek <polacek@redhat.com>
14615 Marc Glisse <marc.glisse@inria.fr>
14617 PR tree-optimization/57980
14618 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
14619 when creating -1 constant.
14621 2013-08-10 Jan Hubicka <jh@suse.cz>
14623 Workaround binutils PR14342.
14624 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
14625 (gimple_init_edge_profiler): Likewise.
14626 (gimple_gen_ic_func_profiler): Likewise.
14628 2013-08-09 Jan Hubicka <jh@suse.cz>
14630 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
14632 2013-08-09 Xinliang David Li <davidxl@google.com>
14634 * config/i386/stringop.def: New file.
14635 * config/i386/stringop.opt: New file.
14636 * config/i386/i386-opts.h: Include stringopt.def.
14637 * config/i386/i386.opt: Include stringopt.opt.
14638 * config/i386/i386.c (ix86_option_override_internal):
14639 Override default size based stringop inline strategies with options.
14640 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
14643 2013-08-09 Jan Hubicka <jh@suse.cz>
14645 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
14647 2013-08-09 Jan Hubicka <jh@suse.cz>
14649 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
14651 (dump_cgraph_node): Dump profile-id.
14652 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
14653 and common_target_probability.
14654 * lto-cgraph.c (lto_output_edge): Stream common targets.
14655 (lto_output_node): Stream profile ids.
14656 (input_node): Stream profile ids.
14657 (input_edge): Stream common targets.
14658 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
14659 * ipa.c: Include value-prof.h
14660 (ipa_profile_generate_summary): Turn indirect call statement histograms
14661 into common targets.
14662 (ipa_profile): Turn common targets into speculative edges.
14664 2013-08-09 Jan Hubicka <jh@suse.cz>
14666 * cgraph.h (cgraph_node): Add profile_id.
14667 * value-prof.c (cgraph_node_map): Turn into pointer_map.
14668 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
14669 (del_node_map): Update.
14670 (find_func_by_funcdef_no): Replace by ...
14671 (find_func_by_profile_id): ... this one.
14672 (gimple_ic_transform): Do not remove useful histograms when
14673 speculation is not done; dump info when indirect call removal
14675 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
14676 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
14677 (__gcov_indirect_call_profiler_v2): .. this one.
14678 * profile.h (init_node_map): Update.
14679 * coverage.c (coverage_compute_profile_id): New function.
14680 * coverage.h (coverage_compute_profile_id): Declare.
14681 * tree-profile.c (init_ic_make_global_vars): Make
14682 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
14683 (gimple_init_edge_profiler): Update prototype of
14684 __gcov_indirect_call_profiler.
14685 (gimple_gen_ic_func_profiler): Simplify.
14686 (tree_profiling): Use init_node_map
14688 2013-08-09 Jan Hubicka <jh@suse.cz>
14690 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
14691 non-speculative refs.
14692 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
14693 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
14694 (cgraph_set_call_stmt): Likewise.
14695 (cgraph_create_edge_1): Fix release checking compilatoin;
14696 clear lto_stmt_uid.
14697 (cgraph_free_edge): Free indirect info.
14698 (cgraph_turn_edge_to_speculative): New function.
14699 (cgraph_speculative_call_info): New function.
14700 (cgraph_make_edge_direct): Return direct edge; handle speculation.
14701 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
14702 (dump_cgraph_node): Dump speculation.
14703 (verify_edge_count_and_frequency): Accept speculative edges.
14704 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
14705 (verify_cgraph_node): Handle speculation.
14706 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
14707 (cgraph_set_call_stmt): Update prototype.
14708 (cgraph_make_edge_direct): Update prototype.
14709 (cgraph_speculative_call_info): Declare.
14710 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
14711 to change; update call of ipa_find_references.
14712 * ipa-ref.c (ipa_record_reference): Fix return value; clear
14713 lto_stmt_uid and speculative flags.
14714 (ipa_dump_references): Dump speculation.
14715 (ipa_clone_references): Clone speculative flag.
14716 (ipa_clone_referring): Likewise.
14717 (ipa_clone_ref): New function.
14718 (ipa_find_reference): Look into lto_stmt_uids
14719 (ipa_clear_stmts_in_references): Do not clear speculative calls.
14720 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
14721 (ipa_find_reference): Update declaration.
14722 (ipa_clone_ref): Declare.
14723 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
14724 stream speculative flag.
14725 (lto_output_ref): Stream statements uids and speculation.
14726 (input_ref): Likewise.
14727 (input_edge): Stream speuclation.
14728 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
14729 (cgraph_set_call_stmt_including_clones): Handle speculation.
14730 * ipa-inline.c (heap_edge_removal_hook): New function.
14731 (inline_small_functions): Register it.
14732 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
14733 also initialize refs.
14734 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
14736 (try_make_edge_direct_simple_call): Likewise.
14737 (try_make_edge_direct_simple_call): Likewise.
14738 (update_indirect_edges_after_inlining): Likewise.
14739 (remove_described_reference): Look proper lto_stmt_uid.
14740 (propagate_controlled_uses): Likewise.
14741 (propagate_controlled_uses): Liekwise.
14742 * tree-inline.c (copy_bb): Copy speculative edges.
14743 (redirect_all_calls): New function.
14744 (copy_cfg_body): Do redirection after loop info is updated.
14745 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
14747 2013-08-09 Jan Hubicka <jh@suse.cz>
14749 * lto-streamer-out.c (output_function): Renumber PHIs.
14750 * lto-streamer-in.c (input_function): Likewise.
14752 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
14754 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
14755 (get_lane_unsigned): Likewise.
14756 (dup_lane_scalar): Likewise.
14757 (get_lane): enable for VALL.
14758 * config/aarch64/aarch64-simd.md
14759 (aarch64_dup_lane_scalar<mode>): Remove.
14760 (aarch64_get_lane_signed<mode>): Likewise.
14761 (aarch64_get_lane_unsigned<mode>): Likewise.
14762 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
14763 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
14764 (aarch64_get_lane<mode>): Enable for all vector modes.
14765 (aarch64_get_lanedi): Remove misleading constraints.
14766 * config/aarch64/arm_neon.h
14767 (__aarch64_vget_lane_any): Define.
14768 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
14769 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
14770 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
14771 * config/aarch64/iterators.md (VDQQH): New.
14772 (VDQQHS): Likewise.
14773 (vwcore): Likewise.
14775 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
14777 * configure.ac: Add GAS check for LEON instructions on SPARC.
14778 * configure: Regenerate.
14779 * config.in: Likewise.
14780 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
14782 * config/sparc/sparc.opt (LEON, LEON3): New masks.
14783 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
14785 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
14786 (AS_LEON_FLAG): New macro.
14787 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
14788 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
14789 Deal with LEON and LEON3 for the memory model.
14790 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
14791 (atomic_compare_and_swap<mode>_1): Likewise.
14792 (*atomic_compare_and_swap<mode>_1): Likewise.
14794 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14796 * config/arm/neon.md (vcond): Fix floating-point vector
14797 comparisons against 0.
14799 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
14801 * lra-constraints.c (emit_spill_move): Remove assert.
14802 (process_alt_operands): Add more debugging
14803 output. Increase reject for spilling into memory. Decrease
14804 reject for reloading scratch.
14805 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
14807 2013-08-08 Steve Ellcey <sellcey@mips.com>
14809 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
14810 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
14811 micromips incompatible. Add nan2008.
14812 (MULTILIB_DIRNAMES): Add nan2008.
14813 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14814 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
14815 and micromips incompatible. Add nan2008.
14816 (MULTILIB_DIRNAMES): Add nan2008.
14817 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14819 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
14821 PR rtl-optimization/58079
14822 * combine.c (combine_simplify_rtx): Avoid using SUBST if
14823 simplify_comparison has widened a comparison with an integer.
14825 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14827 * config/arm/neon.md (movmisalign<mode>): Disable when we
14828 don't allow unaligned accesses.
14829 (*movmisalign<mode>_neon_store): Likewise.
14830 (*movmisalign<mode>_neon_load): Likewise.
14831 (*movmisalign<mode>_neon_store): Likewise.
14832 (*movmisalign<mode>_neon_load): Likewise.
14834 2013-08-08 Jan Hubicka <jh@suse.cz>
14836 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
14837 (make_pass_rebuild_cgraph_edges): Also clear references.
14838 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
14839 * ipa-inline-transform.c (inline_transform): Remove all references
14841 * cgraphunit.c (expand_function): Remove all references after
14843 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
14844 (ipa_find_reference): Rewrite to iterator.
14845 (remove_stmt_references): Likewise.
14846 (ipa_clear_stmts_in_references): New function.
14847 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
14848 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
14850 * ipa-split.c (split_function): Remove references in split function.
14852 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14855 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
14856 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
14858 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14861 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
14862 suggested mode for the assignment isn't compatible with the
14863 registers required.
14865 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
14868 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
14870 2013-08-07 Xinliang David Li <davidxl@google.com>
14872 * config/i386/i386.opt: New option -mtune-ctrl=.
14873 * config/i386/x86-tune.def: New file.
14874 * config/i386/i386.h: include x86-tune.def.
14875 * config/i386/i386.c (ix86_option_override_internal):
14876 Parsing -mtune-ctrl= option and set tune features.
14878 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
14881 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
14882 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
14885 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14887 * rtl.h (update_alignments): Declare.
14888 * final.c (grow_label_align): New function extracted from...
14889 (shorten_branches): ...here. Call it.
14890 (update_alignments): New function.
14891 * reorg.c (sibling_labels): New variable.
14892 (get_label_before): Add SIBLING parameter. If it is non-zero, push
14893 the new label along with it onto the sibling_labels vector.
14894 (fill_simple_delay_slots): Adjust call to get_label_before.
14895 (fill_slots_from_thread): Likewise.
14896 (relax_delay_slots): Likewise.
14897 (make_return_insns): Likewise.
14898 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
14900 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14902 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
14903 document its semantics.
14904 (diagnostic_report_diagnostic): Adjust accordingly.
14906 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14908 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
14909 (sparc_option_override): ...and port to new C++ pass API.
14910 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
14912 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
14914 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
14916 2013-08-06 Caroline Tice <cmtice@google.com>
14918 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
14919 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
14920 * tree-pass.h: Add pass_vtable_verify.
14921 * varasm.c (assemble_variable): Add code to properly set the comdat
14922 section and name for the .vtable_map_vars section.
14923 (assemble_vtyv_preinit_initializer): New function.
14924 (default_sectin_type_flags): Make sure .vtable_map_vars section has
14926 * output.h: Add function decl for assemble_vtv_preinit_initializer.
14927 * vtable-verify.c: New file.
14928 * vtable-verify.h: New file.
14929 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
14930 initialiation levels.
14931 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
14932 * passes.def: Insert pass_vtable_verify.
14933 * aclocal.m4: Reorder includes.
14934 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
14935 -fvtv-counts options.
14936 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
14937 as appropriate, if -fvtable-verify=... is used.
14938 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
14939 -fvtable-verify=... is used.
14940 * Makefile.in (OBJS): Add vtable-verify.o to list.
14941 (vtable-verify.o): Add new build rule.
14942 (GTFILES): Add vtable-verify.c to list.
14943 * common.opt (fvtable-verify=): New flag.
14944 (vtv_priority): Values for fvtable-verify= flag.
14945 (fvtv-counts): New flag.
14946 (fvtv-debug): New flag.
14947 * tree.h (save_vtable_map_decl): New extern function decl.
14949 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14951 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
14952 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
14953 (pass_data_rl78_devirt): ...new pass_data instance and...
14954 (make_pass_rl78_devirt): ...new function.
14955 (rl78_asm_file_start): Port pass registration to new C++ API.
14957 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14959 * coretypes.h (rtl_opt_pass): Add.
14960 (gcc::context): Add.
14961 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
14962 (epiphany_init): Port to new C++ pass API.
14963 (epiphany_optimize_mode_switching): Likewise.
14964 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
14965 (pass_manager::get_pass_mode_switching): New.
14966 (pass_manager::get_pass_peephole2): New.
14967 * mode-switching.c (pass_mode_switching): Add clone method.
14968 * recog.c (pass_peephole2): Add clone method.
14969 (pass_split_all_insns): Add clone method.
14971 2013-08-06 David Malcolm <dmalcolm@redhat.com>
14973 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
14974 (mips_option_override): ...here, porting to new C++ API for passes.
14976 2013-08-06 Jan Hubicka <jh@suse.cz>
14978 * cgraph.c (cgraph_get_body): New function based on lto.c
14980 * cgraph.h (cgraph_get_body): Declare.
14981 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
14983 * cgraphunit.c (expand_function): Get body prior expanding.
14984 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
14985 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
14987 * passes.c (do_per_function_toporder): Get body.
14988 * tree-inline.c (expand_call_inline): Get body prior inlining it.
14989 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
14991 2013-08-06 Martin Jambor <mjambor@suse.cz>
14994 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
14995 is not re-finalized. Rename second parameter to no_collect.
14997 2013-08-06 Martin Jambor <mjambor@suse.cz>
14999 PR middle-end/58041
15000 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
15001 MEM_REF has proper alignment information.
15003 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
15006 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
15008 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
15009 rtx (*) (rtx, ...) with insn_gen_fn.
15010 * genoutput.c (output_insn_data): Cast gen_? function pointers to
15011 insn_gen_fn::stored_funcptr. Add initializer braces.
15013 2013-08-05 David Malcolm <dmalcolm@redhat.com>
15015 Rewrite how instances of passes are cloned to remove assumptions
15016 about their sizes (thus allowing pass subclasses to have
15017 additional data fields, albeit non-GC-managed ones at this point).
15019 * passes.c (make_pass_instance): Now that passes have clone
15020 methods, rewrite this function to eliminate XNEW and memcpy
15021 calls that used hardcoded sizes. Since this function no longer
15022 creates pass instances, rename it to...
15023 (add_pass_instance): ...this. Document the old way that passes were
15024 numbered and flagged, and rework this function to continue using it.
15025 (next_pass_1): Add an initial_pass argument for use by
15027 (position_pass): When adding multiple instances of a pass, use
15028 the pass's clone method, rather than relying on the XNEW/memcpy
15029 within the former make_pass_instance (now add_pass_instance).
15030 (pass_manager::pass_manager): When invoking next_pass_1, also supply
15031 the initial instance of the current pass within the pass manager.
15033 2013-08-05 David Malcolm <dmalcolm@redhat.com>
15035 This is the automated part of the conversion of passes from C
15036 structs to C++ classes.
15038 Patch autogenerated by refactor_passes.py from
15039 https://github.com/davidmalcolm/gcc-refactoring-scripts
15040 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
15042 * asan.c (pass_asan): Convert from a global struct to a subclass of
15043 gimple_opt_pass along with...
15044 (pass_data_asan): ...new pass_data instance and...
15045 (make_pass_asan): ...new function.
15046 (pass_asan_O0): Convert from a global struct to a subclass of
15047 gimple_opt_pass along with...
15048 (pass_data_asan_O0): ...new pass_data instance and...
15049 (make_pass_asan_O0): ...new function.
15050 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
15051 subclass of rtl_opt_pass along with...
15052 (pass_data_inc_dec): ...new pass_data instance and...
15053 (make_pass_inc_dec): ...new function.
15054 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
15055 a subclass of rtl_opt_pass along with...
15056 (pass_data_reorder_blocks): ...new pass_data instance and...
15057 (make_pass_reorder_blocks): ...new function.
15058 (pass_duplicate_computed_gotos): Convert from a global struct to a
15059 subclass of rtl_opt_pass along with...
15060 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
15061 (make_pass_duplicate_computed_gotos): ...new function.
15062 (pass_partition_blocks): Convert from a global struct to a subclass of
15063 rtl_opt_pass along with...
15064 (pass_data_partition_blocks): ...new pass_data instance and...
15065 (make_pass_partition_blocks): ...new function.
15066 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
15067 struct to a subclass of rtl_opt_pass along with...
15068 (pass_data_branch_target_load_optimize1): ...new pass_data instance
15070 (make_pass_branch_target_load_optimize1): ...new function.
15071 (pass_branch_target_load_optimize2): Convert from a global struct to a
15072 subclass of rtl_opt_pass along with...
15073 (pass_data_branch_target_load_optimize2): ...new pass_data instance
15075 (make_pass_branch_target_load_optimize2): ...new function.
15076 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
15077 of rtl_opt_pass along with...
15078 (pass_data_jump): ...new pass_data instance and...
15079 (make_pass_jump): ...new function.
15080 (pass_jump2): Convert from a global struct to a subclass of
15081 rtl_opt_pass along with...
15082 (pass_data_jump2): ...new pass_data instance and...
15083 (make_pass_jump2): ...new function.
15084 * cfgexpand.c (pass_expand): Convert from a global struct to a
15085 subclass of rtl_opt_pass along with...
15086 (pass_data_expand): ...new pass_data instance and...
15087 (make_pass_expand): ...new function.
15088 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
15089 of rtl_opt_pass along with...
15090 (pass_data_free_cfg): ...new pass_data instance and...
15091 (make_pass_free_cfg): ...new function.
15092 (pass_into_cfg_layout_mode): Convert from a global struct to a
15093 subclass of rtl_opt_pass along with...
15094 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
15095 (make_pass_into_cfg_layout_mode): ...new function.
15096 (pass_outof_cfg_layout_mode): Convert from a global struct to a
15097 subclass of rtl_opt_pass along with...
15098 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
15099 (make_pass_outof_cfg_layout_mode): ...new function.
15100 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
15101 struct to a subclass of gimple_opt_pass along with...
15102 (pass_data_build_cgraph_edges): ...new pass_data instance and...
15103 (make_pass_build_cgraph_edges): ...new function.
15104 (pass_rebuild_cgraph_edges): Convert from a global struct to a
15105 subclass of gimple_opt_pass along with...
15106 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
15107 (make_pass_rebuild_cgraph_edges): ...new function.
15108 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
15109 subclass of gimple_opt_pass along with...
15110 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
15112 (make_pass_remove_cgraph_callee_edges): ...new function.
15113 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
15114 struct to a subclass of rtl_opt_pass along with...
15115 (pass_data_stack_adjustments): ...new pass_data instance and...
15116 (make_pass_stack_adjustments): ...new function.
15117 * combine.c (pass_combine): Convert from a global struct to a subclass
15118 of rtl_opt_pass along with...
15119 (pass_data_combine): ...new pass_data instance and...
15120 (make_pass_combine): ...new function.
15121 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
15122 global struct to a subclass of rtl_opt_pass along with...
15123 (pass_data_compare_elim_after_reload): ...new pass_data instance
15125 (make_pass_compare_elim_after_reload): ...new function.
15126 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
15127 of rtl_opt_pass along with...
15128 (pass_data_rtl_cprop): ...new pass_data instance and...
15129 (make_pass_rtl_cprop): ...new function.
15130 * cse.c (pass_cse): Convert from a global struct to a subclass of
15131 rtl_opt_pass along with...
15132 (pass_data_cse): ...new pass_data instance and...
15133 (make_pass_cse): ...new function.
15134 (pass_cse2): Convert from a global struct to a subclass of
15135 rtl_opt_pass along with...
15136 (pass_data_cse2): ...new pass_data instance and...
15137 (make_pass_cse2): ...new function.
15138 (pass_cse_after_global_opts): Convert from a global struct to a
15139 subclass of rtl_opt_pass along with...
15140 (pass_data_cse_after_global_opts): ...new pass_data instance and...
15141 (make_pass_cse_after_global_opts): ...new function.
15142 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
15143 of rtl_opt_pass along with...
15144 (pass_data_ud_rtl_dce): ...new pass_data instance and...
15145 (make_pass_ud_rtl_dce): ...new function.
15146 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
15147 rtl_opt_pass along with...
15148 (pass_data_fast_rtl_dce): ...new pass_data instance and...
15149 (make_pass_fast_rtl_dce): ...new function.
15150 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
15151 a subclass of rtl_opt_pass along with...
15152 (pass_data_df_initialize_opt): ...new pass_data instance and...
15153 (make_pass_df_initialize_opt): ...new function.
15154 (pass_df_initialize_no_opt): Convert from a global struct to a
15155 subclass of rtl_opt_pass along with...
15156 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
15157 (make_pass_df_initialize_no_opt): ...new function.
15158 (pass_df_finish): Convert from a global struct to a subclass of
15159 rtl_opt_pass along with...
15160 (pass_data_df_finish): ...new pass_data instance and...
15161 (make_pass_df_finish): ...new function.
15162 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
15163 rtl_opt_pass along with...
15164 (pass_data_rtl_dse1): ...new pass_data instance and...
15165 (make_pass_rtl_dse1): ...new function.
15166 (pass_rtl_dse2): Convert from a global struct to a subclass of
15167 rtl_opt_pass along with...
15168 (pass_data_rtl_dse2): ...new pass_data instance and...
15169 (make_pass_rtl_dse2): ...new function.
15170 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
15171 subclass of rtl_opt_pass along with...
15172 (pass_data_dwarf2_frame): ...new pass_data instance and...
15173 (make_pass_dwarf2_frame): ...new function.
15174 * except.c (pass_set_nothrow_function_flags): Convert from a global
15175 struct to a subclass of rtl_opt_pass along with...
15176 (pass_data_set_nothrow_function_flags): ...new pass_data instance
15178 (make_pass_set_nothrow_function_flags): ...new function.
15179 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
15180 subclass of rtl_opt_pass along with...
15181 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
15183 (make_pass_convert_to_eh_region_ranges): ...new function.
15184 * final.c (pass_compute_alignments): Convert from a global struct to a
15185 subclass of rtl_opt_pass along with...
15186 (pass_data_compute_alignments): ...new pass_data instance and...
15187 (make_pass_compute_alignments): ...new function.
15188 (pass_final): Convert from a global struct to a subclass of
15189 rtl_opt_pass along with...
15190 (pass_data_final): ...new pass_data instance and...
15191 (make_pass_final): ...new function.
15192 (pass_shorten_branches): Convert from a global struct to a subclass of
15193 rtl_opt_pass along with...
15194 (pass_data_shorten_branches): ...new pass_data instance and...
15195 (make_pass_shorten_branches): ...new function.
15196 (pass_clean_state): Convert from a global struct to a subclass of
15197 rtl_opt_pass along with...
15198 (pass_data_clean_state): ...new pass_data instance and...
15199 (make_pass_clean_state): ...new function.
15200 * function.c (pass_instantiate_virtual_regs): Convert from a global
15201 struct to a subclass of rtl_opt_pass along with...
15202 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
15203 (make_pass_instantiate_virtual_regs): ...new function.
15204 (pass_leaf_regs): Convert from a global struct to a subclass of
15205 rtl_opt_pass along with...
15206 (pass_data_leaf_regs): ...new pass_data instance and...
15207 (make_pass_leaf_regs): ...new function.
15208 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
15209 subclass of rtl_opt_pass along with...
15210 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
15212 (make_pass_thread_prologue_and_epilogue): ...new function.
15213 (pass_match_asm_constraints): Convert from a global struct to a
15214 subclass of rtl_opt_pass along with...
15215 (pass_data_match_asm_constraints): ...new pass_data instance and...
15216 (make_pass_match_asm_constraints): ...new function.
15217 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
15218 subclass of rtl_opt_pass along with...
15219 (pass_data_rtl_fwprop): ...new pass_data instance and...
15220 (make_pass_rtl_fwprop): ...new function.
15221 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
15222 rtl_opt_pass along with...
15223 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
15224 (make_pass_rtl_fwprop_addr): ...new function.
15225 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
15226 rtl_opt_pass along with...
15227 (pass_data_rtl_pre): ...new pass_data instance and...
15228 (make_pass_rtl_pre): ...new function.
15229 (pass_rtl_hoist): Convert from a global struct to a subclass of
15230 rtl_opt_pass along with...
15231 (pass_data_rtl_hoist): ...new pass_data instance and...
15232 (make_pass_rtl_hoist): ...new function.
15233 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
15234 subclass of gimple_opt_pass along with...
15235 (pass_data_lower_cf): ...new pass_data instance and...
15236 (make_pass_lower_cf): ...new function.
15237 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
15238 from a global struct to a subclass of gimple_opt_pass along with...
15239 (pass_data_strength_reduction): ...new pass_data instance and...
15240 (make_pass_strength_reduction): ...new function.
15241 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
15242 of rtl_opt_pass along with...
15243 (pass_data_rtl_ifcvt): ...new pass_data instance and...
15244 (make_pass_rtl_ifcvt): ...new function.
15245 (pass_if_after_combine): Convert from a global struct to a subclass of
15246 rtl_opt_pass along with...
15247 (pass_data_if_after_combine): ...new pass_data instance and...
15248 (make_pass_if_after_combine): ...new function.
15249 (pass_if_after_reload): Convert from a global struct to a subclass of
15250 rtl_opt_pass along with...
15251 (pass_data_if_after_reload): ...new pass_data instance and...
15252 (make_pass_if_after_reload): ...new function.
15253 * init-regs.c (pass_initialize_regs): Convert from a global struct to
15254 a subclass of rtl_opt_pass along with...
15255 (pass_data_initialize_regs): ...new pass_data instance and...
15256 (make_pass_initialize_regs): ...new function.
15257 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
15258 of ipa_opt_pass_d along with...
15259 (pass_data_ipa_cp): ...new pass_data instance and...
15260 (make_pass_ipa_cp): ...new function.
15261 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
15262 global struct to a subclass of gimple_opt_pass along with...
15263 (pass_data_inline_parameters): ...new pass_data instance and...
15264 (make_pass_inline_parameters): ...new function.
15265 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
15266 subclass of gimple_opt_pass along with...
15267 (pass_data_early_inline): ...new pass_data instance and...
15268 (make_pass_early_inline): ...new function.
15269 (pass_ipa_inline): Convert from a global struct to a subclass of
15270 ipa_opt_pass_d along with...
15271 (pass_data_ipa_inline): ...new pass_data instance and...
15272 (make_pass_ipa_inline): ...new function.
15273 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
15274 struct to a subclass of gimple_opt_pass along with...
15275 (pass_data_local_pure_const): ...new pass_data instance and...
15276 (make_pass_local_pure_const): ...new function.
15277 (pass_ipa_pure_const): Convert from a global struct to a subclass of
15278 ipa_opt_pass_d along with...
15279 (pass_data_ipa_pure_const): ...new pass_data instance and...
15280 (make_pass_ipa_pure_const): ...new function.
15281 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
15282 to a subclass of ipa_opt_pass_d along with...
15283 (pass_data_ipa_reference): ...new pass_data instance and...
15284 (make_pass_ipa_reference): ...new function.
15285 * ipa-split.c (pass_split_functions): Convert from a global struct to
15286 a subclass of gimple_opt_pass along with...
15287 (pass_data_split_functions): ...new pass_data instance and...
15288 (make_pass_split_functions): ...new function.
15289 (pass_feedback_split_functions): Convert from a global struct to a
15290 subclass of gimple_opt_pass along with...
15291 (pass_data_feedback_split_functions): ...new pass_data instance and...
15292 (make_pass_feedback_split_functions): ...new function.
15293 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
15294 global struct to a subclass of simple_ipa_opt_pass along with...
15295 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
15297 (make_pass_ipa_function_and_variable_visibility): ...new function.
15298 (pass_ipa_free_inline_summary): Convert from a global struct to a
15299 subclass of simple_ipa_opt_pass along with...
15300 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
15301 (make_pass_ipa_free_inline_summary): ...new function.
15302 (pass_ipa_whole_program_visibility): Convert from a global struct to a
15303 subclass of ipa_opt_pass_d along with...
15304 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
15306 (make_pass_ipa_whole_program_visibility): ...new function.
15307 (pass_ipa_profile): Convert from a global struct to a subclass of
15308 ipa_opt_pass_d along with...
15309 (pass_data_ipa_profile): ...new pass_data instance and...
15310 (make_pass_ipa_profile): ...new function.
15311 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
15312 ipa_opt_pass_d along with...
15313 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
15314 (make_pass_ipa_cdtor_merge): ...new function.
15315 * ira.c (pass_ira): Convert from a global struct to a subclass of
15316 rtl_opt_pass along with...
15317 (pass_data_ira): ...new pass_data instance and...
15318 (make_pass_ira): ...new function.
15319 (pass_reload): Convert from a global struct to a subclass of
15320 rtl_opt_pass along with...
15321 (pass_data_reload): ...new pass_data instance and...
15322 (make_pass_reload): ...new function.
15323 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
15324 subclass of rtl_opt_pass along with...
15325 (pass_data_cleanup_barriers): ...new pass_data instance and...
15326 (make_pass_cleanup_barriers): ...new function.
15327 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
15328 of rtl_opt_pass along with...
15329 (pass_data_loop2): ...new pass_data instance and...
15330 (make_pass_loop2): ...new function.
15331 (pass_rtl_loop_init): Convert from a global struct to a subclass of
15332 rtl_opt_pass along with...
15333 (pass_data_rtl_loop_init): ...new pass_data instance and...
15334 (make_pass_rtl_loop_init): ...new function.
15335 (pass_rtl_loop_done): Convert from a global struct to a subclass of
15336 rtl_opt_pass along with...
15337 (pass_data_rtl_loop_done): ...new pass_data instance and...
15338 (make_pass_rtl_loop_done): ...new function.
15339 (pass_rtl_move_loop_invariants): Convert from a global struct to a
15340 subclass of rtl_opt_pass along with...
15341 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
15342 (make_pass_rtl_move_loop_invariants): ...new function.
15343 (pass_rtl_unswitch): Convert from a global struct to a subclass of
15344 rtl_opt_pass along with...
15345 (pass_data_rtl_unswitch): ...new pass_data instance and...
15346 (make_pass_rtl_unswitch): ...new function.
15347 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
15348 subclass of rtl_opt_pass along with...
15349 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
15351 (make_pass_rtl_unroll_and_peel_loops): ...new function.
15352 (pass_rtl_doloop): Convert from a global struct to a subclass of
15353 rtl_opt_pass along with...
15354 (pass_data_rtl_doloop): ...new pass_data instance and...
15355 (make_pass_rtl_doloop): ...new function.
15356 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
15357 a subclass of rtl_opt_pass along with...
15358 (pass_data_lower_subreg): ...new pass_data instance and...
15359 (make_pass_lower_subreg): ...new function.
15360 (pass_lower_subreg2): Convert from a global struct to a subclass of
15361 rtl_opt_pass along with...
15362 (pass_data_lower_subreg2): ...new pass_data instance and...
15363 (make_pass_lower_subreg2): ...new function.
15364 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
15365 struct to a subclass of ipa_opt_pass_d along with...
15366 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
15367 (make_pass_ipa_lto_gimple_out): ...new function.
15368 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
15369 of ipa_opt_pass_d along with...
15370 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
15371 (make_pass_ipa_lto_finish_out): ...new function.
15372 * mode-switching.c (pass_mode_switching): Convert from a global struct
15373 to a subclass of rtl_opt_pass along with...
15374 (pass_data_mode_switching): ...new pass_data instance and...
15375 (make_pass_mode_switching): ...new function.
15376 * modulo-sched.c (pass_sms): Convert from a global struct to a
15377 subclass of rtl_opt_pass along with...
15378 (pass_data_sms): ...new pass_data instance and...
15379 (make_pass_sms): ...new function.
15380 * omp-low.c (pass_expand_omp): Convert from a global struct to a
15381 subclass of gimple_opt_pass along with...
15382 (pass_data_expand_omp): ...new pass_data instance and...
15383 (make_pass_expand_omp): ...new function.
15384 (pass_lower_omp): Convert from a global struct to a subclass of
15385 gimple_opt_pass along with...
15386 (pass_data_lower_omp): ...new pass_data instance and...
15387 (make_pass_lower_omp): ...new function.
15388 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
15389 of gimple_opt_pass along with...
15390 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
15391 (make_pass_diagnose_omp_blocks): ...new function.
15392 * passes.c (pass_early_local_passes): Convert from a global struct to
15393 a subclass of simple_ipa_opt_pass along with...
15394 (pass_data_early_local_passes): ...new pass_data instance and...
15395 (make_pass_early_local_passes): ...new function.
15396 (pass_all_early_optimizations): Convert from a global struct to a
15397 subclass of gimple_opt_pass along with...
15398 (pass_data_all_early_optimizations): ...new pass_data instance and...
15399 (make_pass_all_early_optimizations): ...new function.
15400 (pass_all_optimizations): Convert from a global struct to a subclass
15401 of gimple_opt_pass along with...
15402 (pass_data_all_optimizations): ...new pass_data instance and...
15403 (make_pass_all_optimizations): ...new function.
15404 (pass_all_optimizations_g): Convert from a global struct to a subclass
15405 of gimple_opt_pass along with...
15406 (pass_data_all_optimizations_g): ...new pass_data instance and...
15407 (make_pass_all_optimizations_g): ...new function.
15408 (pass_rest_of_compilation): Convert from a global struct to a subclass
15409 of rtl_opt_pass along with...
15410 (pass_data_rest_of_compilation): ...new pass_data instance and...
15411 (make_pass_rest_of_compilation): ...new function.
15412 (pass_postreload): Convert from a global struct to a subclass of
15413 rtl_opt_pass along with...
15414 (pass_data_postreload): ...new pass_data instance and...
15415 (make_pass_postreload): ...new function.
15416 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
15417 subclass of rtl_opt_pass along with...
15418 (pass_data_gcse2): ...new pass_data instance and...
15419 (make_pass_gcse2): ...new function.
15420 * postreload.c (pass_postreload_cse): Convert from a global struct to
15421 a subclass of rtl_opt_pass along with...
15422 (pass_data_postreload_cse): ...new pass_data instance and...
15423 (make_pass_postreload_cse): ...new function.
15424 * predict.c (pass_profile): Convert from a global struct to a subclass
15425 of gimple_opt_pass along with...
15426 (pass_data_profile): ...new pass_data instance and...
15427 (make_pass_profile): ...new function.
15428 (pass_strip_predict_hints): Convert from a global struct to a subclass
15429 of gimple_opt_pass along with...
15430 (pass_data_strip_predict_hints): ...new pass_data instance and...
15431 (make_pass_strip_predict_hints): ...new function.
15432 * recog.c (pass_peephole2): Convert from a global struct to a subclass
15433 of rtl_opt_pass along with...
15434 (pass_data_peephole2): ...new pass_data instance and...
15435 (make_pass_peephole2): ...new function.
15436 (pass_split_all_insns): Convert from a global struct to a subclass of
15437 rtl_opt_pass along with...
15438 (pass_data_split_all_insns): ...new pass_data instance and...
15439 (make_pass_split_all_insns): ...new function.
15440 (pass_split_after_reload): Convert from a global struct to a subclass
15441 of rtl_opt_pass along with...
15442 (pass_data_split_after_reload): ...new pass_data instance and...
15443 (make_pass_split_after_reload): ...new function.
15444 (pass_split_before_regstack): Convert from a global struct to a
15445 subclass of rtl_opt_pass along with...
15446 (pass_data_split_before_regstack): ...new pass_data instance and...
15447 (make_pass_split_before_regstack): ...new function.
15448 (pass_split_before_sched2): Convert from a global struct to a subclass
15449 of rtl_opt_pass along with...
15450 (pass_data_split_before_sched2): ...new pass_data instance and...
15451 (make_pass_split_before_sched2): ...new function.
15452 (pass_split_for_shorten_branches): Convert from a global struct to a
15453 subclass of rtl_opt_pass along with...
15454 (pass_data_split_for_shorten_branches): ...new pass_data instance
15456 (make_pass_split_for_shorten_branches): ...new function.
15457 * ree.c (pass_ree): Convert from a global struct to a subclass of
15458 rtl_opt_pass along with...
15459 (pass_data_ree): ...new pass_data instance and...
15460 (make_pass_ree): ...new function.
15461 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
15462 subclass of rtl_opt_pass along with...
15463 (pass_data_stack_regs): ...new pass_data instance and...
15464 (make_pass_stack_regs): ...new function.
15465 (pass_stack_regs_run): Convert from a global struct to a subclass of
15466 rtl_opt_pass along with...
15467 (pass_data_stack_regs_run): ...new pass_data instance and...
15468 (make_pass_stack_regs_run): ...new function.
15469 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
15470 subclass of rtl_opt_pass along with...
15471 (pass_data_cprop_hardreg): ...new pass_data instance and...
15472 (make_pass_cprop_hardreg): ...new function.
15473 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
15474 subclass of rtl_opt_pass along with...
15475 (pass_data_reginfo_init): ...new pass_data instance and...
15476 (make_pass_reginfo_init): ...new function.
15477 * regmove.c (pass_regmove): Convert from a global struct to a subclass
15478 of rtl_opt_pass along with...
15479 (pass_data_regmove): ...new pass_data instance and...
15480 (make_pass_regmove): ...new function.
15481 * regrename.c (pass_regrename): Convert from a global struct to a
15482 subclass of rtl_opt_pass along with...
15483 (pass_data_regrename): ...new pass_data instance and...
15484 (make_pass_regrename): ...new function.
15485 * reorg.c (pass_delay_slots): Convert from a global struct to a
15486 subclass of rtl_opt_pass along with...
15487 (pass_data_delay_slots): ...new pass_data instance and...
15488 (make_pass_delay_slots): ...new function.
15489 (pass_machine_reorg): Convert from a global struct to a subclass of
15490 rtl_opt_pass along with...
15491 (pass_data_machine_reorg): ...new pass_data instance and...
15492 (make_pass_machine_reorg): ...new function.
15493 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
15494 of rtl_opt_pass along with...
15495 (pass_data_sched): ...new pass_data instance and...
15496 (make_pass_sched): ...new function.
15497 (pass_sched2): Convert from a global struct to a subclass of
15498 rtl_opt_pass along with...
15499 (pass_data_sched2): ...new pass_data instance and...
15500 (make_pass_sched2): ...new function.
15501 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
15502 to a subclass of rtl_opt_pass along with...
15503 (pass_data_stack_ptr_mod): ...new pass_data instance and...
15504 (make_pass_stack_ptr_mod): ...new function.
15505 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
15506 to a subclass of rtl_opt_pass along with...
15507 (pass_data_rtl_store_motion): ...new pass_data instance and...
15508 (make_pass_rtl_store_motion): ...new function.
15509 * tracer.c (pass_tracer): Convert from a global struct to a subclass
15510 of gimple_opt_pass along with...
15511 (pass_data_tracer): ...new pass_data instance and...
15512 (make_pass_tracer): ...new function.
15513 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
15514 to a subclass of gimple_opt_pass along with...
15515 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
15516 (make_pass_diagnose_tm_blocks): ...new function.
15517 (pass_lower_tm): Convert from a global struct to a subclass of
15518 gimple_opt_pass along with...
15519 (pass_data_lower_tm): ...new pass_data instance and...
15520 (make_pass_lower_tm): ...new function.
15521 (pass_tm_init): Convert from a global struct to a subclass of
15522 gimple_opt_pass along with...
15523 (pass_data_tm_init): ...new pass_data instance and...
15524 (make_pass_tm_init): ...new function.
15525 (pass_tm_mark): Convert from a global struct to a subclass of
15526 gimple_opt_pass along with...
15527 (pass_data_tm_mark): ...new pass_data instance and...
15528 (make_pass_tm_mark): ...new function.
15529 (pass_tm_edges): Convert from a global struct to a subclass of
15530 gimple_opt_pass along with...
15531 (pass_data_tm_edges): ...new pass_data instance and...
15532 (make_pass_tm_edges): ...new function.
15533 (pass_tm_memopt): Convert from a global struct to a subclass of
15534 gimple_opt_pass along with...
15535 (pass_data_tm_memopt): ...new pass_data instance and...
15536 (make_pass_tm_memopt): ...new function.
15537 (pass_ipa_tm): Convert from a global struct to a subclass of
15538 simple_ipa_opt_pass along with...
15539 (pass_data_ipa_tm): ...new pass_data instance and...
15540 (make_pass_ipa_tm): ...new function.
15541 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
15542 subclass of gimple_opt_pass along with...
15543 (pass_data_call_cdce): ...new pass_data instance and...
15544 (make_pass_call_cdce): ...new function.
15545 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
15546 subclass of gimple_opt_pass along with...
15547 (pass_data_build_cfg): ...new pass_data instance and...
15548 (make_pass_build_cfg): ...new function.
15549 (pass_split_crit_edges): Convert from a global struct to a subclass of
15550 gimple_opt_pass along with...
15551 (pass_data_split_crit_edges): ...new pass_data instance and...
15552 (make_pass_split_crit_edges): ...new function.
15553 (pass_warn_function_return): Convert from a global struct to a
15554 subclass of gimple_opt_pass along with...
15555 (pass_data_warn_function_return): ...new pass_data instance and...
15556 (make_pass_warn_function_return): ...new function.
15557 (pass_warn_function_noreturn): Convert from a global struct to a
15558 subclass of gimple_opt_pass along with...
15559 (pass_data_warn_function_noreturn): ...new pass_data instance and...
15560 (make_pass_warn_function_noreturn): ...new function.
15561 (pass_warn_unused_result): Convert from a global struct to a subclass
15562 of gimple_opt_pass along with...
15563 (pass_data_warn_unused_result): ...new pass_data instance and...
15564 (make_pass_warn_unused_result): ...new function.
15565 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
15566 a subclass of gimple_opt_pass along with...
15567 (pass_data_merge_phi): ...new pass_data instance and...
15568 (make_pass_merge_phi): ...new function.
15569 * tree-complex.c (pass_lower_complex): Convert from a global struct to
15570 a subclass of gimple_opt_pass along with...
15571 (pass_data_lower_complex): ...new pass_data instance and...
15572 (make_pass_lower_complex): ...new function.
15573 (pass_lower_complex_O0): Convert from a global struct to a subclass of
15574 gimple_opt_pass along with...
15575 (pass_data_lower_complex_O0): ...new pass_data instance and...
15576 (make_pass_lower_complex_O0): ...new function.
15577 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
15578 subclass of gimple_opt_pass along with...
15579 (pass_data_lower_eh): ...new pass_data instance and...
15580 (make_pass_lower_eh): ...new function.
15581 (pass_refactor_eh): Convert from a global struct to a subclass of
15582 gimple_opt_pass along with...
15583 (pass_data_refactor_eh): ...new pass_data instance and...
15584 (make_pass_refactor_eh): ...new function.
15585 (pass_lower_resx): Convert from a global struct to a subclass of
15586 gimple_opt_pass along with...
15587 (pass_data_lower_resx): ...new pass_data instance and...
15588 (make_pass_lower_resx): ...new function.
15589 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
15590 of gimple_opt_pass along with...
15591 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
15592 (make_pass_lower_eh_dispatch): ...new function.
15593 (pass_cleanup_eh): Convert from a global struct to a subclass of
15594 gimple_opt_pass along with...
15595 (pass_data_cleanup_eh): ...new pass_data instance and...
15596 (make_pass_cleanup_eh): ...new function.
15597 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
15598 to a subclass of simple_ipa_opt_pass along with...
15599 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
15600 (make_pass_ipa_lower_emutls): ...new function.
15601 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
15602 a subclass of gimple_opt_pass along with...
15603 (pass_data_if_conversion): ...new pass_data instance and...
15604 (make_pass_if_conversion): ...new function.
15605 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
15606 subclass of gimple_opt_pass along with...
15607 (pass_data_build_ssa): ...new pass_data instance and...
15608 (make_pass_build_ssa): ...new function.
15609 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
15610 global struct to a subclass of gimple_opt_pass along with...
15611 (pass_data_loop_distribution): ...new pass_data instance and...
15612 (make_pass_loop_distribution): ...new function.
15613 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
15614 subclass of gimple_opt_pass along with...
15615 (pass_data_mudflap_1): ...new pass_data instance and...
15616 (make_pass_mudflap_1): ...new function.
15617 (pass_mudflap_2): Convert from a global struct to a subclass of
15618 gimple_opt_pass along with...
15619 (pass_data_mudflap_2): ...new pass_data instance and...
15620 (make_pass_mudflap_2): ...new function.
15621 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
15622 subclass of gimple_opt_pass along with...
15623 (pass_data_mudflap_1): ...new pass_data instance and...
15624 (make_pass_mudflap_1): ...new function.
15625 (pass_mudflap_2): Convert from a global struct to a subclass of
15626 gimple_opt_pass along with...
15627 (pass_data_mudflap_2): ...new pass_data instance and...
15628 (make_pass_mudflap_2): ...new function.
15629 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
15630 gimple_opt_pass along with...
15631 (pass_data_nrv): ...new pass_data instance and...
15632 (make_pass_nrv): ...new function.
15633 (pass_return_slot): Convert from a global struct to a subclass of
15634 gimple_opt_pass along with...
15635 (pass_data_return_slot): ...new pass_data instance and...
15636 (make_pass_return_slot): ...new function.
15637 * tree-object-size.c (pass_object_sizes): Convert from a global struct
15638 to a subclass of gimple_opt_pass along with...
15639 (pass_data_object_sizes): ...new pass_data instance and...
15640 (make_pass_object_sizes): ...new function.
15641 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
15642 global struct to a subclass of gimple_opt_pass along with...
15643 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
15645 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15646 (pass_fixup_cfg): Convert from a global struct to a subclass of
15647 gimple_opt_pass along with...
15648 (pass_data_fixup_cfg): ...new pass_data instance and...
15649 (make_pass_fixup_cfg): ...new function.
15650 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
15651 (make_pass_mudflap_1): ...new function.
15652 (pass_mudflap_2): Replace declaration with that of...
15653 (make_pass_mudflap_2): ...new function.
15654 (pass_asan): Replace declaration with that of...
15655 (make_pass_asan): ...new function.
15656 (pass_asan_O0): Replace declaration with that of...
15657 (make_pass_asan_O0): ...new function.
15658 (pass_tsan): Replace declaration with that of...
15659 (make_pass_tsan): ...new function.
15660 (pass_tsan_O0): Replace declaration with that of...
15661 (make_pass_tsan_O0): ...new function.
15662 (pass_lower_cf): Replace declaration with that of...
15663 (make_pass_lower_cf): ...new function.
15664 (pass_refactor_eh): Replace declaration with that of...
15665 (make_pass_refactor_eh): ...new function.
15666 (pass_lower_eh): Replace declaration with that of...
15667 (make_pass_lower_eh): ...new function.
15668 (pass_lower_eh_dispatch): Replace declaration with that of...
15669 (make_pass_lower_eh_dispatch): ...new function.
15670 (pass_lower_resx): Replace declaration with that of...
15671 (make_pass_lower_resx): ...new function.
15672 (pass_build_cfg): Replace declaration with that of...
15673 (make_pass_build_cfg): ...new function.
15674 (pass_early_tree_profile): Replace declaration with that of...
15675 (make_pass_early_tree_profile): ...new function.
15676 (pass_cleanup_eh): Replace declaration with that of...
15677 (make_pass_cleanup_eh): ...new function.
15678 (pass_sra): Replace declaration with that of...
15679 (make_pass_sra): ...new function.
15680 (pass_sra_early): Replace declaration with that of...
15681 (make_pass_sra_early): ...new function.
15682 (pass_early_ipa_sra): Replace declaration with that of...
15683 (make_pass_early_ipa_sra): ...new function.
15684 (pass_tail_recursion): Replace declaration with that of...
15685 (make_pass_tail_recursion): ...new function.
15686 (pass_tail_calls): Replace declaration with that of...
15687 (make_pass_tail_calls): ...new function.
15688 (pass_tree_loop): Replace declaration with that of...
15689 (make_pass_tree_loop): ...new function.
15690 (pass_tree_loop_init): Replace declaration with that of...
15691 (make_pass_tree_loop_init): ...new function.
15692 (pass_lim): Replace declaration with that of...
15693 (make_pass_lim): ...new function.
15694 (pass_tree_unswitch): Replace declaration with that of...
15695 (make_pass_tree_unswitch): ...new function.
15696 (pass_predcom): Replace declaration with that of...
15697 (make_pass_predcom): ...new function.
15698 (pass_iv_canon): Replace declaration with that of...
15699 (make_pass_iv_canon): ...new function.
15700 (pass_scev_cprop): Replace declaration with that of...
15701 (make_pass_scev_cprop): ...new function.
15702 (pass_empty_loop): Replace declaration with that of...
15703 (make_pass_empty_loop): ...new function.
15704 (pass_record_bounds): Replace declaration with that of...
15705 (make_pass_record_bounds): ...new function.
15706 (pass_graphite): Replace declaration with that of...
15707 (make_pass_graphite): ...new function.
15708 (pass_graphite_transforms): Replace declaration with that of...
15709 (make_pass_graphite_transforms): ...new function.
15710 (pass_if_conversion): Replace declaration with that of...
15711 (make_pass_if_conversion): ...new function.
15712 (pass_loop_distribution): Replace declaration with that of...
15713 (make_pass_loop_distribution): ...new function.
15714 (pass_vectorize): Replace declaration with that of...
15715 (make_pass_vectorize): ...new function.
15716 (pass_slp_vectorize): Replace declaration with that of...
15717 (make_pass_slp_vectorize): ...new function.
15718 (pass_complete_unroll): Replace declaration with that of...
15719 (make_pass_complete_unroll): ...new function.
15720 (pass_complete_unrolli): Replace declaration with that of...
15721 (make_pass_complete_unrolli): ...new function.
15722 (pass_parallelize_loops): Replace declaration with that of...
15723 (make_pass_parallelize_loops): ...new function.
15724 (pass_loop_prefetch): Replace declaration with that of...
15725 (make_pass_loop_prefetch): ...new function.
15726 (pass_iv_optimize): Replace declaration with that of...
15727 (make_pass_iv_optimize): ...new function.
15728 (pass_tree_loop_done): Replace declaration with that of...
15729 (make_pass_tree_loop_done): ...new function.
15730 (pass_ch): Replace declaration with that of...
15731 (make_pass_ch): ...new function.
15732 (pass_ccp): Replace declaration with that of...
15733 (make_pass_ccp): ...new function.
15734 (pass_phi_only_cprop): Replace declaration with that of...
15735 (make_pass_phi_only_cprop): ...new function.
15736 (pass_build_ssa): Replace declaration with that of...
15737 (make_pass_build_ssa): ...new function.
15738 (pass_build_alias): Replace declaration with that of...
15739 (make_pass_build_alias): ...new function.
15740 (pass_build_ealias): Replace declaration with that of...
15741 (make_pass_build_ealias): ...new function.
15742 (pass_dominator): Replace declaration with that of...
15743 (make_pass_dominator): ...new function.
15744 (pass_dce): Replace declaration with that of...
15745 (make_pass_dce): ...new function.
15746 (pass_dce_loop): Replace declaration with that of...
15747 (make_pass_dce_loop): ...new function.
15748 (pass_cd_dce): Replace declaration with that of...
15749 (make_pass_cd_dce): ...new function.
15750 (pass_call_cdce): Replace declaration with that of...
15751 (make_pass_call_cdce): ...new function.
15752 (pass_merge_phi): Replace declaration with that of...
15753 (make_pass_merge_phi): ...new function.
15754 (pass_split_crit_edges): Replace declaration with that of...
15755 (make_pass_split_crit_edges): ...new function.
15756 (pass_pre): Replace declaration with that of...
15757 (make_pass_pre): ...new function.
15758 (pass_profile): Replace declaration with that of...
15759 (make_pass_profile): ...new function.
15760 (pass_strip_predict_hints): Replace declaration with that of...
15761 (make_pass_strip_predict_hints): ...new function.
15762 (pass_lower_complex_O0): Replace declaration with that of...
15763 (make_pass_lower_complex_O0): ...new function.
15764 (pass_lower_complex): Replace declaration with that of...
15765 (make_pass_lower_complex): ...new function.
15766 (pass_lower_vector): Replace declaration with that of...
15767 (make_pass_lower_vector): ...new function.
15768 (pass_lower_vector_ssa): Replace declaration with that of...
15769 (make_pass_lower_vector_ssa): ...new function.
15770 (pass_lower_omp): Replace declaration with that of...
15771 (make_pass_lower_omp): ...new function.
15772 (pass_diagnose_omp_blocks): Replace declaration with that of...
15773 (make_pass_diagnose_omp_blocks): ...new function.
15774 (pass_expand_omp): Replace declaration with that of...
15775 (make_pass_expand_omp): ...new function.
15776 (pass_expand_omp_ssa): Replace declaration with that of...
15777 (make_pass_expand_omp_ssa): ...new function.
15778 (pass_object_sizes): Replace declaration with that of...
15779 (make_pass_object_sizes): ...new function.
15780 (pass_strlen): Replace declaration with that of...
15781 (make_pass_strlen): ...new function.
15782 (pass_fold_builtins): Replace declaration with that of...
15783 (make_pass_fold_builtins): ...new function.
15784 (pass_stdarg): Replace declaration with that of...
15785 (make_pass_stdarg): ...new function.
15786 (pass_early_warn_uninitialized): Replace declaration with that of...
15787 (make_pass_early_warn_uninitialized): ...new function.
15788 (pass_late_warn_uninitialized): Replace declaration with that of...
15789 (make_pass_late_warn_uninitialized): ...new function.
15790 (pass_cse_reciprocals): Replace declaration with that of...
15791 (make_pass_cse_reciprocals): ...new function.
15792 (pass_cse_sincos): Replace declaration with that of...
15793 (make_pass_cse_sincos): ...new function.
15794 (pass_optimize_bswap): Replace declaration with that of...
15795 (make_pass_optimize_bswap): ...new function.
15796 (pass_optimize_widening_mul): Replace declaration with that of...
15797 (make_pass_optimize_widening_mul): ...new function.
15798 (pass_warn_function_return): Replace declaration with that of...
15799 (make_pass_warn_function_return): ...new function.
15800 (pass_warn_function_noreturn): Replace declaration with that of...
15801 (make_pass_warn_function_noreturn): ...new function.
15802 (pass_cselim): Replace declaration with that of...
15803 (make_pass_cselim): ...new function.
15804 (pass_phiopt): Replace declaration with that of...
15805 (make_pass_phiopt): ...new function.
15806 (pass_forwprop): Replace declaration with that of...
15807 (make_pass_forwprop): ...new function.
15808 (pass_phiprop): Replace declaration with that of...
15809 (make_pass_phiprop): ...new function.
15810 (pass_tree_ifcombine): Replace declaration with that of...
15811 (make_pass_tree_ifcombine): ...new function.
15812 (pass_dse): Replace declaration with that of...
15813 (make_pass_dse): ...new function.
15814 (pass_nrv): Replace declaration with that of...
15815 (make_pass_nrv): ...new function.
15816 (pass_rename_ssa_copies): Replace declaration with that of...
15817 (make_pass_rename_ssa_copies): ...new function.
15818 (pass_sink_code): Replace declaration with that of...
15819 (make_pass_sink_code): ...new function.
15820 (pass_fre): Replace declaration with that of...
15821 (make_pass_fre): ...new function.
15822 (pass_check_data_deps): Replace declaration with that of...
15823 (make_pass_check_data_deps): ...new function.
15824 (pass_copy_prop): Replace declaration with that of...
15825 (make_pass_copy_prop): ...new function.
15826 (pass_vrp): Replace declaration with that of...
15827 (make_pass_vrp): ...new function.
15828 (pass_uncprop): Replace declaration with that of...
15829 (make_pass_uncprop): ...new function.
15830 (pass_return_slot): Replace declaration with that of...
15831 (make_pass_return_slot): ...new function.
15832 (pass_reassoc): Replace declaration with that of...
15833 (make_pass_reassoc): ...new function.
15834 (pass_rebuild_cgraph_edges): Replace declaration with that of...
15835 (make_pass_rebuild_cgraph_edges): ...new function.
15836 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
15837 (make_pass_remove_cgraph_callee_edges): ...new function.
15838 (pass_build_cgraph_edges): Replace declaration with that of...
15839 (make_pass_build_cgraph_edges): ...new function.
15840 (pass_local_pure_const): Replace declaration with that of...
15841 (make_pass_local_pure_const): ...new function.
15842 (pass_tracer): Replace declaration with that of...
15843 (make_pass_tracer): ...new function.
15844 (pass_warn_unused_result): Replace declaration with that of...
15845 (make_pass_warn_unused_result): ...new function.
15846 (pass_diagnose_tm_blocks): Replace declaration with that of...
15847 (make_pass_diagnose_tm_blocks): ...new function.
15848 (pass_lower_tm): Replace declaration with that of...
15849 (make_pass_lower_tm): ...new function.
15850 (pass_tm_init): Replace declaration with that of...
15851 (make_pass_tm_init): ...new function.
15852 (pass_tm_mark): Replace declaration with that of...
15853 (make_pass_tm_mark): ...new function.
15854 (pass_tm_memopt): Replace declaration with that of...
15855 (make_pass_tm_memopt): ...new function.
15856 (pass_tm_edges): Replace declaration with that of...
15857 (make_pass_tm_edges): ...new function.
15858 (pass_split_functions): Replace declaration with that of...
15859 (make_pass_split_functions): ...new function.
15860 (pass_feedback_split_functions): Replace declaration with that of...
15861 (make_pass_feedback_split_functions): ...new function.
15862 (pass_strength_reduction): Replace declaration with that of...
15863 (make_pass_strength_reduction): ...new function.
15864 (pass_ipa_lower_emutls): Replace declaration with that of...
15865 (make_pass_ipa_lower_emutls): ...new function.
15866 (pass_ipa_function_and_variable_visibility): Replace declaration with
15868 (make_pass_ipa_function_and_variable_visibility): ...new function.
15869 (pass_ipa_tree_profile): Replace declaration with that of...
15870 (make_pass_ipa_tree_profile): ...new function.
15871 (pass_early_local_passes): Replace declaration with that of...
15872 (make_pass_early_local_passes): ...new function.
15873 (pass_ipa_whole_program_visibility): Replace declaration with that
15875 (make_pass_ipa_whole_program_visibility): ...new function.
15876 (pass_ipa_lto_gimple_out): Replace declaration with that of...
15877 (make_pass_ipa_lto_gimple_out): ...new function.
15878 (pass_ipa_increase_alignment): Replace declaration with that of...
15879 (make_pass_ipa_increase_alignment): ...new function.
15880 (pass_ipa_inline): Replace declaration with that of...
15881 (make_pass_ipa_inline): ...new function.
15882 (pass_ipa_free_lang_data): Replace declaration with that of...
15883 (make_pass_ipa_free_lang_data): ...new function.
15884 (pass_ipa_free_inline_summary): Replace declaration with that of...
15885 (make_pass_ipa_free_inline_summary): ...new function.
15886 (pass_ipa_cp): Replace declaration with that of...
15887 (make_pass_ipa_cp): ...new function.
15888 (pass_ipa_reference): Replace declaration with that of...
15889 (make_pass_ipa_reference): ...new function.
15890 (pass_ipa_pure_const): Replace declaration with that of...
15891 (make_pass_ipa_pure_const): ...new function.
15892 (pass_ipa_pta): Replace declaration with that of...
15893 (make_pass_ipa_pta): ...new function.
15894 (pass_ipa_lto_finish_out): Replace declaration with that of...
15895 (make_pass_ipa_lto_finish_out): ...new function.
15896 (pass_ipa_tm): Replace declaration with that of...
15897 (make_pass_ipa_tm): ...new function.
15898 (pass_ipa_profile): Replace declaration with that of...
15899 (make_pass_ipa_profile): ...new function.
15900 (pass_ipa_cdtor_merge): Replace declaration with that of...
15901 (make_pass_ipa_cdtor_merge): ...new function.
15902 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
15904 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15905 (pass_init_datastructures): Replace declaration with that of...
15906 (make_pass_init_datastructures): ...new function.
15907 (pass_fixup_cfg): Replace declaration with that of...
15908 (make_pass_fixup_cfg): ...new function.
15909 (pass_expand): Replace declaration with that of...
15910 (make_pass_expand): ...new function.
15911 (pass_instantiate_virtual_regs): Replace declaration with that of...
15912 (make_pass_instantiate_virtual_regs): ...new function.
15913 (pass_rtl_fwprop): Replace declaration with that of...
15914 (make_pass_rtl_fwprop): ...new function.
15915 (pass_rtl_fwprop_addr): Replace declaration with that of...
15916 (make_pass_rtl_fwprop_addr): ...new function.
15917 (pass_jump): Replace declaration with that of...
15918 (make_pass_jump): ...new function.
15919 (pass_jump2): Replace declaration with that of...
15920 (make_pass_jump2): ...new function.
15921 (pass_lower_subreg): Replace declaration with that of...
15922 (make_pass_lower_subreg): ...new function.
15923 (pass_cse): Replace declaration with that of...
15924 (make_pass_cse): ...new function.
15925 (pass_fast_rtl_dce): Replace declaration with that of...
15926 (make_pass_fast_rtl_dce): ...new function.
15927 (pass_ud_rtl_dce): Replace declaration with that of...
15928 (make_pass_ud_rtl_dce): ...new function.
15929 (pass_rtl_dce): Replace declaration with that of...
15930 (make_pass_rtl_dce): ...new function.
15931 (pass_rtl_dse1): Replace declaration with that of...
15932 (make_pass_rtl_dse1): ...new function.
15933 (pass_rtl_dse2): Replace declaration with that of...
15934 (make_pass_rtl_dse2): ...new function.
15935 (pass_rtl_dse3): Replace declaration with that of...
15936 (make_pass_rtl_dse3): ...new function.
15937 (pass_rtl_cprop): Replace declaration with that of...
15938 (make_pass_rtl_cprop): ...new function.
15939 (pass_rtl_pre): Replace declaration with that of...
15940 (make_pass_rtl_pre): ...new function.
15941 (pass_rtl_hoist): Replace declaration with that of...
15942 (make_pass_rtl_hoist): ...new function.
15943 (pass_rtl_store_motion): Replace declaration with that of...
15944 (make_pass_rtl_store_motion): ...new function.
15945 (pass_cse_after_global_opts): Replace declaration with that of...
15946 (make_pass_cse_after_global_opts): ...new function.
15947 (pass_rtl_ifcvt): Replace declaration with that of...
15948 (make_pass_rtl_ifcvt): ...new function.
15949 (pass_into_cfg_layout_mode): Replace declaration with that of...
15950 (make_pass_into_cfg_layout_mode): ...new function.
15951 (pass_outof_cfg_layout_mode): Replace declaration with that of...
15952 (make_pass_outof_cfg_layout_mode): ...new function.
15953 (pass_loop2): Replace declaration with that of...
15954 (make_pass_loop2): ...new function.
15955 (pass_rtl_loop_init): Replace declaration with that of...
15956 (make_pass_rtl_loop_init): ...new function.
15957 (pass_rtl_move_loop_invariants): Replace declaration with that of...
15958 (make_pass_rtl_move_loop_invariants): ...new function.
15959 (pass_rtl_unswitch): Replace declaration with that of...
15960 (make_pass_rtl_unswitch): ...new function.
15961 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
15962 (make_pass_rtl_unroll_and_peel_loops): ...new function.
15963 (pass_rtl_doloop): Replace declaration with that of...
15964 (make_pass_rtl_doloop): ...new function.
15965 (pass_rtl_loop_done): Replace declaration with that of...
15966 (make_pass_rtl_loop_done): ...new function.
15967 (pass_web): Replace declaration with that of...
15968 (make_pass_web): ...new function.
15969 (pass_cse2): Replace declaration with that of...
15970 (make_pass_cse2): ...new function.
15971 (pass_df_initialize_opt): Replace declaration with that of...
15972 (make_pass_df_initialize_opt): ...new function.
15973 (pass_df_initialize_no_opt): Replace declaration with that of...
15974 (make_pass_df_initialize_no_opt): ...new function.
15975 (pass_reginfo_init): Replace declaration with that of...
15976 (make_pass_reginfo_init): ...new function.
15977 (pass_inc_dec): Replace declaration with that of...
15978 (make_pass_inc_dec): ...new function.
15979 (pass_stack_ptr_mod): Replace declaration with that of...
15980 (make_pass_stack_ptr_mod): ...new function.
15981 (pass_initialize_regs): Replace declaration with that of...
15982 (make_pass_initialize_regs): ...new function.
15983 (pass_combine): Replace declaration with that of...
15984 (make_pass_combine): ...new function.
15985 (pass_if_after_combine): Replace declaration with that of...
15986 (make_pass_if_after_combine): ...new function.
15987 (pass_ree): Replace declaration with that of...
15988 (make_pass_ree): ...new function.
15989 (pass_partition_blocks): Replace declaration with that of...
15990 (make_pass_partition_blocks): ...new function.
15991 (pass_match_asm_constraints): Replace declaration with that of...
15992 (make_pass_match_asm_constraints): ...new function.
15993 (pass_regmove): Replace declaration with that of...
15994 (make_pass_regmove): ...new function.
15995 (pass_split_all_insns): Replace declaration with that of...
15996 (make_pass_split_all_insns): ...new function.
15997 (pass_fast_rtl_byte_dce): Replace declaration with that of...
15998 (make_pass_fast_rtl_byte_dce): ...new function.
15999 (pass_lower_subreg2): Replace declaration with that of...
16000 (make_pass_lower_subreg2): ...new function.
16001 (pass_mode_switching): Replace declaration with that of...
16002 (make_pass_mode_switching): ...new function.
16003 (pass_sms): Replace declaration with that of...
16004 (make_pass_sms): ...new function.
16005 (pass_sched): Replace declaration with that of...
16006 (make_pass_sched): ...new function.
16007 (pass_ira): Replace declaration with that of...
16008 (make_pass_ira): ...new function.
16009 (pass_reload): Replace declaration with that of...
16010 (make_pass_reload): ...new function.
16011 (pass_clean_state): Replace declaration with that of...
16012 (make_pass_clean_state): ...new function.
16013 (pass_branch_prob): Replace declaration with that of...
16014 (make_pass_branch_prob): ...new function.
16015 (pass_value_profile_transformations): Replace declaration with that
16017 (make_pass_value_profile_transformations): ...new function.
16018 (pass_postreload_cse): Replace declaration with that of...
16019 (make_pass_postreload_cse): ...new function.
16020 (pass_gcse2): Replace declaration with that of...
16021 (make_pass_gcse2): ...new function.
16022 (pass_split_after_reload): Replace declaration with that of...
16023 (make_pass_split_after_reload): ...new function.
16024 (pass_branch_target_load_optimize1): Replace declaration with that
16026 (make_pass_branch_target_load_optimize1): ...new function.
16027 (pass_thread_prologue_and_epilogue): Replace declaration with that
16029 (make_pass_thread_prologue_and_epilogue): ...new function.
16030 (pass_stack_adjustments): Replace declaration with that of...
16031 (make_pass_stack_adjustments): ...new function.
16032 (pass_peephole2): Replace declaration with that of...
16033 (make_pass_peephole2): ...new function.
16034 (pass_if_after_reload): Replace declaration with that of...
16035 (make_pass_if_after_reload): ...new function.
16036 (pass_regrename): Replace declaration with that of...
16037 (make_pass_regrename): ...new function.
16038 (pass_cprop_hardreg): Replace declaration with that of...
16039 (make_pass_cprop_hardreg): ...new function.
16040 (pass_reorder_blocks): Replace declaration with that of...
16041 (make_pass_reorder_blocks): ...new function.
16042 (pass_branch_target_load_optimize2): Replace declaration with that
16044 (make_pass_branch_target_load_optimize2): ...new function.
16045 (pass_leaf_regs): Replace declaration with that of...
16046 (make_pass_leaf_regs): ...new function.
16047 (pass_split_before_sched2): Replace declaration with that of...
16048 (make_pass_split_before_sched2): ...new function.
16049 (pass_compare_elim_after_reload): Replace declaration with that of...
16050 (make_pass_compare_elim_after_reload): ...new function.
16051 (pass_sched2): Replace declaration with that of...
16052 (make_pass_sched2): ...new function.
16053 (pass_stack_regs): Replace declaration with that of...
16054 (make_pass_stack_regs): ...new function.
16055 (pass_stack_regs_run): Replace declaration with that of...
16056 (make_pass_stack_regs_run): ...new function.
16057 (pass_df_finish): Replace declaration with that of...
16058 (make_pass_df_finish): ...new function.
16059 (pass_compute_alignments): Replace declaration with that of...
16060 (make_pass_compute_alignments): ...new function.
16061 (pass_duplicate_computed_gotos): Replace declaration with that of...
16062 (make_pass_duplicate_computed_gotos): ...new function.
16063 (pass_variable_tracking): Replace declaration with that of...
16064 (make_pass_variable_tracking): ...new function.
16065 (pass_free_cfg): Replace declaration with that of...
16066 (make_pass_free_cfg): ...new function.
16067 (pass_machine_reorg): Replace declaration with that of...
16068 (make_pass_machine_reorg): ...new function.
16069 (pass_cleanup_barriers): Replace declaration with that of...
16070 (make_pass_cleanup_barriers): ...new function.
16071 (pass_delay_slots): Replace declaration with that of...
16072 (make_pass_delay_slots): ...new function.
16073 (pass_split_for_shorten_branches): Replace declaration with that of...
16074 (make_pass_split_for_shorten_branches): ...new function.
16075 (pass_split_before_regstack): Replace declaration with that of...
16076 (make_pass_split_before_regstack): ...new function.
16077 (pass_convert_to_eh_region_ranges): Replace declaration with that
16079 (make_pass_convert_to_eh_region_ranges): ...new function.
16080 (pass_shorten_branches): Replace declaration with that of...
16081 (make_pass_shorten_branches): ...new function.
16082 (pass_set_nothrow_function_flags): Replace declaration with that of...
16083 (make_pass_set_nothrow_function_flags): ...new function.
16084 (pass_dwarf2_frame): Replace declaration with that of...
16085 (make_pass_dwarf2_frame): ...new function.
16086 (pass_final): Replace declaration with that of...
16087 (make_pass_final): ...new function.
16088 (pass_rtl_seqabstr): Replace declaration with that of...
16089 (make_pass_rtl_seqabstr): ...new function.
16090 (pass_release_ssa_names): Replace declaration with that of...
16091 (make_pass_release_ssa_names): ...new function.
16092 (pass_early_inline): Replace declaration with that of...
16093 (make_pass_early_inline): ...new function.
16094 (pass_inline_parameters): Replace declaration with that of...
16095 (make_pass_inline_parameters): ...new function.
16096 (pass_update_address_taken): Replace declaration with that of...
16097 (make_pass_update_address_taken): ...new function.
16098 (pass_convert_switch): Replace declaration with that of...
16099 (make_pass_convert_switch): ...new function.
16100 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
16101 to a subclass of simple_ipa_opt_pass along with...
16102 (pass_data_ipa_tree_profile): ...new pass_data instance and...
16103 (make_pass_ipa_tree_profile): ...new function.
16104 * tree-sra.c (pass_sra_early): Convert from a global struct to a
16105 subclass of gimple_opt_pass along with...
16106 (pass_data_sra_early): ...new pass_data instance and...
16107 (make_pass_sra_early): ...new function.
16108 (pass_sra): Convert from a global struct to a subclass of
16109 gimple_opt_pass along with...
16110 (pass_data_sra): ...new pass_data instance and...
16111 (make_pass_sra): ...new function.
16112 (pass_early_ipa_sra): Convert from a global struct to a subclass of
16113 gimple_opt_pass along with...
16114 (pass_data_early_ipa_sra): ...new pass_data instance and...
16115 (make_pass_early_ipa_sra): ...new function.
16116 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
16117 subclass of gimple_opt_pass along with...
16118 (pass_data_ccp): ...new pass_data instance and...
16119 (make_pass_ccp): ...new function.
16120 (pass_fold_builtins): Convert from a global struct to a subclass of
16121 gimple_opt_pass along with...
16122 (pass_data_fold_builtins): ...new pass_data instance and...
16123 (make_pass_fold_builtins): ...new function.
16124 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
16125 subclass of gimple_opt_pass along with...
16126 (pass_data_copy_prop): ...new pass_data instance and...
16127 (make_pass_copy_prop): ...new function.
16128 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
16129 global struct to a subclass of gimple_opt_pass along with...
16130 (pass_data_rename_ssa_copies): ...new pass_data instance and...
16131 (make_pass_rename_ssa_copies): ...new function.
16132 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
16133 subclass of gimple_opt_pass along with...
16134 (pass_data_dce): ...new pass_data instance and...
16135 (make_pass_dce): ...new function.
16136 (pass_dce_loop): Convert from a global struct to a subclass of
16137 gimple_opt_pass along with...
16138 (pass_data_dce_loop): ...new pass_data instance and...
16139 (make_pass_dce_loop): ...new function.
16140 (pass_cd_dce): Convert from a global struct to a subclass of
16141 gimple_opt_pass along with...
16142 (pass_data_cd_dce): ...new pass_data instance and...
16143 (make_pass_cd_dce): ...new function.
16144 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
16145 subclass of gimple_opt_pass along with...
16146 (pass_data_dominator): ...new pass_data instance and...
16147 (make_pass_dominator): ...new function.
16148 (pass_phi_only_cprop): Convert from a global struct to a subclass of
16149 gimple_opt_pass along with...
16150 (pass_data_phi_only_cprop): ...new pass_data instance and...
16151 (make_pass_phi_only_cprop): ...new function.
16152 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
16153 subclass of gimple_opt_pass along with...
16154 (pass_data_dse): ...new pass_data instance and...
16155 (make_pass_dse): ...new function.
16156 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
16157 a subclass of gimple_opt_pass along with...
16158 (pass_data_forwprop): ...new pass_data instance and...
16159 (make_pass_forwprop): ...new function.
16160 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
16161 struct to a subclass of gimple_opt_pass along with...
16162 (pass_data_tree_ifcombine): ...new pass_data instance and...
16163 (make_pass_tree_ifcombine): ...new function.
16164 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
16165 subclass of gimple_opt_pass along with...
16166 (pass_data_ch): ...new pass_data instance and...
16167 (make_pass_ch): ...new function.
16168 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
16169 subclass of gimple_opt_pass along with...
16170 (pass_data_tree_loop): ...new pass_data instance and...
16171 (make_pass_tree_loop): ...new function.
16172 (pass_tree_loop_init): Convert from a global struct to a subclass of
16173 gimple_opt_pass along with...
16174 (pass_data_tree_loop_init): ...new pass_data instance and...
16175 (make_pass_tree_loop_init): ...new function.
16176 (pass_lim): Convert from a global struct to a subclass of
16177 gimple_opt_pass along with...
16178 (pass_data_lim): ...new pass_data instance and...
16179 (make_pass_lim): ...new function.
16180 (pass_tree_unswitch): Convert from a global struct to a subclass of
16181 gimple_opt_pass along with...
16182 (pass_data_tree_unswitch): ...new pass_data instance and...
16183 (make_pass_tree_unswitch): ...new function.
16184 (pass_predcom): Convert from a global struct to a subclass of
16185 gimple_opt_pass along with...
16186 (pass_data_predcom): ...new pass_data instance and...
16187 (make_pass_predcom): ...new function.
16188 (pass_vectorize): Convert from a global struct to a subclass of
16189 gimple_opt_pass along with...
16190 (pass_data_vectorize): ...new pass_data instance and...
16191 (make_pass_vectorize): ...new function.
16192 (pass_graphite): Convert from a global struct to a subclass of
16193 gimple_opt_pass along with...
16194 (pass_data_graphite): ...new pass_data instance and...
16195 (make_pass_graphite): ...new function.
16196 (pass_graphite_transforms): Convert from a global struct to a subclass
16197 of gimple_opt_pass along with...
16198 (pass_data_graphite_transforms): ...new pass_data instance and...
16199 (make_pass_graphite_transforms): ...new function.
16200 (pass_check_data_deps): Convert from a global struct to a subclass of
16201 gimple_opt_pass along with...
16202 (pass_data_check_data_deps): ...new pass_data instance and...
16203 (make_pass_check_data_deps): ...new function.
16204 (pass_iv_canon): Convert from a global struct to a subclass of
16205 gimple_opt_pass along with...
16206 (pass_data_iv_canon): ...new pass_data instance and...
16207 (make_pass_iv_canon): ...new function.
16208 (pass_scev_cprop): Convert from a global struct to a subclass of
16209 gimple_opt_pass along with...
16210 (pass_data_scev_cprop): ...new pass_data instance and...
16211 (make_pass_scev_cprop): ...new function.
16212 (pass_record_bounds): Convert from a global struct to a subclass of
16213 gimple_opt_pass along with...
16214 (pass_data_record_bounds): ...new pass_data instance and...
16215 (make_pass_record_bounds): ...new function.
16216 (pass_complete_unroll): Convert from a global struct to a subclass of
16217 gimple_opt_pass along with...
16218 (pass_data_complete_unroll): ...new pass_data instance and...
16219 (make_pass_complete_unroll): ...new function.
16220 (pass_complete_unrolli): Convert from a global struct to a subclass of
16221 gimple_opt_pass along with...
16222 (pass_data_complete_unrolli): ...new pass_data instance and...
16223 (make_pass_complete_unrolli): ...new function.
16224 (pass_parallelize_loops): Convert from a global struct to a subclass
16225 of gimple_opt_pass along with...
16226 (pass_data_parallelize_loops): ...new pass_data instance and...
16227 (make_pass_parallelize_loops): ...new function.
16228 (pass_loop_prefetch): Convert from a global struct to a subclass of
16229 gimple_opt_pass along with...
16230 (pass_data_loop_prefetch): ...new pass_data instance and...
16231 (make_pass_loop_prefetch): ...new function.
16232 (pass_iv_optimize): Convert from a global struct to a subclass of
16233 gimple_opt_pass along with...
16234 (pass_data_iv_optimize): ...new pass_data instance and...
16235 (make_pass_iv_optimize): ...new function.
16236 (pass_tree_loop_done): Convert from a global struct to a subclass of
16237 gimple_opt_pass along with...
16238 (pass_data_tree_loop_done): ...new pass_data instance and...
16239 (make_pass_tree_loop_done): ...new function.
16240 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
16241 struct to a subclass of gimple_opt_pass along with...
16242 (pass_data_cse_reciprocals): ...new pass_data instance and...
16243 (make_pass_cse_reciprocals): ...new function.
16244 (pass_cse_sincos): Convert from a global struct to a subclass of
16245 gimple_opt_pass along with...
16246 (pass_data_cse_sincos): ...new pass_data instance and...
16247 (make_pass_cse_sincos): ...new function.
16248 (pass_optimize_bswap): Convert from a global struct to a subclass of
16249 gimple_opt_pass along with...
16250 (pass_data_optimize_bswap): ...new pass_data instance and...
16251 (make_pass_optimize_bswap): ...new function.
16252 (pass_optimize_widening_mul): Convert from a global struct to a
16253 subclass of gimple_opt_pass along with...
16254 (pass_data_optimize_widening_mul): ...new pass_data instance and...
16255 (make_pass_optimize_widening_mul): ...new function.
16256 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
16257 subclass of gimple_opt_pass along with...
16258 (pass_data_phiopt): ...new pass_data instance and...
16259 (make_pass_phiopt): ...new function.
16260 (pass_cselim): Convert from a global struct to a subclass of
16261 gimple_opt_pass along with...
16262 (pass_data_cselim): ...new pass_data instance and...
16263 (make_pass_cselim): ...new function.
16264 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
16265 subclass of gimple_opt_pass along with...
16266 (pass_data_phiprop): ...new pass_data instance and...
16267 (make_pass_phiprop): ...new function.
16268 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
16269 subclass of gimple_opt_pass along with...
16270 (pass_data_pre): ...new pass_data instance and...
16271 (make_pass_pre): ...new function.
16272 (pass_fre): Convert from a global struct to a subclass of
16273 gimple_opt_pass along with...
16274 (pass_data_fre): ...new pass_data instance and...
16275 (make_pass_fre): ...new function.
16276 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
16277 subclass of gimple_opt_pass along with...
16278 (pass_data_reassoc): ...new pass_data instance and...
16279 (make_pass_reassoc): ...new function.
16280 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
16281 subclass of gimple_opt_pass along with...
16282 (pass_data_sink_code): ...new pass_data instance and...
16283 (make_pass_sink_code): ...new function.
16284 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
16285 subclass of gimple_opt_pass along with...
16286 (pass_data_strlen): ...new pass_data instance and...
16287 (make_pass_strlen): ...new function.
16288 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
16289 struct to a subclass of gimple_opt_pass along with...
16290 (pass_data_build_alias): ...new pass_data instance and...
16291 (make_pass_build_alias): ...new function.
16292 (pass_build_ealias): Convert from a global struct to a subclass of
16293 gimple_opt_pass along with...
16294 (pass_data_build_ealias): ...new pass_data instance and...
16295 (make_pass_build_ealias): ...new function.
16296 (pass_ipa_pta): Convert from a global struct to a subclass of
16297 simple_ipa_opt_pass along with...
16298 (pass_data_ipa_pta): ...new pass_data instance and...
16299 (make_pass_ipa_pta): ...new function.
16300 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
16301 subclass of gimple_opt_pass along with...
16302 (pass_data_uncprop): ...new pass_data instance and...
16303 (make_pass_uncprop): ...new function.
16304 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
16305 global struct to a subclass of gimple_opt_pass along with...
16306 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
16307 (make_pass_late_warn_uninitialized): ...new function.
16308 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
16309 to a subclass of gimple_opt_pass along with...
16310 (pass_data_init_datastructures): ...new pass_data instance and...
16311 (make_pass_init_datastructures): ...new function.
16312 (pass_early_warn_uninitialized): Convert from a global struct to a
16313 subclass of gimple_opt_pass along with...
16314 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
16315 (make_pass_early_warn_uninitialized): ...new function.
16316 (pass_update_address_taken): Convert from a global struct to a
16317 subclass of gimple_opt_pass along with...
16318 (pass_data_update_address_taken): ...new pass_data instance and...
16319 (make_pass_update_address_taken): ...new function.
16320 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
16321 struct to a subclass of gimple_opt_pass along with...
16322 (pass_data_release_ssa_names): ...new pass_data instance and...
16323 (make_pass_release_ssa_names): ...new function.
16324 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
16325 subclass of gimple_opt_pass along with...
16326 (pass_data_stdarg): ...new pass_data instance and...
16327 (make_pass_stdarg): ...new function.
16328 * tree-switch-conversion.c (pass_convert_switch): Convert from a
16329 global struct to a subclass of gimple_opt_pass along with...
16330 (pass_data_convert_switch): ...new pass_data instance and...
16331 (make_pass_convert_switch): ...new function.
16332 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
16333 to a subclass of gimple_opt_pass along with...
16334 (pass_data_tail_recursion): ...new pass_data instance and...
16335 (make_pass_tail_recursion): ...new function.
16336 (pass_tail_calls): Convert from a global struct to a subclass of
16337 gimple_opt_pass along with...
16338 (pass_data_tail_calls): ...new pass_data instance and...
16339 (make_pass_tail_calls): ...new function.
16340 * tree-vect-generic.c (pass_lower_vector): Convert from a global
16341 struct to a subclass of gimple_opt_pass along with...
16342 (pass_data_lower_vector): ...new pass_data instance and...
16343 (make_pass_lower_vector): ...new function.
16344 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
16345 gimple_opt_pass along with...
16346 (pass_data_lower_vector_ssa): ...new pass_data instance and...
16347 (make_pass_lower_vector_ssa): ...new function.
16348 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
16349 to a subclass of gimple_opt_pass along with...
16350 (pass_data_slp_vectorize): ...new pass_data instance and...
16351 (make_pass_slp_vectorize): ...new function.
16352 (pass_ipa_increase_alignment): Convert from a global struct to a
16353 subclass of simple_ipa_opt_pass along with...
16354 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
16355 (make_pass_ipa_increase_alignment): ...new function.
16356 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
16357 gimple_opt_pass along with...
16358 (pass_data_vrp): ...new pass_data instance and...
16359 (make_pass_vrp): ...new function.
16360 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
16361 subclass of simple_ipa_opt_pass along with...
16362 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
16363 (make_pass_ipa_free_lang_data): ...new function.
16364 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
16365 gimple_opt_pass along with...
16366 (pass_data_tsan): ...new pass_data instance and...
16367 (make_pass_tsan): ...new function.
16368 (pass_tsan_O0): Convert from a global struct to a subclass of
16369 gimple_opt_pass along with...
16370 (pass_data_tsan_O0): ...new pass_data instance and...
16371 (make_pass_tsan_O0): ...new function.
16372 * var-tracking.c (pass_variable_tracking): Convert from a global
16373 struct to a subclass of rtl_opt_pass along with...
16374 (pass_data_variable_tracking): ...new pass_data instance and...
16375 (make_pass_variable_tracking): ...new function.
16376 * web.c (pass_web): Convert from a global struct to a subclass of
16377 rtl_opt_pass along with...
16378 (pass_data_web): ...new pass_data instance and...
16379 (make_pass_web): ...new function.
16380 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
16381 declaration with that of...
16382 (make_pass_mode_switch_use): ...new function.
16383 (pass_resolve_sw_modes): Replace declaration with that of...
16384 (make_pass_resolve_sw_modes): ...new function.
16385 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
16386 from a global struct to a subclass of rtl_opt_pass along with...
16387 (pass_data_mode_switch_use): ...new pass_data instance and...
16388 (make_pass_mode_switch_use): ...new function.
16389 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
16390 from a global struct to a subclass of rtl_opt_pass along with...
16391 (pass_data_resolve_sw_modes): ...new pass_data instance and...
16392 (make_pass_resolve_sw_modes): ...new function.
16393 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
16394 struct to a subclass of rtl_opt_pass along with...
16395 (pass_data_insert_vzeroupper): ...new pass_data instance and...
16396 (make_pass_insert_vzeroupper): ...new function.
16397 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
16398 global struct to a subclass of rtl_opt_pass along with...
16399 (pass_data_work_around_errata): ...new pass_data instance and...
16400 (make_pass_work_around_errata): ...new function.
16401 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
16402 struct to a subclass of rtl_opt_pass along with...
16403 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
16404 (make_pass_mips_machine_reorg2): ...new function.
16406 2013-08-05 David Malcolm <dmalcolm@redhat.com>
16408 * passes.c (pass_manager::operator new): New.
16410 2013-08-05 David Malcolm <dmalcolm@redhat.com>
16412 Handwritten part of conversion of passes to C++ classes.
16414 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
16415 (toplev.o): Add dep on PASS_MANAGER_H.
16416 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
16417 of early local pases to reflect this moving from a global to a
16418 member of gcc::pass_manager.
16419 (cgraph_add_new_function): Likewise.
16420 * lto-cgraph.c (lto_output_node): Update for conversion of
16421 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
16422 * passes.c (opt_pass::clone): New.
16423 (opt_pass::gate): New.
16424 (opt_pass::execute): New.
16425 (opt_pass::opt_pass): New.
16426 (pass_manager::execute_early_local_passes): New.
16427 (pass_manager::execute_pass_mode_switching): new.
16428 (finish_optimization_passes): Convert to...
16429 (pass_manager::finish_optimization_passes): ...this.
16430 (finish_optimization_passes): Update for conversion of passes to
16432 (register_dump_files_1): Use has_gate since we cannot portably
16433 check a vtable entry against NULL.
16434 (dump_one_pass): Likewise.
16435 (ipa_write_summaries_2): Likewise.
16436 (ipa_write_optimization_summaries_1): Likewise.
16437 (ipa_read_summaries_1): Likewise.
16438 (ipa_read_optimization_summaries_1): Likewise.
16439 (execute_ipa_stmt_fixups): Likewise.
16440 (pass_manager::pass_manager): Rewrite pass-creation, invoking
16441 pass-creation functions rather than wiring up globals, and
16442 storing the results in fields of pass_manager generated using
16443 pass-instances.def.
16444 (pass_manager::dump_profile_report): Update for conversion of
16445 passes to C++ classes.
16446 (pass_manager::execute_ipa_summary_passes): Likewise.
16447 (execute_one_ipa_transform_pass): Likewise.
16448 (execute_one_pass): Use has_gate and has_execute since we cannot
16449 portably check a vtable entry against NULL.
16450 * pass_manager.h (pass_manager::finish_optimization_passes): New.
16451 (pass_manager): Use pass-instances.def to add fields for the
16452 various pass instances.
16453 * toplev.c (finalize): Update for move of
16454 finish_optimization_passes to a method of gcc::pass_manager.
16455 * toplev.h (finish_optimization_passes): Move to method of class
16457 * tree-pass.h (struct pass_data): New.
16458 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
16459 (opt_pass::gate): Convert to virtual function.
16460 (opt_pass::~opt_pass): New.
16461 (opt_pass::clone): New.
16462 (opt_pass::execute): Convert to virtual function.
16463 (opt_pass::opt_pass): New.
16464 (opt_pass::ctxt_): new.
16465 (gimple_opt_pass): Convert to subclass of opt_pass.
16466 (gimple_opt_pass::gimple_opt_pass): New.
16467 (rtl_opt_pass): Convert to subclass of opt_pass.
16468 (rtl_opt_pass::rtl_opt_pass): New.
16469 (ipa_opt_pass_d): Convert to subclass of opt_pass.
16470 (ipa_opt_pass_d::ipa_opt_pass_d): New.
16471 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
16472 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
16473 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
16474 invocation of pass_mode_switching to reflect this moving from a
16475 global to a member of gcc::pass_manager.
16476 (ix86_option_override): Rework how pass_insert_vzeroupper is
16477 added to the pass_manager to reflect autogenerated changes.
16478 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
16480 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
16482 PR rtl-optimization/57708
16483 * recog.c (peep2_find_free_register): Validate all regs in a
16486 2013-08-05 Jan Hubicka <jh@suse.cz>
16489 * cgraph.c (verify_cgraph_node): Accept local flags from other
16491 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
16492 (function_and_variable_visibility): Likewise.
16493 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
16495 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16497 * graph.c (init_graph_slim_pretty_print): Remove.
16498 (print_graph_cfg): Do not call it. Use local pretty printer.
16499 (start_graph_dump): Likewise.
16501 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16503 * gimple-pretty-print.c (buffer): Remove.
16504 (initialized): Likewise.
16505 (maybe_init_pretty_print): Likewise.
16506 (print_gimple_stmt): Do not call it. Use non-static local
16507 pretty_printer variable.
16508 (print_gimple_expr): Likewise.
16509 (print_gimple_seq): Likewise.
16510 (gimple_dump_bb): Likewise.
16512 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16514 * asan.c (asan_pp): Remove.
16515 (asan_pp_initialized): Likewise.
16516 (asan_pp_initialize): Likewise.
16517 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
16518 (asan_emit_stack_protection): Tidy. Use local pretty printer.
16519 (asan_add_global): Likewise.
16521 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
16523 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
16524 * diagnostic.h (diagnostic_flush_buffer): Adjust.
16525 * pretty-print.c (pp_formatted_text_data): Likewise.
16526 (pp_indent): Rename from pp_base_indent.
16527 (pp_format): Rename from pp_base_format.
16528 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
16529 (pp_format_verbatim): Rename from pp_base_format_verbatim.
16530 (pp_flush): Rename from pp_base_flush.
16531 (pp_set_line_maximum_length): Rename from
16532 pp_base_set_line_maximum_length.
16533 (pp_clear_output_area): Rename from pp_base_clear_output_area.
16534 (pp_set_prefix): Rename from pp_base_set_prefix.
16535 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
16536 (pp_emit_prefix): Rename from pp_base_emit_prefix.
16537 (pp_append_text): Rename from pp_base_append_text.
16538 (pp_formatted_text): Rename from pp_base_formatted_text.
16539 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
16540 (pp_remaining_character_count_for_line): Rename from
16541 pp_base_remaining_character_count_for_line.
16542 (pp_newline): Rename from pp_base_newline.
16543 (pp_character): Rename from pp_base_character.
16544 (pp_string): Rename from pp_base_string.
16545 (pp_maybe_space): Rename from pp_base_maybe_space.
16546 * asan.c (asan_pp_string): Adjust.
16547 (asan_emit_stack_protection): Likewise.
16548 (asan_add_global): Likewise.
16549 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
16550 * tree-mudflap.c (mf_varname_tree): Likewise.
16551 * tree-pretty-print.c (pp_tree_identifier): Rename from
16552 pp_base_tree_identifier.
16553 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
16554 Declare as function.
16556 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
16558 * pretty-print.h (pp_bar_bar): New.
16559 (pp_ampersand_ampersand): Likewise.
16560 (pp_less_equal): Likewise.
16561 (pp_greater_equal): Likewise.
16562 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
16563 printer functions instead of pp_string or operators and punctuators.
16564 (dump_gimple_call): Likewise.
16565 (dump_gimple_omp_for): Likewise.
16566 (dump_gimple_transaction): Likewise.
16567 (dump_gimple_phi): Likewise.
16568 (pp_gimple_stmt_1): Likewise.
16569 * sched-vis.c (print_insn): Likewise.
16570 * tree-mudflap.c (mf_varname_tree): Likewise.
16571 * tree-pretty-print.c (dump_block_node): Likewise.
16572 (dump_generic_node): Likewise.
16574 2013-08-02 Jan Hubicka <jh@suse.cz>
16576 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
16578 * lto-streamer-out.c (tree_is_indexable): Results decls and
16579 parm decls are not indexable.
16580 (DFS_write_tree_body): Do not follow args and results.
16581 (hash_tree): Likewise.
16582 (output_functions): Rearrange so struct function is needed
16583 only when real body is output; be able to also ouptut abstract
16584 functions; output DECL_ARGUMENTS and DECL_RESULT.
16585 (lto_output): When not in WPA, ale store abstract functions.
16586 (write_symbol): Do not care about RESULT_DECL.
16587 (output_symbol_p): Handle correctly sbtract decls.
16588 * lto-streamer-in.c (input_function): Rearrange so struct
16589 function can be NULL at entry; allow streaming of
16590 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
16591 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
16592 sanity check during LTO.
16593 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
16594 RESULT_DECl and DECL_ARGUMENTS.
16595 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
16598 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
16600 * pretty-print.h (pp_underscore): New.
16602 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
16603 printer functions instead of pp_character.
16604 (dump_binary_rhs): Likewise.
16605 (dump_ternary_rhs): Likewise.
16606 (dump_gimple_call_args): Likewise.
16607 (pp_points_to_solution): Likewise.
16608 (dump_gimple_call): Likewise.
16609 (dump_gimple_switch): Likewise.
16610 (dump_gimple_cond): Likewise.
16611 (dump_gimple_bind): Likewise.
16612 (dump_gimple_try): Likewise.
16613 (dump_gimple_omp_for): Likewise.
16614 (dump_gimple_omp_continue): Likewise.
16615 (dump_gimple_omp_single): Likewise.
16616 (dump_gimple_omp_sections): Likewise.
16617 (dump_gimple_omp_block): Likewise.
16618 (dump_gimple_omp_critical): Likewise.
16619 (dump_gimple_transaction): Likewise.
16620 (dump_gimple_asm): Likewise.
16621 (dump_gimple_phi): Likewise.
16622 (dump_gimple_omp_parallel): Likewise.
16623 (dump_gimple_omp_task): Likewise.
16624 (dump_gimple_omp_atomic_load): Likewise.
16625 (dump_gimple_omp_atomic_store): Likewise.
16626 (dump_gimple_mem_ops): Likewise.
16627 (pp_gimple_stmt_1): Likewise.
16628 (pp_cfg_jump): Likewise.
16629 (dump_implicit_edges): Likewise.
16630 (gimple_dump_bb_for_graph): Likewise.
16631 * graph.c (draw_cfg_node): Likewise.
16632 * langhooks.c (lhd_print_error_function): Likewise.
16633 * sched-vis.c (print_exp): Likewise.
16634 (print_value): Likewise.
16635 (print_pattern): Likewise.
16636 (print_insn): Likewise.
16637 (rtl_dump_bb_for_graph): Likewise.
16638 * tree-pretty-print.c (dump_function_declaration): Likewise.
16639 (dump_array_domain): Likewise.
16640 (dump_omp_clause): Likewise.
16641 (dump_location): Likewise.
16642 (dump_generic_node): Likewise.
16643 (print_struct_decl): Likewise.
16644 * diagnostic.c (diagnostic_show_locus): Use pp_space.
16646 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16648 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
16649 candidate table when replacing a candidate statement.
16650 (replace_rhs_if_not_dup): Likewise.
16651 (replace_one_candidate): Likewise.
16653 2013-08-02 Jan Hubicka <jh@suse.cz>
16654 Martin Liska <marxin.liska@gmail.com>
16656 * cgraphunit.c (add_new_function): Fix logic when adding from
16658 (assemble_thunk): Rename to ...
16659 (expand_thunk); .. this one; export; get it working with
16660 general functions; make produced gimple valid.
16661 * cgraph.h (expand_thunk): Declare.
16663 2013-08-02 Jan Hubicka <jh@suse.cz>
16665 * ipa-cp.c (gather_context_independent_values): Use
16666 ipa_get_param_move_cost.
16667 (get_replacement_map): Remove PARAM; move parameter folding
16669 (create_specialized_node): Update.
16670 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
16671 assert that we have gimple body; update move_cost.
16672 (count_formal_params): Assert that we have gimple body.
16673 (ipa_dump_param): New function.
16674 (ipa_alloc_node_params): Break out from ...
16675 (ipa_initialize_node_params): ... here.
16676 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
16677 (ipa_write_node_info): Stream move costs.
16678 (ipa_read_node_info): Read move costs.
16679 (ipa_update_after_lto_read): Do not recompute node params.
16680 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
16681 (ipa_get_param): Check we are not in WPA.
16682 (ipa_get_param_move_cost): New.
16683 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
16684 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
16685 parm numbers to be present.
16687 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
16689 PR rtl-optimization/58048
16690 * lra-constraints.c (process_alt_operands): Don't check asm
16691 operand on register.
16693 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
16695 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
16696 the implied StoreLoad barrier for atomic operations if before.
16698 2013-08-02 Jan Hubicka <jh@suse.cz>
16699 Martin Liska <marxin.liska@gmail.com>
16701 * cgraph.c (cgraph_function_body_availability): Do not check
16703 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
16704 symtab_node_availability): Declare.
16705 * ipa.c (can_replace_by_local_alias): New.
16706 (function_and_variable_visibility): Use it.
16707 * symtab.c (symtab_for_node_and_aliases,
16708 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
16710 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
16712 PR rtl-optimization/57963
16713 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
16714 (lra_constraints): Use them.
16716 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
16718 * config/arm/types.md (define_attr "type"): Add "load_acq"
16720 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
16722 (cortex_a53_store1): Likewise.
16724 2013-08-01 Jan Hubicka <jh@suse.cz>
16726 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
16727 partitions are not needed.
16729 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
16731 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
16732 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
16733 MAYBE_NON_Q_CLASS_P where appropriate.
16735 2013-08-01 Jan Hubicka <jh@suse.cz>
16737 * cgraph.h (release_function_body): Declare.
16738 * tree.c (free_lang_data_in_decl): Free, parameters and return values
16739 of unused delcarations.
16741 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16743 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
16744 RTL form when subtracting a constant.
16746 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16748 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
16749 Generate canonical plus rtx with negated immediate instead of minus
16751 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
16753 2013-08-01 Jan Hubicka <jh@suse.cz>
16755 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
16756 (cgraph_release_function_body): Likewise.
16757 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
16758 * cgraph.h (cgrpah_node): Rename abstract_and_needed
16759 to used_as_abstract_origin.
16760 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
16761 symbols used as abstract origins.
16762 * cgraphunit.c (analyze_functions): Update.
16763 * ipa.c (symtab_remove_unreachable_nodes): Recompute
16764 used_as_abstract_origin.
16765 * tree-inline.c (tree_function_versioning): Update
16766 used_as_abstract_origin; be ready for DECL_RESULT and
16767 DECL_ARGUMENTS to be NULL.
16769 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
16770 for abstract functions.
16771 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
16774 2013-08-01 Jan Hubicka <jh@suse.cz>
16776 * profile.c (compute_value_histograms): Fix thinko.
16778 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16780 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
16781 aarch-common-protos.h to extra_headers.
16782 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16783 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
16784 * config/aarch64/t-aarch64 (aarch-common.o): Define.
16786 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16788 * config/aarch64/aarch64.md (define_attr "type"): Delete.
16789 Include "../arm/types.md". Define "type" attribute for all patterns.
16790 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
16793 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
16795 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
16796 to support power8 load fusion.
16797 (fusion_gpr_mem_load): Likewise.
16799 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
16801 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
16802 declarations for power8 load fusion.
16803 (emit_fusion_gpr_load): Likewise.
16805 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
16806 tuning for power8, turn on fusion mode by default. Turn on sign
16807 extending fusion mode if normal fusion mode is on, and we are at
16809 (fusion_gpr_load_p): New function, return true if we can fuse an
16810 addis instruction with a dependent load to a GPR.
16811 (emit_fusion_gpr_load): Emit the instructions for power8 load
16814 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
16815 (VSX load fusion peepholes): New peepholes to fuse together an
16816 addi instruction with a VSX load instruction.
16818 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
16819 peepholes to fuse an addis instruction with a load to a GPR base
16820 register. If we are supporting sign extending fusions, convert
16821 sign extending loads to zero extending loads and add an explicit
16824 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16826 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
16827 aarch-common-protos.h to extra_headers.
16828 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16829 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
16830 (arm_early_store_addr_dep): Likewise.
16831 (arm_no_early_alu_shift_dep): Likewise.
16832 (arm_no_early_alu_shift_value_dep): Likewise.
16833 (arm_no_early_mul_dep): Likewise.
16834 (arm_no_early_store_addr_dep): Likewise.
16835 (arm_mac_accumulator_is_mul_result): Likewise.
16836 (arm_mac_accumulator_is_result): Likewise.
16837 * config/arm/aarch-common.c: ... here. New file.
16838 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
16840 (arm_early_store_addr_dep): Likewise.
16841 (arm_no_early_alu_shift_dep): Likewise.
16842 (arm_no_early_alu_shift_value_dep): Likewise.
16843 (arm_no_early_mul_dep): Likewise.
16844 (arm_no_early_store_addr_dep): Likewise.
16845 (arm_mac_accumulator_is_mul_result): Likewise.
16846 (arm_mac_accumulator_is_result): Likewise.
16847 * config/arm/aarch-common-protos.h: ... here. New file.
16848 * config/arm/t-arm (aarch-common.o): Define.
16850 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16852 * config/arm/arm.md: Include new file "types.md".
16853 (define_attr "type"): Move from here to ...
16854 (define_attr "mul32"): Likewise.
16855 (define_attr "mul64"): Likewise.
16856 * config/arm/types.md: ... here. New file.
16858 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
16860 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
16861 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
16863 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16865 * gen-pass-instances.awk: Fix offset of substr().
16867 2013-07-31 David Malcolm <dmalcolm@redhat.com>
16869 * Makefile.in (pass-instances.def): New.
16870 (passes.o): Replace dependency on passes.def with one on
16873 * gen-pass-instances.awk: New.
16875 * passes.c (pass_manager::pass_manager): Use pass-instances.def
16876 rather than passes.def, updating local definition of NEXT_PASS
16877 macro to add an extra NUM parameter (currently unused).
16879 2013-07-30 David Malcolm <dmalcolm@redhat.com>
16881 * Makefile.in (PASS_MANAGER_H): New.
16882 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
16883 (passes.o): Likewise.
16884 (statistics.o): Likewise.
16885 (cgraphunit.o): Likewise.
16886 (context.o): Depend on PASS_MANAGER_H.
16888 * pass_manager.h: New.
16890 * cgraphunit.c (cgraph_add_new_function): Update for moves
16891 of globals to fields of pass_manager.
16892 (analyze_function): Likewise.
16893 (expand_function): Likewise.
16894 (ipa_passes): Likewise.
16895 (compile): Likewise.
16897 * context.c (context::context): New.
16898 * context.h (context::context): New.
16899 (context::get_passes): New.
16900 (context::passes_): New.
16902 * lto-cgraph.c (input_node): Update for moves of globals to
16903 fields of pass_manager.
16905 * passes.c (all_passes): Remove, in favor of a field of the
16906 same name within the new class pass_manager.
16907 (all_small_ipa_passes): Likewise.
16908 (all_lowering_passes): Likewise.
16909 (all_regular_ipa_passes): Likewise.
16910 (all_late_ipa_passes): Likewise.
16911 (all_lto_gen_passes): Likewise.
16912 (passes_by_id): Likewise.
16913 (passes_by_id_size): Likewise.
16914 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16915 the new class pass_manager.
16916 (set_pass_for_id): Convert to...
16917 (pass_manager::set_pass_for_id): ...method.
16918 (get_pass_for_id): Convert to...
16919 (pass_manager::get_pass_for_id): ...method.
16920 (register_one_dump_file): Move body of implementation into...
16921 (pass_manager::register_one_dump_file): ...here.
16922 (register_dump_files_1): Convert to...
16923 (pass_manager::register_dump_files_1): ...method.
16924 (register_dump_files): Convert to...
16925 (pass_manager::register_dump_files): ...method.
16926 (create_pass_tab): Update for moves of globals to fields of
16928 (dump_passes): Move body of implementation into...
16929 (pass_manager::dump_passes): ...here.
16930 (register_pass): Move body of implementation into...
16931 (pass_manager::register_pass): ...here.
16932 (init_optimization_passes): Convert into...
16933 (pass_manager::pass_manager): ...constructor for new
16934 pass_manager class, and initialize the pass_lists array.
16935 (check_profile_consistency): Update for moves of globals to
16936 fields of pass_manager.
16937 (dump_profile_report): Move body of implementation into...
16938 (pass_manager::dump_profile_report): ...here.
16939 (ipa_write_summaries_1): Update for moves of pass lists from
16940 being globals to fields of pass_manager.
16941 (ipa_write_optimization_summaries): Likewise.
16942 (ipa_read_summaries): Likewise.
16943 (ipa_read_optimization_summaries): Likewise.
16944 (execute_all_ipa_stmt_fixups): Likewise.
16946 * statistics.c (statistics_fini): Update for moves of globals to
16947 fields of pass_manager.
16949 * toplev.c (general_init): Replace call to
16950 init_optimization_passes with construction of the pass_manager
16953 * tree-pass.h (all_passes): Remove, in favor of a field of the
16954 same name within the new class pass_manager.
16955 (all_small_ipa_passes): Likewise.
16956 (all_lowering_passes): Likewise.
16957 (all_regular_ipa_passes): Likewise.
16958 (all_lto_gen_passes): Likewise.
16959 (all_late_ipa_passes): Likewise.
16960 (passes_by_id): Likewise.
16961 (passes_by_id_size): Likewise.
16962 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16963 the new class pass_manager.
16964 (get_pass_for_id): Remove.
16966 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16968 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
16971 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16973 * arm.md (mulhi3): New expand pattern.
16975 2013-07-30 Jan Hubicka <jh@suse.cz>
16976 Martin Liska <marxin.liska@gmail.com>
16978 * profile.c (compute_value_histograms): Do not ICE when
16979 there is mismatch only on some counters.
16981 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16983 PR rtl-optimization/57637
16984 * function.c (move_insn_for_shrink_wrap): Also check the
16985 GEN set of the LIVE problem for the liveness analysis
16986 if it exists, otherwise give up.
16988 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16990 PR tree-optimization/57993
16991 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
16992 replaced statement in the candidate table.
16993 (phi_add_costs): Return infinite cost when the hidden basis does
16994 not dominate all phis on which the candidate is dependent.
16995 (replace_one_candidate): Record replaced statement in the
16998 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
17000 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
17001 (ashlv2si3): New expander.
17002 (*ashlv2si3_i): New define_insn_and_split.
17003 * predicates.md (float_operation): Allow patterns with three
17004 basic sub-patterns.
17006 PR rtl-optimization/58021
17007 * mode-switching.c (create_pre_exit): Always split off preceding
17008 insns if we are not at the basic block head.
17010 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
17012 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
17013 (UCLIBC_DYNAMIC_LINKER): New macro.
17014 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
17016 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
17017 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
17019 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
17020 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
17021 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
17022 for SF and DF modes. Use ieee_quad_format for TF mode.
17023 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
17024 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
17025 (mips_option_override): Handle `-mnan=legacy'.
17026 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
17027 `-mabs=2008' and `-mnan=2008'.
17028 (OPTION_DEFAULT_SPECS): Add "nan" default.
17029 (ASM_SPEC): Handle `-mnan='.
17030 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
17031 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
17032 comment accordingly.
17033 (neg<mode>2): Likewise.
17034 * config/mips/mips.opt (mabs, mnan): New options.
17035 * doc/install.texi (Configuration): Document `--with-nan=' option.
17036 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
17038 (MIPS Options): Document them.
17039 * config.gcc <mips*-*-*>: Handle `--with-nan='.
17040 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
17041 * configure: Regenerate.
17042 * config.in: Regenerate.
17044 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
17046 * config/i386/i386.md (float post-reload splitters): Do not check
17047 for subregs of SSE registers.
17049 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
17050 H.J. Lu <hongjiu.lu@intel.com>
17054 * config/i386/i386.md (post-reload splitter
17055 to avoid partial SSE reg dependency stalls): New pattern.
17057 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
17059 * config/s390/s390.md ("movcc"): Swap load and store instructions.
17061 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
17063 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
17064 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
17066 2013-07-26 Cary Coutant <ccoutant@google.com>
17068 * dwarf2out.c (die_checksum_ordered): Don't include template
17069 instantiations in signature.
17070 (is_template_parameter): New function.
17071 (is_template_instantiation): New function.
17072 (generate_skeleton_bottom_up): Don't include template instantiations
17074 (generate_skeleton): Likewise.
17075 (break_out_comdat_types): Move recursive call to break out nested
17077 (prune_unused_types_mark_generic_parms_dies): Call
17078 is_template_parameter.
17080 2013-07-26 Ian Bolton <ian.bolton@arm.com>
17082 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
17083 uses vector registers.
17084 * config/aarch64/iterators.md: Add attributes rtn and vas.
17086 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17087 Richard Earnshaw <richard.earnshaw@arm.com>
17089 * combine.c (simplify_comparison): Re-canonicalize operands
17091 * config/arm/arm.md (movcond_addsi): New splitter.
17093 2013-07-25 Sterling Augustine <saugustine@google.com>
17095 * dwarf2out.c (size_of_pubnames): Move code to...
17096 (include_pubname_in_output): ...here. New.
17097 (want_pubnames): Rearrange.
17098 (output_pubnames): Call include_pubname_in_output. Move assertion.
17100 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
17102 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
17104 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
17107 * doc/extend.texi: Remove obsolete builtins. Fix
17108 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
17110 2013-07-25 Jan Hubicka <jh@suse.cz>
17112 * cgraph.c (release_function_body): Break out from ...
17113 (cgraph_release_function_body): ... this one; also release DECL_RESULT
17114 and DECL_ARGUMENTS.
17115 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
17116 old_tree in the map.
17117 (create_specialized_node): Update.
17118 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
17120 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
17121 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
17122 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
17123 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
17126 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17128 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
17129 addsi3_carryin_alt2_<optab>): Correct output template.
17131 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17133 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
17134 Adjust for arm_restrict_it.
17135 Remove trailing whitespace.
17137 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
17139 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
17140 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
17142 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
17144 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
17146 PR rtl-optimization/57960
17147 * lra-constraints.c (process_alt_operands): Use the right mode
17148 when checking strict_low.
17150 2013-07-25 Jan Hubicka <jh@suse.cz>
17152 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
17153 * cgraph.c (cgraph_remove_node): Do not release function body
17154 when in cgraph streaming.
17155 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
17156 in other partitions are not considered reachable; fix handling of
17159 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17161 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
17163 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17168 * config/arm/arm.md ("*sibcall_insn): Replace use of
17169 Ss with US. Adjust output for v5 and v4t.
17170 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
17172 * config/arm/constraints.md ("Ss"): Rename to US.
17174 2013-07-25 Terry Guo <terry.guo@arm.com>
17176 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
17177 shift_add/shift_sub0/shift_sub1 RTXs.
17179 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
17180 Anton Blanchard <anton@au1.ibm.com>
17182 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
17183 (altivec_vpks<VI_char>ss): Likewise.
17184 (altivec_vpks<VI_char>us): Likewise.
17185 (altivec_vpku<VI_char>us): Likewise.
17186 (altivec_vpku<VI_char>um): Likewise.
17188 2013-07-24 David Malcolm <dmalcolm@redhat.com>
17190 Introduce context class.
17192 * Makefile.in (CONTEXT_H): New.
17193 (OBJS): Add context.o.
17194 (toplev.o): Add CONTEXT_H to dependencies.
17197 * toplev.c (general_init): Create the singleton gcc::context instance.
17203 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
17205 PR rtl-optimization/57968
17206 * mode-switching.c (create_pre_exit): Allow instructions that
17207 don't set a return register to need a non-exit mode.
17209 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
17210 Anton Blanchard <anton@au1.ibm.com>
17212 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
17213 operands to vperm for little endian.
17214 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
17215 of lvsl to create the control mask for a vperm for little endian.
17217 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17218 Anton Blanchard <anton@au1.ibm.com>
17220 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
17221 two operands for little-endian.
17223 2013-07-23 Steve Ellcey <sellcey@mips.com>
17225 * config/mips/mips.c (mips_case_values_threshold): New.
17226 (TARGET_CASE_VALUES_THRESHOLD): Define.
17228 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17229 Anton Blanchard <anton@au1.ibm.com>
17231 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
17232 selection of field for vector splat in little endian mode.
17234 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
17236 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
17237 expanders to rs6000.md.
17238 (ior<mode>3): Likewise.
17239 (and<mode>3): Likewise.
17240 (one_cmpl<mode>2): Likewise.
17241 (nor<mode>3): Likewise.
17242 (andc<mode>3): Likewise.
17243 (eqv<mode>3): Likewise.
17244 (nand<mode>3): Likewise.
17245 (orc<mode>3): Likewise.
17247 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
17250 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
17251 to split multi-word logical operations.
17252 (rs6000_split_logical_di): Likewise.
17253 (rs6000_split_logical): Likewise.
17255 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
17256 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
17257 and allow TImode operations in 32-bit.
17258 (vsx_and<mode>3_64bit): Likewise.
17259 (vsx_ior<mode>3_32bit): Likewise.
17260 (vsx_ior<mode>3_64bit): Likewise.
17261 (vsx_xor<mode>3_32bit): Likewise.
17262 (vsx_xor<mode>3_64bit): Likewise.
17263 (vsx_one_cmpl<mode>2_32bit): Likewise.
17264 (vsx_one_cmpl<mode>2_64bit): Likewise.
17265 (vsx_nor<mode>3_32bit): Likewise.
17266 (vsx_nor<mode>3_64bit): Likewise.
17267 (vsx_andc<mode>3_32bit): Likewise.
17268 (vsx_andc<mode>3_64bit): Likewise.
17269 (vsx_eqv<mode>3_32bit): Likewise.
17270 (vsx_eqv<mode>3_64bit): Likewise.
17271 (vsx_nand<mode>3_32bit): Likewise.
17272 (vsx_nand<mode>3_64bit): Likewise.
17273 (vsx_orc<mode>3_32bit): Likewise.
17274 (vsx_orc<mode>3_64bit): Likewise.
17276 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
17277 logical types in GPRs.
17279 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
17280 logical insns to rs6000.md, and allow TImode operations in
17282 (altivec_ior<mode>3): Likewise.
17283 (altivec_xor<mode>3): Likewise.
17284 (altivec_one_cmpl<mode>2): Likewise.
17285 (altivec_nor<mode>3): Likewise.
17286 (altivec_andc<mode>3): Likewise.
17288 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
17289 attributes for moving the 128-bit logical operations into
17291 (BOOL_REGS_OUTPUT): Likewise.
17292 (BOOL_REGS_OP1): Likewise.
17293 (BOOL_REGS_OP2): Likewise.
17294 (BOOL_REGS_UNARY): Likewise.
17295 (BOOL_REGS_AND_CR0): Likewise.
17296 (one_cmpl<mode>2): Add support for DI logical operations on
17297 32-bit, splitting the operations to 32-bit.
17298 (anddi3): Likewise.
17299 (iordi3): Likewise.
17300 (xordi3): Likewise.
17301 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
17302 changes to combine the 32/64-bit code, allow logical operations on
17303 TI mode in 32-bit, and to use similar match_operator patterns like
17304 scalar mode uses. Combine the Altivec and VSX code for logical
17305 operations, and move it here.
17306 (ior<mode>3, 128-bit types): Likewise.
17307 (xor<mode>3, 128-bit types): Likewise.
17308 (one_cmpl<mode>3, 128-bit types): Likewise.
17309 (nor<mode>3, 128-bit types): Likewise.
17310 (andc<mode>3, 128-bit types): Likewise.
17311 (eqv<mode>3, 128-bit types): Likewise.
17312 (nand<mode>3, 128-bit types): Likewise.
17313 (orc<mode>3, 128-bit types): Likewise.
17314 (and<mode>3_internal): Likewise.
17315 (bool<mode>3_internal): Likewise.
17316 (boolc<mode>3_internal1): Likewise.
17317 (boolc<mode>3_internal2): Likewise.
17318 (boolcc<mode>3_internal1): Likewise.
17319 (boolcc<mode>3_internal2): Likewise.
17320 (eqv<mode>3_internal1): Likewise.
17321 (eqv<mode>3_internal2): Likewise.
17322 (one_cmpl1<mode>3_internal): Likewise.
17324 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
17326 * config/microblaze/microblaze.c (microblaze_expand_prologue):
17327 Rename flag_stack_usage to flag_stack_usage_info.
17329 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
17331 * config/microblaze/sync.md: New file.
17332 * config/microblaze/microblaze.md: Include sync.md
17333 * config/microblaze/microblaze.c: Add print_operand 'y'.
17334 * config/microblaze/constraints.md: Add memory_contraint
17335 'Q' which is a single register.
17337 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
17339 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
17341 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
17343 * reload.c (find_reloads): Exit loop once we find this operand
17344 cannot be reloaded somehow for this alternative.
17346 * reload.c (find_reloads): Exit loop once we find a hard register.
17348 * rtlanal.c (computed_jump_p): Exit loop once we find label
17351 * i386.c (ix86_pad_returns): Exit loop after setting replace.
17353 * cfgloopmanip.c (remove_path): Exit loop after setting
17356 * gensupport.c (subst_dup): Avoid loop if code is not
17357 MATCH_DUP nor MATCH_OP_DUP.
17359 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
17361 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
17363 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17365 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
17366 true for SP_REGNUM if mode == ptr_mode.
17367 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
17368 with value R0_REGNUM + 31.
17370 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17372 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
17373 pad pointer-typed argument downward.
17375 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17377 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
17378 and __ILP32__ when the ILP32 model is in use.
17380 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17382 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
17383 (aarch64_load_symref_appropriately): In the case of
17384 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
17385 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
17386 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
17387 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
17388 if 'mode' doesn't equal to 'ptr_mode'.
17389 (aarch64_output_mi_thunk): Add an assertion on the alignment of
17390 'vcall_offset'; change to call aarch64_emit_move differently depending
17391 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
17392 to calculate the upper bound of 'vcall_offset'.
17393 (aarch64_cannot_force_const_mem): Change to also return true if
17395 (aarch64_legitimize_reload_address): In the case of large
17396 displacements, add new local variable 'xmode' and an assertion
17397 based on it; change to use 'xmode' to generate the new rtx and
17399 (aarch64_asm_trampoline_template): Change to generate the template
17400 differently depending on TARGET_ILP32 or not; change to use
17401 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
17402 (aarch64_trampoline_size): Removed.
17403 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
17404 and replace immediate literals with it. Change to use 'ptr_mode'
17405 instead of 'DImode' and call convert_memory_address if the mode
17406 of 'fnaddr' doesn't equal to 'ptr_mode'.
17407 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
17409 (aarch64_elf_asm_destructor): Likewise.
17410 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
17411 on TARGET_ILP32 instead of aarch64_trampoline_size.
17412 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
17413 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
17414 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
17415 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
17416 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
17417 (storewb_pair<GPI:mode>_<P:mode>): ... this.
17418 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
17419 depending on the value of 'mode'.
17420 (add_losym_<mode>): New.
17421 (ldr_got_small_<mode>): New, based on ldr_got_small.
17422 (ldr_got_small): Remove.
17423 (ldr_got_small_sidi): New.
17424 * config/aarch64/iterators.md (P): New.
17425 (PTR): Change to 'ptr_mode' in the condition.
17427 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17429 * config.gcc (aarch64*-*-*): Support --with-abi.
17430 (aarch64*-*-elf): Support --with-multilib-list.
17431 (aarch64*-*-linux*): Likewise.
17432 (supported_defaults): Add abi to aarch64*-*-*.
17433 * configure.ac: Mention AArch64 for --with-multilib-list.
17434 * configure: Re-generated.
17435 * config/aarch64/biarchilp32.h: New file.
17436 * config/aarch64/biarchlp64.h: New file.
17437 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
17439 (MULTILIB_DEFAULTS): Ditto.
17440 (DRIVER_SELF_SPECS): Ditto.
17441 (ASM_SPEC): Update to also substitute -mabi.
17442 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
17443 file whose name depends on -mabi= and -mbig-endian.
17444 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
17446 (POINTER_SIZE): New define.
17447 (POINTERS_EXTEND_UNSIGNED): Ditto.
17448 (enum aarch64_abi_type): New enumeration tag.
17449 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
17450 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
17451 (TARGET_ILP32): New define.
17452 * config/aarch64/aarch64.opt (mabi): New.
17453 (aarch64_abi): New.
17454 (ilp32, lp64): New values for -mabi.
17455 * config/aarch64/t-aarch64 (comma): New define.
17456 (MULTILIB_OPTIONS): Ditto.
17457 (MULTILIB_DIRNAMES): Ditto.
17458 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
17459 * doc/invoke.texi: Document -mabi for AArch64.
17461 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
17463 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
17465 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
17466 Anton Blanchard <anton@au1.ibm.com>
17468 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
17469 endianness when selecting field to splat.
17471 2013-07-22 Eric Christopher <echristo@gmail.com>
17473 * dwarf2out.c (die_odr_checksum): New function to use
17474 CHECKSUM_ macros and ULEB128 for DIE tag.
17475 (generate_type_signature): Use.
17477 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
17479 * config.gcc (sparc*-*-*): Accept leon3 processor.
17480 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
17481 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
17482 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
17483 * config/sparc/sparc.opt (enum processor_type): Add leon3.
17484 (mfix-ut699): Adjust comment.
17485 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
17486 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
17487 (CPP_CPU_SPEC): Likewise.
17488 (ASM_CPU_SPEC): Likewise.
17489 * config/sparc/sparc.c (leon3_cost): New constant.
17490 (sparc_option_override): Add leon3 support.
17491 (mem_ref): New function.
17492 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
17493 (sparc_do_work_around_errata): Look into the instruction in the delay
17494 slot and adjust accordingly. Add fix for the data cache nullify issues
17495 of the UT699. Change insertion position for the NOP.
17496 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
17497 (leon3_load): New reservation.
17498 (leon_store): Bump latency to 2.
17499 (grfpu): New automaton.
17500 (grfpu_alu): New unit.
17501 (grfpu_ds): Likewise.
17502 (leon_fp_alu): Adjust.
17503 (leon_fp_mult): Delete.
17504 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
17505 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
17506 * config/sparc/sparc.md (cpu): Add leon3.
17507 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
17508 (swapsi): Likewise.
17509 (atomic_test_and_set): Likewise.
17510 (ldstub): Likewise.
17512 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
17514 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
17515 default for R5900 targets.
17516 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
17517 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
17518 * config/mips/mips.c (mips_option_override): Report an error for
17519 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
17520 for -march=r5900 -mhard-float.
17522 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
17524 * df-problems.c (can_move_insns_across): Exit loop once we
17525 find a non-fixed, non-global register.
17527 * ipa-pure-const.c (propagate_nothrow): Exit loop after
17530 * omega.c (omega_eliminate_red): Break after setting red_found.
17531 (omega_problem_has_red_equations): Similarly after setting found.
17532 (omega_query_variable): Similarly after setting coupled.
17534 2013-07-22 Marek Polacek <polacek@redhat.com>
17536 * gimplify.c: Don't include gimple.h twice.
17538 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17540 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
17541 instead of TARGET_THUMB1.
17542 (Pz): New constraint.
17543 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
17545 (compare_negsi_si): Likewise.
17546 (compare_addsi2_op0): Likewise.
17547 (compare_addsi2_op1): Likewise.
17548 (addsi3_carryin_<optab>): Likewise.
17549 (addsi3_carryin_alt2_<optab>): Likewise.
17550 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
17551 for arm_restrict_it.
17552 (subsi3_carryin): Likewise.
17553 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
17554 (minmax_arithsi): Disable for arm_restrict_it.
17555 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
17556 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
17557 (satsi_<SAT:code>_shift): Likewise.
17558 (arm_shiftsi3): Add alternative for 16-bit encoding.
17559 (arm32_movhf): Disable for arm_restrict_it.
17560 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
17561 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
17563 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
17565 * config/arm/arm.md (attribute "insn"): Delete.
17566 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
17567 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
17568 (not_shiftsi): Update for attribute change.
17569 (not_shiftsi_compare0): Likewise.
17570 (not_shiftsi_compare0_scratch): Likewise.
17571 (arm_one_cmplsi2): Likewise.
17572 (thumb1_one_cmplsi2): Likewise.
17573 (notsi_compare0): Likewise.
17574 (notsi_compare0_scratch): Likewise.
17575 (thumb1_movdi_insn): Likewise.
17576 (arm_movsi_insn): Likewise.
17577 (movhi_insn_arch4): Likewise.
17578 (movhi_bytes): Likewise.
17579 (arm_movqi_insn): Likewise.
17580 (thumb1_movqi_insn): Likewise.
17581 (arm32_movhf): Likewise.
17582 (thumb1_movhf): Likewise.
17583 (arm_movsf_soft_insn): Likewise.
17584 (thumb1_movsf_insn): Likewise.
17585 (thumb_movdf_insn): Likewise.
17586 (movsicc_insn): Likewise.
17587 (movsfcc_soft_insn): Likewise.
17588 (and_scc): Likewise.
17589 (cond_move): Likewise.
17590 (if_move_not): Likewise.
17591 (if_not_move): Likewise.
17592 (if_shift_move): Likewise.
17593 (if_move_shift): Likewise.
17594 (if_shift_shift): Likewise.
17595 (if_not_arith): Likewise.
17596 (if_arith_not): Likewise.
17597 (cond_move_not): Likewise.
17598 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
17599 (neon_mov<mode>): Likewise.
17600 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
17601 (thumb2_movsi_vfp): Likewise.
17602 (movsf_vfp): Likewise.
17603 (thumb2_movsf_vfp): Likewise.
17604 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
17606 (cortexa7_older_only): Likewise.
17607 (cortexa7_younger): Likewise.
17608 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
17609 (1020alu_shift_op): Likewise.
17610 (1020alu_shift_reg_op): Likewise.
17611 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
17612 (alu_shift_op): Likewise.
17613 (alu_shift_reg_op): Likewise.
17614 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
17615 (11_alu_shift_op): Likewise.
17616 (11_alu_shift_reg_op): Likewise.
17617 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
17618 (9_alu_shift_reg_op): Likewise.
17619 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17621 (cortex_a15_alu_shift): Likewise.
17622 (cortex_a15_alu_shift_reg): Likewise.
17623 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
17624 (cortex_a5_alu_shift): Likewise.
17625 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
17627 (cortex_a53_alu_shift): Likewise.
17628 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17630 (cortex_a7_alu_reg): Likewise.
17631 (cortex_a7_alu_shift): Likewise.
17632 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17633 (cortex_a8_alu_shift): Likewise.
17634 (cortex_a8_alu_shift_reg): Likewise.
17635 (cortex_a8_mov): Likewise.
17636 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
17637 (cortex_a9_dp_shift): Likewise.
17638 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
17639 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
17640 (cortex_r4_mov): Likewise.
17641 (cortex_r4_alu_shift): Likewise.
17642 (cortex_r4_alu_shift_reg): Likewise.
17643 * config/arm/fa526.md (526_alu_op): Update for attribute change.
17644 (526_alu_shift_op): Likewise.
17645 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
17646 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
17647 (626te_alu_shift_op): Likewise.
17648 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
17649 (726te_alu_op): Likewise.
17650 (726te_alu_shift_op): Likewise.
17651 (726te_alu_shift_reg_op): Likewise.
17652 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
17653 (mp626_alu_shift_op): Likewise.
17654 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
17655 (pj4_alu_e1_conds): Likewise.
17656 (pj4_alu): Likewise.
17657 (pj4_alu_conds): Likewise.
17658 (pj4_shift): Likewise.
17659 (pj4_shift_conds): Likewise.
17660 (pj4_alu_shift): Likewise.
17661 (pj4_alu_shift_conds): Likewise.
17663 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17665 * config/arm/predicates.md (shiftable_operator_strict_it):
17667 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
17668 Disable cond_exec version for arm_restrict_it.
17669 (thumb2_smaxsi3): Convert to generate cond_exec.
17670 (thumb2_sminsi3): Likewise.
17671 (thumb32_umaxsi3): Likewise.
17672 (thumb2_uminsi3): Likewise.
17673 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
17674 (thumb2_neg_abssi2): Likewise.
17675 (thumb2_mov_scc): Add alternative for 16-bit encoding.
17676 (thumb2_movsicc_insn): Adjust alternatives.
17677 (thumb2_mov_negscc): Disable for arm_restrict_it.
17678 (thumb2_mov_negscc_strict_it): New pattern.
17679 (thumb2_mov_notscc_strict_it): New pattern.
17680 (thumb2_mov_notscc): Disable for arm_restrict_it.
17681 (thumb2_ior_scc): Likewise.
17682 (thumb2_ior_scc_strict_it): New pattern.
17683 (thumb2_cond_move): Adjust for arm_restrict_it.
17684 (thumb2_cond_arith): Disable for arm_restrict_it.
17685 (thumb2_cond_arith_strict_it): New pattern.
17686 (thumb2_cond_sub): Adjust for arm_restrict_it.
17687 (thumb2_movcond): Likewise.
17688 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
17689 (thumb2_zero_extendhisi2_v6): Likewise.
17690 (thumb2_zero_extendqisi2_v6): Likewise.
17691 (orsi_notsi_si): Likewise.
17692 (orsi_not_shiftsi_si): Likewise.
17694 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
17696 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
17697 instruction sequence is 1 byte shorter.
17699 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
17701 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
17702 it is not needed after split.
17704 2013-07-20 Iain Sandoe <iain@codesourcery.com>
17707 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
17708 second label for nonlocal goto receivers. Don't output pic base labels
17709 unless we're producing PIC; mark that action unreachable().
17710 (ix86_save_reg): If the function contains a nonlocal label, save the
17712 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
17713 * config/darwin.c (emitted_pic_label_num): New GTY.
17714 (update_pic_label_number_if_needed): New.
17715 (machopic_output_function_base_name): Adjust for nonlocal receiver
17717 (machopic_should_output_picbase_label): New.
17718 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
17719 (nonlocal_goto_receiver): New insn and split.
17721 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
17723 * config/aarch64/aarch64-builtins.c
17724 (aarch64_fold_builtin): Fold abs in all modes.
17725 * config/aarch64/aarch64-simd-builtins.def
17726 (abs): Enable for all modes.
17727 * config/aarch64/arm_neon.h
17728 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
17729 (vabs_f64): Add missing intrinsic.
17731 2013-07-19 Ian Bolton <ian.bolton@arm.com>
17733 * config/aarch64/arm_neon.h (vabs_s64): New function
17735 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
17738 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
17739 * config/avr/avr.md (adjust_len): Add `round'.
17740 * config/avr/avr-protos.h (avr_out_round): New prototype.
17741 (avr_out_plus): Add `out_label' argument.
17742 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
17743 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
17744 Handle the case where `insn' is just a pattern.
17745 (avr_out_bitop): Handle the case where `insn' is just a pattern.
17746 (avr_out_round): New function.
17747 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
17749 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
17751 * config/microblaze/microblaze.c (microblaze_expand_prologue):
17752 Add check for flag_stack_usage to handle -fstack-usage support
17754 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
17756 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
17757 interaction for new Power8 flags and VSX.
17759 2013-07-18 Sriraman Tallam <tmsriram@google.com>
17761 PR middle-end/57698
17762 * tree-inline.c (expand_call_inline): Emit errors during
17763 early_inlining only if optimization is not turned on.
17765 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17769 * passes.c (init_optimization_passes): Move the construction of
17770 the pass hierarchy into a new passes.def file.
17772 * Makefile.in (passes.o): Add dependency on passes.def.
17774 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17776 * passes.c (init_optimization_passes): Introduce macros for
17777 constructing the tree of passes (INSERT_PASSES_AFTER,
17778 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
17779 TERMINATE_PASS_LIST).
17781 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
17782 Wei Mi <wmi@google.com>
17784 PR rtl-optimization/57878
17785 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
17787 (reload_pseudo_compare_func): Check nregs first for reload
17790 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17792 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
17794 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
17796 * read-rtl.c (validate_const_int): Once an invalid character is
17797 seen, quit the loop.
17799 * gengtype.c (write_roots): Similarly once we find the "deletable"
17800 or "if_marked" option.
17802 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17804 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
17805 "xtab" and "sat". Move value "clz" from here to ...
17806 (attriubte "type"): ... here.
17807 (satsi_<SAT:code>): Delete "insn" attribute.
17808 (satsi_<SAT:code>_shift): Likewise.
17809 (arm_zero_extendqisi2addsi): Likewise.
17810 (arm_extendqisi2addsi): Likewise.
17811 (clzsi2): Update for attribute changes.
17812 (rbitsi2): Likewise.
17813 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
17815 (arm_usatsihi): Likewise.
17816 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17818 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17820 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
17821 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
17822 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
17823 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
17824 in alphabetical order.
17825 (attribute "core_cycles"): Update for attribute changes.
17826 (arm_addsi3): Likewise.
17827 (addsi3_compare0): Likewise.
17828 (addsi3_compare0_scratch): Likewise.
17829 (addsi3_compare_op1): Likewise.
17830 (addsi3_compare_op2): Likewise.
17831 (compare_addsi2_op0): Likewise.
17832 (compare_addsi2_op1): Likewise.
17833 (addsi3_carryin_shift_<optab>): Likewise.
17834 (subsi3_carryin_shift): Likewise.
17835 (rsbsi3_carryin_shift): Likewise.
17836 (arm_subsi3_insn): Likewise.
17837 (subsi3_compare0): Likewise.
17838 (subsi3_compare): Likewise.
17839 (arm_andsi3_insn): Likewise.
17840 (thumb1_andsi3_insn): Likewise.
17841 (andsi3_compare0): Likewise.
17842 (andsi3_compare0_scratch): Likewise.
17843 (zeroextractsi_compare0_scratch
17844 (andsi_not_shiftsi_si): Likewise.
17845 (iorsi3_insn): Likewise.
17846 (iorsi3_compare0): Likewise.
17847 (iorsi3_compare0_scratch): Likewise.
17848 (arm_xorsi3): Likewise.
17849 (thumb1_xorsi3_insn): Likewise.
17850 (xorsi3_compare0): Likewise.
17851 (xorsi3_compare0_scratch): Likewise.
17852 (satsi_<SAT:code>_shift): Likewise.
17854 (arm_shiftsi3): Likewise.
17855 (shiftsi3_compare0): Likewise.
17856 (not_shiftsi): Likewise.
17857 (not_shiftsi_compare0): Likewise.
17858 (not_shiftsi_compare0_scratch): Likewise.
17859 (arm_one_cmplsi2): Likewise.
17860 (thumb_one_complsi2): Likewise.
17861 (notsi_compare0): Likewise.
17862 (notsi_compare0_scratch): Likewise.
17863 (thumb1_zero_extendhisi2): Likewise.
17864 (arm_zero_extendhisi2): Likewise.
17865 (arm_zero_extendhisi2_v6): Likewise.
17866 (arm_zero_extendhisi2addsi): Likewise.
17867 (thumb1_zero_extendqisi2): Likewise.
17868 (thumb1_zero_extendqisi2_v6): Likewise.
17869 (arm_zero_extendqisi2): Likewise.
17870 (arm_zero_extendqisi2_v6): Likewise.
17871 (arm_zero_extendqisi2addsi): Likewise.
17872 (thumb1_extendhisi2): Likewise.
17873 (arm_extendhisi2): Likewise.
17874 (arm_extendhisi2_v6): Likewise.
17875 (arm_extendqisi): Likewise.
17876 (arm_extendqisi_v6): Likewise.
17877 (arm_extendqisi2addsi): Likewise.
17878 (thumb1_extendqisi2): Likewise.
17879 (thumb1_movdi_insn): Likewise.
17880 (arm_movsi_insn): Likewise.
17881 (movsi_compare0): Likewise.
17882 (movhi_insn_arch4): Likewise.
17883 (movhi_bytes): Likewise.
17884 (arm_movqi_insn): Likewise.
17885 (thumb1_movqi_insn): Likewise.
17886 (arm32_movhf): Likewise.
17887 (thumb1_movhf): Likewise.
17888 (arm_movsf_soft_insn): Likewise.
17889 (thumb1_movsf_insn): Likewise.
17890 (movdf_soft_insn): Likewise.
17891 (thumb_movdf_insn): Likewise.
17892 (arm_cmpsi_insn): Likewise.
17893 (cmpsi_shiftsi): Likewise.
17894 (cmpsi_shiftsi_swp): Likewise.
17895 (arm_cmpsi_negshiftsi_si): Likewise.
17896 (movsicc_insn): Likewise.
17897 (movsfcc_soft_insn): Likewise.
17898 (arith_shiftsi): Likewise.
17899 (arith_shiftsi_compare0
17900 (arith_shiftsi_compare0_scratch
17901 (sub_shiftsi): Likewise.
17902 (sub_shiftsi_compare0
17903 (sub_shiftsi_compare0_scratch
17904 (and_scc): Likewise.
17905 (cond_move): Likewise.
17906 (if_plus_move): Likewise.
17907 (if_move_plus): Likewise.
17908 (if_move_not): Likewise.
17909 (if_not_move): Likewise.
17910 (if_shift_move): Likewise.
17911 (if_move_shift): Likewise.
17912 (if_shift_shift): Likewise.
17913 (if_not_arith): Likewise.
17914 (if_arith_not): Likewise.
17915 (cond_move_not): Likewise.
17916 (thumb1_ashlsi3): Set type attribute.
17917 (thumb1_ashrsi3): Likewise.
17918 (thumb1_lshrsi3): Likewise.
17919 (thumb1_rotrsi3): Likewise.
17920 (shiftsi3_compare0_scratch): Likewise.
17921 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
17922 (neon_mov<mode>): Likewise.
17923 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
17925 (thumb2_movsi_insn): Likewise.
17926 (thumb2_cmpsi_neg_shiftsi): Likewise.
17927 (thumb2_extendqisi_v6): Likewise.
17928 (thumb2_zero_extendhisi2_v6): Likewise.
17929 (thumb2_zero_extendqisi2_v6): Likewise.
17930 (thumb2_shiftsi3_short): Likewise.
17931 (thumb2_addsi3_compare0_scratch): Likewise.
17932 (orsi_not_shiftsi_si): Likewise.
17933 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
17934 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
17936 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
17937 (1020alu_shift_op): Likewise.
17938 (1020alu_shift_reg_op): Likewise.
17939 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
17940 (alu_shift_op): Likewise.
17941 (alu_shift_reg_op): Likewise.
17942 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
17943 (11_alu_shift_op): Likewise.
17944 (11_alu_shift_reg_op): Likewise.
17945 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
17946 (9_alu_shift_reg_op): Likewise.
17947 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17949 (cortex_a15_alu_shift): Likewise.
17950 (cortex_a15_alu_shift_reg): Likewise.
17951 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
17953 (cortex_a5_alu_shift): Likewise.
17954 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
17956 (cortex_a53_alu_shift): Likewise.
17957 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17959 (cortex_a7_alu_reg): Likewise.
17960 (cortex_a7_alu_shift): Likewise.
17961 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
17963 (cortex_a8_alu_shift): Likewise.
17964 (cortex_a8_alu_shift_reg): Likewise.
17965 (cortex_a8_mov): Likewise.
17966 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
17967 (cortex_a9_dp_shift): Likewise.
17968 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
17970 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
17972 (cortex_r4_mov): Likewise.
17973 (cortex_r4_alu_shift): Likewise.
17974 (cortex_r4_alu_shift_reg): Likewise.
17975 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
17976 (526_alu_shift_op): Likewise.
17977 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
17978 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
17979 (626te_alu_shift_op): Likewise.
17980 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
17981 (726te_alu_op): Likewise.
17982 (726te_alu_shift_op): Likewise.
17983 (726te_alu_shift_reg_op): Likewise.
17984 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
17985 (mp626_alu_shift_op): Likewise.
17986 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
17987 (pj4_alu_e1_conds): Likewise.
17988 (pj4_alu): Likewise.
17989 (pj4_alu_conds): Likewise.
17990 (pj4_shift): Likewise.
17991 (pj4_shift_conds): Likewise.
17992 (pj4_alu_shift): Likewise.
17993 (pj4_alu_shift_conds): Likewise.
17994 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
17996 (cortexa7_older_only): Likewise.
17997 (cortexa7_younger): Likewise.
17999 2013-07-18 David Malcolm <dmalcolm@redhat.com>
18001 * ipa-pure-const.c (generate_summary): Rename to...
18002 (pure_const_generate_summary): ... this.
18004 2013-07-17 Iain Sandoe <iain@codesourcery.com>
18006 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
18008 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
18011 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
18014 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18016 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
18017 enabled without -march=zEC12.
18018 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
18021 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
18023 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
18024 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
18025 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
18026 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
18027 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
18028 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
18029 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
18031 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
18032 and ISA_HAS_NMADD3_NMSUB3.
18033 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
18034 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
18035 (nmsub4<mode>, nmsub3<mode>): Likewise.
18036 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
18038 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
18040 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
18041 TARGET_MIPS5400 checking.
18043 2013-07-16 Jakub Jelinek <jakub@redhat.com>
18044 Peter Bergner <bergner@vnet.ibm.com>
18046 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
18047 registers in the comment.
18048 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
18049 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
18050 rather than FIRST_PSEUDO_REGISTERS.
18052 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
18054 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
18055 enable extra ISA flags with TARGET_HTM.
18057 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
18059 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
18062 2013-07-15 Cong Hou <congh@google.com>
18064 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
18065 in compare function for sorting.
18067 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
18069 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
18070 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
18071 * config/rs6000/rs6000.opt: Add -mhtm option.
18072 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
18073 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
18074 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
18075 __HTM__ if the HTM instructions are available.
18076 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
18077 htm_spr_reg_operand): New define_predicates.
18078 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
18079 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
18081 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
18082 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
18083 HTM builtin functions.
18084 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
18085 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
18086 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
18087 (rs6000_builtin_mask_calculate): Likewise.
18088 (rs6000_option_override_internal): Likewise.
18089 (bdesc_htm): Add new HTM builtin support.
18090 (htm_spr_num): New function.
18091 (htm_spr_regno): Likewise.
18092 (rs6000_htm_spr_icode): Likewise.
18093 (htm_expand_builtin): Likewise.
18094 (htm_init_builtins): Likewise.
18095 (rs6000_expand_builtin): Add support for HTM builtin functions.
18096 (rs6000_init_builtins): Likewise.
18097 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
18099 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
18100 (TARGET_HTM, MASK_HTM): Define macros.
18101 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
18102 (FIXED_REGISTERS): Likewise.
18103 (CALL_USED_REGISTERS): Likewise.
18104 (CALL_REALLY_USED_REGISTERS): Likewise.
18105 (REG_ALLOC_ORDER): Likewise.
18106 (enum reg_class): Likewise.
18107 (REG_CLASS_NAMES): Likewise.
18108 (REG_CLASS_CONTENTS): Likewise.
18109 (REGISTER_NAMES): Likewise.
18110 (ADDITIONAL_REGISTER_NAMES): Likewise.
18111 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
18112 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
18113 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
18114 * config/rs6000/htm.md: New file.
18115 * config/rs6000/htmintrin.h: New file.
18116 * config/rs6000/htmxlintrin.h: New file.
18118 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
18120 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
18121 Define SYMBOL_TINY_GOT, update comment.
18122 * config/aarch64/aarch64.c
18123 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
18124 (aarch64_expand_mov_immediate): Likewise.
18125 (aarch64_print_operand): Likewise.
18126 (aarch64_classify_symbol): Likewise.
18127 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
18128 (ldr_got_tiny): Define.
18130 2013-07-13 Tobias Grosser <tobias@grosser.es>
18132 PR tree-optimization/54094
18133 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
18134 scheduling dimension for the parallelism check from the polyhedral
18135 information in the AST.
18136 * graphite-dependences.c (carries_deps): Do not assume the schedule is
18139 2013-07-13 Jason Merrill <jason@redhat.com>
18141 * print-tree.c (debug_vec_tree): Use debug_raw.
18142 (debug_raw (vec<tree, va_gc> &)): New.
18143 (debug_raw (vec<tree, va_gc> *)): New.
18144 * tree.h: Declare them.
18146 2013-07-13 Bin Cheng <bin.cheng@arm.com>
18148 * ifcvt.c (ifcvt_after_combine): New static variable.
18149 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
18151 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
18152 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
18153 rest_of_handle_if_after_reload): Pass new argument for if_convert.
18155 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
18157 * config/mips/mips.c (mips_expand_call): Remove empty statement.
18159 2013-07-12 Michael Matz <matz@suse.de>
18161 PR middle-end/55771
18162 * convert.c (convert_to_real): Reject non-float inner types.
18164 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
18166 * config/aarch64/aarch64-protos.h
18167 (aarch64_simd_immediate_valid_for_move): Remove.
18168 * config/aarch64/aarch64.c (simd_immediate_info): New member.
18169 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
18171 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
18173 2013-07-11 Steve Ellcey <sellcey@mips.com>
18175 * config/mips/mips.c (mips_conditional_register_usage): Do not
18176 use t[0-7] registers in MIPS16 mode when optimizing for size.
18178 2013-07-11 Sriraman Tallam <tmsriram@google.com>
18180 * config/i386/i386.c (dispatch_function_versions): Fix array
18181 indexing of function_version_info to match actual_versions.
18183 2013-07-11 Teresa Johnson <tejohnson@google.com>
18185 * vec.h (struct va_gc): Move release out-of-line.
18186 (va_gc::release): Call ggc_free on released vec.
18188 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
18190 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
18191 Require GOT register as additional operand in UNSPEC.
18192 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
18193 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
18194 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
18195 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
18196 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
18197 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
18198 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
18200 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
18203 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
18204 name seen by assembler/linker rather if available.
18206 2013-07-11 Andreas Schwab <schwab@suse.de>
18208 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
18210 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
18212 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
18214 2013-07-10 Joseph Myers <joseph@codesourcery.com>
18216 * doc/tm.texi.in: Move hook documentation to ....
18217 * target.def: ... here.
18219 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
18220 text on @hook line.
18221 * doc/tm.texi: Regenerate.
18223 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
18226 * doc/invoke.texi: Document -Wconditionally-supported.
18228 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
18231 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
18234 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
18237 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
18238 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
18239 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
18240 Remove duplicate devices.
18241 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
18242 * config/avr/t-multilib: Regenerate.
18243 * config/avr/avr-tables.opt: Regenerate.
18244 * doc/avr-mmcu.texi: Regenerate.
18246 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
18249 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
18251 2013-07-10 Graham Stott <graham.stott@btinternet.com>
18253 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
18254 the cost of MULT when optimizing for size.
18256 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18258 * config/cr16/cr16-protos.h: Don't include target.h.
18260 2013-07-09 Joseph Myers <joseph@codesourcery.com>
18262 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
18263 adjust register size for TDmode and TFmode for VSX registers.
18265 2013-07-08 Kai Tietz <ktietz@redhat.com>
18268 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
18269 hook_bool_const_tree_true.
18271 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18273 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
18274 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
18275 * config/s390/s390.md: Define FPR*_REGNUM constants.
18276 Fix FPR2_REGNUM constant (18 -> 17).
18277 ("*trunc<BFP:mode><DFP_ALL:mode>2")
18278 ("*trunc<DFP_ALL:mode><BFP:mode>2")
18279 ("trunc<BFP:mode><DFP_ALL:mode>2")
18280 ("trunc<DFP_ALL:mode><BFP:mode>2")
18281 ("*extend<BFP:mode><DFP_ALL:mode>2")
18282 ("*extend<DFP_ALL:mode><BFP:mode>2")
18283 ("extend<BFP:mode><DFP_ALL:mode>2")
18284 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
18287 2013-07-08 Graham Stott <graham.stott@btinternet.com>
18289 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
18291 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18293 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
18294 and cfun_fpr_bit_p to cfun_fpr_save_p.
18295 (s390_frame_area, s390_register_info, s390_frame_info)
18296 (s390_emit_prologue, s390_emit_epilogue)
18297 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
18299 * config/s390/s390.h: Define *_REGNUM macros for floating point
18302 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
18304 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
18306 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
18308 PR rtl-optimization/57786
18309 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
18310 and break out of the loop when it is set to false.
18312 2013-07-08 Jakub Jelinek <jakub@redhat.com>
18315 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
18316 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
18317 (const_int 63)) 0)).
18318 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
18319 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
18320 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
18322 PR rtl-optimization/57829
18323 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
18324 mask bits outside of mode are just sign-extension from mode to HWI.
18326 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
18328 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
18329 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
18330 adjust_address instead of change_address to keep info about alignment.
18331 (emit_strmov): Remove.
18332 (emit_memmov): New function.
18333 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
18334 (expand_movmem_epilogue): Likewise and return updated rtx for
18336 (expand_constant_movmem_prologue): Likewise and return updated rtx for
18337 destination and source.
18338 (decide_alignment): Refactor, handle vector_loop.
18339 (ix86_expand_movmem): Likewise.
18340 (ix86_expand_setmem): Likewise.
18341 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
18343 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
18345 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
18346 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
18348 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
18350 * config/i386/sse.md (sse_movlhps): Change alternative 3
18351 of operand 2 to "m".
18353 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
18356 * config/i386/sse.md (iptr): New mode attribute.
18357 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
18358 (<sse>_vm<plusminus_insn><mode>3): Ditto.
18359 (<sse>_vmmul<mode>3): Ditto.
18360 (<sse>_vmdiv<mode>3): Ditto.
18361 (sse_vmrcpv4sf2): Ditto.
18362 (<sse>_vmsqrt<mode>2): Ditto.
18363 (sse_vmrsqrtv4sf2): Ditto.
18364 (<sse>_vm<code><mode>3): Ditto.
18365 (avx_vmcmp<mode>3): Ditto.
18366 (<sse>_vmmaskcmp<mode>3): Ditto.
18367 (<sse>_comi): Ditto.
18368 (<sse>_ucomi): Ditto.
18369 (*xop_vmfrcz_<mode>): Ditto.
18370 (*fmai_fmadd_<mode>): Ditto.
18371 (*fmai_fmsub_<mode>): Ditto.
18372 (*fmai_fnmadd_<mode>): Ditto.
18373 (*fmai_fnmsub_<mode>): Ditto.
18374 (*fma4i_vmfmadd_<mode>): Ditto.
18375 (*fma4i_vmfmsub_<mode>): Ditto.
18376 (*fma4i_vmfnmadd_<mode>): Ditto.
18377 (*fma4i_vmfnmsub_<mode>): Ditto.
18378 (*xop_vmfrcz_<mode>): Ditto.
18379 (sse_cvtps2pi): Ditto.
18380 (sse_cvttps2pi): Ditto.
18381 (sse_cvtss2si): Ditto.
18382 (sse_cvtss2si_2): Ditto.
18383 (sse_cvtss2siq_2): Ditto.
18384 (sse_cvttss2si): Ditto.
18385 (sse_cvttss2siq): Ditto.
18386 (sse_cvtsd2si): Ditto.
18387 (sse_cvtsd2si_2): Ditto.
18388 (sse_cvtsd2siq_2): Ditto.
18389 (sse_cvttsd2si): Ditto.
18390 (sse_cvttsd2siq): Ditto.
18391 (sse_cvtsd2ss): Ditto.
18392 (sse_cvtss2sd): Ditto.
18393 (avx2_pbroadcast<mode>): Ditto.
18394 (avx2_pbroadcast<mode>_1): Ditto.
18395 (*avx_vperm_broadcast_v4sf): Ditto.
18397 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
18398 (sse_movlhps): Ditto.
18399 (sse_storehps): Ditto.
18400 (sse_loadhps): Ditto.
18401 (sse_storelps): Ditto.
18402 (sse_loadlps): Ditto.
18403 (*vec_concatv4sf): Ditto.
18404 (*vec_interleave_highv2df): Ditto.
18405 (*vec_interleave_lowv2df): Ditto.
18406 (*vec_extractv2df_1_sse): Ditto.
18407 (*vec_extractv2df_0_sse): Ditto.
18408 (sse2_storelpd): Ditto.
18409 (sse2_loadlpd): Ditto.
18410 (sse2_movsd): Ditto.
18411 (*vec_concatv4si): Ditto.
18412 (vec_concatv2di): Ditto.
18414 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
18415 for Intel asm dialect.
18416 (mmx_punpcklwd): Ditto.
18417 (mmx_punpckldq): Ditto.
18419 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
18420 for intel assembler dialect.
18422 2013-07-06 Jakub Jelinek <jakub@redhat.com>
18425 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
18426 for BUILT_IN_C{LZ,LRSB}*.
18427 * tree.h (CASE_INT_FN): Add FN##IMAX case.
18428 * tree-vrp.c (extract_range_basic): Handle
18429 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
18430 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
18431 fall thru to code calling set_value*.
18432 * builtins.c (expand_builtin): Remove *IMAX cases.
18433 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
18434 if width is bigger than 2*HWI.
18436 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
18438 PR rtl-optimization/55342
18439 * lra-int.h (lra_subreg_reload_pseudos): New.
18440 * lra.c: Add undoing optional reloads to the block diagram.
18441 (lra_subreg_reload_pseudos): New.
18442 (lra_optional_reload_pseudos): Change comments.
18443 (lra): Init and clear lra_subreg_reload_pseudos. Clear
18444 lra_optional_reload_pseudos after undo transformations.
18445 * lra-assigns.c (pseudo_prefix_title): New.
18446 (lra_setup_reg_renumber): Use it.
18447 (spill_for): Ditto. Check subreg reload pseudos too.
18448 (assign_by_spills): Consider subreg reload pseudos too.
18449 * lra-constraints.c (simplify_operand_subreg): Use
18450 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
18451 (curr_insn_transform): Recognize and do optional reloads.
18452 (undo_optional_reloads): New.
18453 (lra_undo_inheritance): Call undo_optional_reloads.
18455 2013-07-05 Thomas Quinot <quinot@adacore.com>
18457 * tree-complex.c (expand_complex_operations_1): Fix typo.
18459 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
18461 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
18462 (tune_params): New member 'const vec_costs'.
18463 * config/aarch64/aarch64.c (generic_vector_cost): New.
18464 (generic_tunings): New member 'generic_vector_cost'.
18465 (aarch64_builtin_vectorization_cost): New.
18466 (aarch64_add_stmt_cost): New.
18467 (TARGET_VECTORIZE_ADD_STMT_COST): New.
18468 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
18470 2013-07-03 Jakub Jelinek <jakub@redhat.com>
18473 * config/i386/predicates.md (vsib_address_operand): Disallow
18474 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
18476 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
18478 PR middle-end/55030
18479 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
18480 expand_builtin_setjmp_receiver.
18481 (expand_label): Adjust, call expand_builtin_setjmp_receiver
18482 with NULL for the label parameter.
18483 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
18484 the frame-pointer. Adjust comments.
18485 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
18486 only if LABEL is non-NULL.
18488 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
18490 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
18491 (ARM_ABI_AAPCS64): Ditto.
18493 (ARM_DEFAULT_ABI): Ditto.
18495 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
18497 * config/aarch64/aarch64-builtins.c
18498 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
18499 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
18501 * config/aarch64/aarch64-simd.md
18502 (aarch64_ld1<VALL:mode>): New.
18503 (aarch64_st1<VALL:mode>): Likewise.
18504 * config/aarch64/arm_neon.h
18505 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
18507 2013-07-02 Sriraman Tallam <tmsriram@google.com>
18509 * config/i386/i386.c (gate_insert_vzeroupper): Check if
18511 (ix86_option_override_internal):Turn on all -mavx target flags by
18512 default as they are dependent on AVX anyway.
18514 2013-07-02 Cary Coutant <ccoutant@google.com>
18516 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
18517 deterministic hash.
18518 (loc_checksum_ordered): Likewise.
18519 (hash_loc_operands): Remove inline keyword.
18521 2013-07-02 Jakub Jelinek <jakub@redhat.com>
18523 PR tree-optimization/57741
18524 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
18525 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
18526 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
18527 Allow REAL_CST step_exprs if flag_associative_math.
18528 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
18530 2013-07-02 Ian Bolton <ian.bolton@arm.com>
18532 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
18534 2013-07-02 Ian Bolton <ian.bolton@arm.com>
18536 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
18538 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18540 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
18542 (iorsi3_insn): Likewise.
18543 (arm_xorsi3): Likewise.
18545 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
18547 * arm.md (attribute "wtype"): Delete. Move attribute values from here
18549 (attribute "type"): ... here, and prefix with "wmmx_".
18550 (attribute "core_cycles"): Update for attribute changes.
18551 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
18552 (tbcstv4hi): Likewise.
18553 (tbcstv2si): Likewise.
18554 (iwmmxt_iordi3): Likewise.
18555 (iwmmxt_xordi3): Likewise.
18556 (iwmmxt_anddi3): Likewise.
18557 (iwmmxt_nanddi3): Likewise.
18558 (iwmmxt_arm_movdi): Likewise.
18559 (iwmmxt_movsi_insn): Likewise.
18560 (mov<mode>_internal): Likewise.
18561 (and<mode>3_iwmmxt): Likewise.
18562 (ior<mode>3_iwmmxt): Likewise.
18563 (xor<mode>3_iwmmxt): Likewise.
18564 (add<mode>3_iwmmxt): Likewise.
18565 (ssaddv8qi3): Likewise.
18566 (ssaddv4hi3): Likewise.
18567 (ssaddv2si3): Likewise.
18568 (usaddv8qi3): Likewise.
18569 (usaddv4hi3): Likewise.
18570 (usaddv2si3): Likewise.
18571 (sub<mode>3_iwmmxt): Likewise.
18572 (sssubv8qi3): Likewise.
18573 (sssubv4hi3): Likewise.
18574 (sssubv2si3): Likewise.
18575 (ussubv8qi3): Likewise.
18576 (ussubv4hi3): Likewise.
18577 (ussubv2si3): Likewise.
18578 (mulv4hi3_iwmmxt): Likewise.
18579 (smulv4hi3_highpart): Likewise.
18580 (umulv4hi3_highpart): Likewise.
18581 (iwmmxt_wmacs): Likewise.
18582 (iwmmxt_wmacsz): Likewise.
18583 (iwmmxt_wmacu): Likewise.
18584 (iwmmxt_wmacuz): Likewise.
18585 (iwmmxt_clrdi): Likewise.
18586 (iwmmxt_clrv8qi): Likewise.
18587 (iwmmxt_clr4hi): Likewise.
18588 (iwmmxt_clr2si): Likewise.
18589 (iwmmxt_uavgrndv8qi3): Likewise.
18590 (iwmmxt_uavgrndv4hi3): Likewise.
18591 (iwmmxt_uavgv8qi3): Likewise.
18592 (iwmmxt_uavgv4hi3): Likewise.
18593 (iwmmxt_tinsrb): Likewise.
18594 (iwmmxt_tinsrh): Likewise.
18595 (iwmmxt_tinsrw): Likewise.
18596 (iwmmxt_textrmub): Likewise.
18597 (iwmmxt_textrmsb): Likewise.
18598 (iwmmxt_textrmuh): Likewise.
18599 (iwmmxt_textrmsh): Likewise.
18600 (iwmmxt_textrmw): Likewise.
18601 (iwmxxt_wshufh): Likewise.
18602 (eqv8qi3): Likewise.
18603 (eqv4hi3): Likewise.
18604 (eqv2si3): Likewise.
18605 (gtuv8qi3): Likewise.
18606 (gtuv4hi3): Likewise.
18607 (gtuv2si3): Likewise.
18608 (gtv8qi3): Likewise.
18609 (gtv4hi3): Likewise.
18610 (gtv2si3): Likewise.
18611 (smax<mode>3_iwmmxt): Likewise.
18612 (umax<mode>3_iwmmxt): Likewise.
18613 (smin<mode>3_iwmmxt): Likewise.
18614 (umin<mode>3_iwmmxt): Likewise.
18615 (iwmmxt_wpackhss): Likewise.
18616 (iwmmxt_wpackwss): Likewise.
18617 (iwmmxt_wpackdss): Likewise.
18618 (iwmmxt_wpackhus): Likewise.
18619 (iwmmxt_wpackwus): Likewise.
18620 (iwmmxt_wpackdus): Likewise.
18621 (iwmmxt_wunpckihb): Likewise.
18622 (iwmmxt_wunpckihh): Likewise.
18623 (iwmmxt_wunpckihw): Likewise.
18624 (iwmmxt_wunpckilb): Likewise.
18625 (iwmmxt_wunpckilh): Likewise.
18626 (iwmmxt_wunpckilw): Likewise.
18627 (iwmmxt_wunpckehub): Likewise.
18628 (iwmmxt_wunpckehuh): Likewise.
18629 (iwmmxt_wunpckehuw): Likewise.
18630 (iwmmxt_wunpckehsb): Likewise.
18631 (iwmmxt_wunpckehsh): Likewise.
18632 (iwmmxt_wunpckehsw): Likewise.
18633 (iwmmxt_wunpckelub): Likewise.
18634 (iwmmxt_wunpckeluh): Likewise.
18635 (iwmmxt_wunpckeluw): Likewise.
18636 (iwmmxt_wunpckelsb): Likewise.
18637 (iwmmxt_wunpckelsh): Likewise.
18638 (iwmmxt_wunpckelsw): Likewise.
18639 (ror<mode>3): Likewise.
18640 (ashr<mode>3_iwmmxt): Likewise.
18641 (lshr<mode>3_iwmmxt): Likewise.
18642 (ashl<mode>3_iwmmxt): Likewise.
18643 (ror<mode>3_di): Likewise.
18644 (ashr<mode>3_di): Likewise.
18645 (lshr<mode>3_di): Likewise.
18646 (ashl<mode>3_di): Likewise.
18647 (iwmmxt_wmadds): Likewise.
18648 (iwmmxt_wmaddu): Likewise.
18649 (iwmmxt_tmia): Likewise.
18650 (iwmmxt_tmiaph): Likewise.
18651 (iwmmxt_tmiabb): Likewise.
18652 (iwmmxt_tmiatb): Likewise.
18653 (iwmmxt_tmiabt): Likewise.
18654 (iwmmxt_tmiatt): Likewise.
18655 (iwmmxt_tmovmskb): Likewise.
18656 (iwmmxt_tmovmskh): Likewise.
18657 (iwmmxt_tmovmskw): Likewise.
18658 (iwmmxt_waccb): Likewise.
18659 (iwmmxt_wacch): Likewise.
18660 (iwmmxt_waccw): Likewise.
18661 (iwmmxt_waligni): Likewise.
18662 (iwmmxt_walignr): Likewise.
18663 (iwmmxt_walignr0): Likewise.
18664 (iwmmxt_walignr1): Likewise.
18665 (iwmmxt_walignr2): Likewise.
18666 (iwmmxt_walignr3): Likewise.
18667 (iwmmxt_wsadb): Likewise.
18668 (iwmmxt_wsadh): Likewise.
18669 (iwmmxt_wsadbz): Likewise.
18670 (iwmmxt_wsadhz): Likewise.
18671 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
18672 (iwmmxt_wabsdiffb): Likewise.
18673 (iwmmxt_wabsdiffh): Likewise.
18674 (iwmmxt_wabsdiffw): Likewise.
18675 (iwmmxt_waddsubhx): Likewise
18676 (iwmmxt_wsubaddhx): Likewise.
18677 (addc<mode>3): Likewise.
18678 (iwmmxt_avg4): Likewise.
18679 (iwmmxt_avg4r): Likewise.
18680 (iwmmxt_wmaddsx): Likewise.
18681 (iwmmxt_wmaddux): Likewise.
18682 (iwmmxt_wmaddsn): Likewise.
18683 (iwmmxt_wmaddun): Likewise.
18684 (iwmmxt_wmulwsm): Likewise.
18685 (iwmmxt_wmulwum): Likewise.
18686 (iwmmxt_wmulsmr): Likewise.
18687 (iwmmxt_wmulumr): Likewise.
18688 (iwmmxt_wmulwsmr): Likewise.
18689 (iwmmxt_wmulwumr): Likewise.
18690 (iwmmxt_wmulwl): Likewise.
18691 (iwmmxt_wqmulm): Likewise.
18692 (iwmmxt_wqmulwm): Likewise.
18693 (iwmmxt_wqmulmr): Likewise.
18694 (iwmmxt_wqmulwmr): Likewise.
18695 (iwmmxt_waddbhusm): Likewise.
18696 (iwmmxt_waddbhusl): Likewise.
18697 (iwmmxt_wqmiabb): Likewise.
18698 (iwmmxt_wqmiabt): Likewise.
18699 (iwmmxt_wqmiatb): Likewise.
18700 (iwmmxt_wqmiatt): Likewise.
18701 (iwmmxt_wqmiabbn): Likewise.
18702 (iwmmxt_wqmiabtn): Likewise.
18703 (iwmmxt_wqmiatbn): Likewise.
18704 (iwmmxt_wqmiattn): Likewise.
18705 (iwmmxt_wmiabb): Likewise.
18706 (iwmmxt_wmiabt): Likewise.
18707 (iwmmxt_wmiatb): Likewise.
18708 (iwmmxt_wmiatt): Likewise.
18709 (iwmmxt_wmiabbn): Likewise.
18710 (iwmmxt_wmiabtn): Likewise.
18711 (iwmmxt_wmiatbn): Likewise.
18712 (iwmmxt_wmiattn): Likewise.
18713 (iwmmxt_wmiawbb): Likewise.
18714 (iwmmxt_wmiawbt): Likewise.
18715 (iwmmxt_wmiawtb): Likewise.
18716 (iwmmxt_wmiawtt): Likewise.
18717 (iwmmxt_wmiawbbn): Likewise.
18718 (iwmmxt_wmiawbtn): Likewise.
18719 (iwmmxt_wmiawtbn): Likewise.
18720 (iwmmxt_wmiawttn): Likewise.
18721 (iwmmxt_wmerge): Likewise.
18722 (iwmmxt_tandc<mode>3): Likewise.
18723 (iwmmxt_torc<mode>3): Likewise.
18724 (iwmmxt_torvsc<mode>3): Likewise.
18725 (iwmmxt_textrc<mode>3): Likewise.
18726 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
18727 (wmmxt_pack): Likewise.
18728 (wmmxt_mult_c1): Likewise.
18729 (wmmxt_mult_c2): Likewise.
18730 (wmmxt_alu_c1): Likewise.
18731 (wmmxt_alu_c2): Likewise.
18732 (wmmxt_alu_c3): Likewise.
18733 (wmmxt_transfer_c1): Likewise.
18734 (wmmxt_transfer_c2): Likewise.
18735 (wmmxt_transfer_c3): Likewise.
18736 (marvell_f_iwmmxt_wstr): Likewise.
18737 (marvell_f_iwmmxt_wldr): Likewise.
18739 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
18741 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
18743 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18746 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18747 * lra-constraints.c (need_for_split_p): Check call used hard regs
18748 living through calls.
18750 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
18751 call used regs for call insn.
18753 2013-06-28 Jakub Jelinek <jakub@redhat.com>
18756 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
18757 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
18759 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18761 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
18763 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18765 * lra-constraints.c (need_for_split_p): Check call used hard regs
18766 living through calls.
18768 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
18771 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
18772 to tie with any other modes. Eliminate Altivec vector mode tests,
18773 since these are a subset of ALTIVEC or VSX vector modes. Simplify
18774 code, to return 0 if testing MODE2 for a condition, if we've
18775 already tested MODE1 for the same condition.
18777 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18779 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
18782 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18784 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
18785 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
18787 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18789 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
18791 (aarch64_symbolic_constant_p): Remove.
18792 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
18793 static. Fix line length and white space.
18794 (aarch64_symbolic_constant_p): Remove.
18795 * config/aarch64/predicates.md (aarch64_valid_symref):
18796 Use aarch64_classify_symbol_expression.
18798 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18800 * config/arm/constraints.md (Ts): New constraint.
18801 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
18803 (compare_scc): Use "Ts" constraint for operand 0.
18804 (ior_scc_scc): Likewise.
18805 (and_scc_scc): Likewise.
18806 (and_scc_scc_nodom): Likewise.
18807 (ior_scc_scc_cmp): Likewise for operand 7.
18808 (and_scc_scc_cmp): Likewise.
18809 * config/arm/thumb2.md (thumb2_movsi_insn):
18810 Add alternatives for 16-bit encodings.
18811 (thumb2_movhi_insn): Likewise.
18812 (thumb2_movsicc_insn): Likewise.
18813 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
18814 (thumb2_negscc): Use "Ts" constraint.
18815 Move mvn instruction outside cond_exec block.
18816 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
18817 for 16-bit encodings.
18819 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18821 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
18823 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
18824 (mulsi3subsi): Likewise.
18825 (mulsidi3adddi): Likewise.
18826 (mulsidi3_v6): Likewise.
18827 (umulsidi3_v6): Likewise.
18828 (umulsidi3adddi_v6): Likewise.
18829 (smulsi3_highpart_v6): Likewise.
18830 (umulsi3_highpart_v6): Likewise.
18831 (mulhisi3tb): Likewise.
18832 (mulhisi3bt): Likewise.
18833 (mulhisi3tt): Likewise.
18834 (maddhisi4): Likewise.
18835 (maddhisi4tb): Likewise.
18836 (maddhisi4tt): Likewise.
18837 (maddhidi4): Likewise.
18838 (maddhidi4tb): Likewise.
18839 (maddhidi4tt): Likewise.
18840 (zeroextractsi_compare0_scratch): Likewise.
18841 (insv_zero): Likewise.
18842 (insv_t2): Likewise.
18843 (anddi_notzesidi_di): Likewise.
18844 (anddi_notsesidi_di): Likewise.
18845 (andsi_notsi_si): Likewise.
18846 (iordi_zesidi_di): Likewise.
18847 (xordi_zesidi_di): Likewise.
18848 (andsi_iorsi3_notsi): Likewise.
18849 (smax_0): Likewise.
18850 (smax_m1): Likewise.
18851 (smin_0): Likewise.
18852 (not_shiftsi): Likewise.
18853 (unaligned_loadsi): Likewise.
18854 (unaligned_loadhis): Likewise.
18855 (unaligned_loadhiu): Likewise.
18856 (unaligned_storesi): Likewise.
18857 (unaligned_storehi): Likewise.
18858 (extv_reg): Likewise.
18859 (extzv_t2): Likewise.
18860 (divsi3): Likewise.
18861 (udivsi3): Likewise.
18862 (arm_zero_extendhisi2addsi): Likewise.
18863 (arm_zero_extendqisi2addsi): Likewise.
18864 (compareqi_eq0): Likewise.
18865 (arm_extendhisi2_v6): Likewise.
18866 (arm_extendqisi2addsi): Likewise.
18867 (arm_movt): Likewise.
18868 (thumb2_ldrd): Likewise.
18869 (thumb2_ldrd_base): Likewise.
18870 (thumb2_ldrd_base_neg): Likewise.
18871 (thumb2_strd): Likewise.
18872 (thumb2_strd_base): Likewise.
18873 (thumb2_strd_base_neg): Likewise.
18874 (arm_negsi2): Add alternative for 16-bit encoding.
18875 (arm_one_cmplsi2): Likewise.
18877 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18879 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
18880 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
18881 (movdfcc): Likewise.
18882 * config/arm/vfp.md (*thumb2_movsf_vfp):
18883 Disable predication for arm_restrict_it.
18884 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
18885 (*thumb2_movdfcc_vfp): Likewise.
18886 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
18887 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
18888 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
18889 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
18890 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
18891 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
18892 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
18893 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
18894 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
18895 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
18896 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
18897 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
18898 Disable predication for arm_restrict_it.
18900 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
18902 * config/i386/bmiintrin.h (_bextr_u32): New.
18903 (_bextr_u64): Ditto.
18905 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
18907 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
18908 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
18909 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
18910 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
18911 * config/mips/n32-elf.h: ...this new file.
18913 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
18916 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
18917 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
18919 2013-06-27 Catherine Moore <clm@codesourcery.com>
18921 * config/mips/mips-tables.opt: Regenerate.
18922 * config/mips/mips-cpus.def: Add m14ke and m14kec.
18923 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
18924 * doc/invoke.texi: Add -m14kc.
18926 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18929 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
18930 constraints of operand 1 and 2.
18933 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
18934 to match RTL canonicalization. Swap predicates and
18935 constraints of operand 1 and 2.
18937 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
18939 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
18940 Process OP_INOUT regs for splitting too.
18942 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18944 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
18945 decl before the loop, initialize to NULL.
18946 (vectorizable_load): Initialize ptr_incr to NULL.
18948 2013-06-27 Martin Jambor <mjambor@suse.cz>
18951 * ipa-ref.h (ipa_maybe_record_reference): Declare.
18952 * ipa-ref.c (ipa_maybe_record_reference): New function.
18953 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
18954 * ipa-cp.c (create_specialized_node): Record potential references from
18956 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
18958 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18960 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
18961 parameter 'mode' of type 'enum machine_mode mode'; change to pass
18962 'mode' to force_reg.
18963 (aarch64_add_offset): Update calls to aarch64_force_temporary.
18964 (aarch64_expand_mov_immediate): Likewise.
18966 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18968 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
18969 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
18971 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18973 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
18974 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18975 (struct machine_function): Add tbegin_p.
18976 (s390_canonicalize_comparison): Fold CC mode compares to
18977 conditional jump if possible.
18978 (s390_emit_jump): Return the emitted jump.
18979 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
18980 Handle CCRAWmode compares.
18981 (s390_option_override): Default to -mhtm if available.
18982 (s390_reg_clobbered_rtx): Handle floating point regs as well.
18983 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
18984 FPRs instead of df_regs_ever_live_p.
18985 (s390_optimize_nonescaping_tx): New function.
18986 (s390_init_frame_layout): Extend clobbered_regs array to cover
18988 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
18989 (s390_expand_tbegin): New function.
18990 (enum s390_builtin): New enum definition.
18991 (code_for_builtin): New array definition.
18992 (s390_init_builtins): New function.
18993 (s390_expand_builtin): New function.
18994 (TARGET_INIT_BUILTINS): Define.
18995 (TARGET_EXPAND_BUILTIN): Define.
18996 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
18997 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
18998 (s390_alc_comparison): Likewise.
18999 * config/s390/s390-modes.def: Add CCRAWmode.
19000 * config/s390/s390.h (processor_flags): Add PF_TX.
19001 (TARGET_CPU_HTM): Define macro.
19002 (TARGET_HTM): Define macro.
19003 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
19004 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
19005 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
19006 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
19007 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
19008 (TBEGIN_MASK, TBEGINC_MASK): New constants.
19009 ("*cc_to_int"): Move up.
19010 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
19011 constants other than 0.
19012 ("*ccraw_to_int"): New insn and splitter definition.
19013 ("tbegin", "tbegin_nofloat", "tbegin_retry")
19014 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
19015 ("tx_assist"): New expander.
19016 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
19017 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
19018 * config/s390/s390.opt: Add -mhtm option.
19019 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
19020 * config/s390/htmxlintrin.h: New file.
19021 * config/s390/htmintrin.h: New file.
19022 * config/s390/s390intrin.h: New file.
19023 * doc/extend.texi: Document htm builtins.
19024 * config.gcc: Add the new header files to extra_headers.
19026 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
19028 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
19029 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
19031 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
19032 Pat Haugen <pthaugen@us.ibm.com>
19033 Peter Bergner <bergner@vnet.ibm.com>
19035 * config/rs6000/power8.md: New.
19036 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
19037 setting for power8 entry.
19038 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
19039 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
19040 test for Power4/Power5 only.
19041 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
19043 (force_new_group): Adjust comment.
19044 * config/rs6000/rs6000.md: Include power8.md.
19046 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
19048 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
19049 * config/arm/arm-protos.h (arm_max_conditional_execute): New
19051 (tune_params): Update comment.
19052 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
19053 (arm_max_conditional_execute): New function.
19054 (thumb2_final_prescan_insn): Use max_insn_skipped and
19055 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
19057 2013-06-25 Jakub Jelinek <jakub@redhat.com>
19059 PR tree-optimization/57705
19060 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
19061 SSA_NAME step, provided that it is not defined inside the loop.
19062 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
19063 (get_initial_def_for_induction): Handle SSA_NAME IV step.
19065 2013-06-25 Martin Jambor <mjambor@suse.cz>
19067 PR middle-end/57670
19068 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
19069 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
19071 (ipa_note_param_call): Initialize member_ptr flag.
19072 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
19073 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
19074 (ipa_write_indirect_edge_info): Stream member_ptr flag.
19075 (ipa_read_indirect_edge_info): Likewise.
19077 2013-06-25 Richard Biener <rguenther@suse.de>
19079 PR middle-end/56977
19080 * passes.c (init_optimization_passes): Move pass_fold_builtins
19081 and pass_dce earlier with -Og.
19083 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
19085 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
19086 <BIT_FIELD_REF>: Remove trailing TAB.
19087 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
19090 2013-06-24 Martin Jambor <mjambor@suse.cz>
19092 PR tree-optimization/57358
19093 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
19094 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
19095 (ipa_analyze_params_uses): Generate pessimistic info when true.
19097 2013-06-24 Martin Jambor <mjambor@suse.cz>
19099 PR tree-optimization/57539
19100 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
19101 global.inlined_to of the new node to it. All callers changed.
19102 * ipa-inline-transform.c (clone_inlined_nodes): New variable
19103 inlining_into, pass it to cgraph_clone_node.
19104 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
19105 ipa_free_edge_args_substructures.
19106 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
19107 rdesc linked list. Do not assert rdesc edges have inlined caller.
19108 Assert we have found an rdesc in the rdesc list.
19110 2013-06-24 Richard Biener <rguenther@suse.de>
19112 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
19113 (pointer_set_lookup): Declare.
19114 (class pointer_map): New template class implementing a
19115 generic pointer to T map.
19116 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
19117 pointer_map<T>::contains, pointer_map<T>::insert,
19118 pointer_map<T>::traverse): New functions.
19119 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
19120 (pointer_set_lookup): New function.
19121 (pointer_set_contains): Use pointer_set_lookup.
19122 (pointer_set_insert): Likewise.
19123 (insert_aux): Remove.
19124 (struct pointer_map_t): Embed a pointer_set_t.
19125 (pointer_map_create): Adjust.
19126 (pointer_map_destroy): Likewise.
19127 (pointer_map_contains): Likewise.
19128 (pointer_map_insert): Likewise.
19129 (pointer_map_traverse): Likewise.
19130 * tree-streamer.h (struct streamer_tree_cache_d): Use a
19131 pointer_map<unsigned> instead of a pointer_map_t.
19132 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
19133 (streamer_tree_cache_lookup): Likewise.
19134 (streamer_tree_cache_create): Likewise.
19135 (streamer_tree_cache_delete): Likewise.
19136 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
19137 pointer_map<unsigned> instead of a pointer_map_t.
19138 (lto_init_tree_ref_encoder): Adjust.
19139 (lto_destroy_tree_ref_encoder): Likewise.
19140 * lto-section-out.c (lto_output_decl_index): Likewise.
19141 (lto_record_function_out_decl_state): Likewise.
19142 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
19144 2013-06-24 Richard Biener <rguenther@suse.de>
19146 PR tree-optimization/57488
19147 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
19149 2013-06-24 Alan Modra <amodra@gmail.com>
19151 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
19152 (gen_easy_altivec_constant): Likewise.
19153 * config/rs6000/predicates.md (easy_vector_constant_add_self,
19154 easy_vector_constant_msb): Likewise.
19156 2013-06-23 Jakub Jelinek <jakub@redhat.com>
19159 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
19160 add missing return true.
19162 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
19165 * config/sh/predicates.md (general_extend_operand): Invoke
19166 general_movsrc_operand for memory operands.
19167 (general_movsrc_operand): Allow reg+reg addressing, do not use
19168 general_operand for memory operands.
19170 2013-06-23 Sriraman Tallam <tmsriram@google.com>
19172 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
19173 when current target options does not apply.
19174 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
19175 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
19176 * config/i386/bmiintrin.h: Pass appropriate target
19177 attributes to header.
19178 * config/i386/mmintrin.h: Ditto.
19179 * config/i386/nmmintrin.h: Ditto.
19180 * config/i386/avx2intrin.h: Ditto.
19181 * config/i386/fxsrintrin.h: Ditto.
19182 * config/i386/tbmintrin.h: Ditto.
19183 * config/i386/xsaveintrin.h: Ditto.
19184 * config/i386/f16cintrin.h: Ditto.
19185 * config/i386/xtestintrin.h: Ditto.
19186 * config/i386/xsaveoptintrin.h: Ditto.
19187 * config/i386/bmi2intrin.h: Ditto.
19188 * config/i386/lzcntintrin.h: Ditto.
19189 * config/i386/smmintrin.h: Ditto.
19190 * config/i386/wmmintrin.h: Ditto.
19191 * config/i386/x86intrin.h: Remove all header include guards.
19192 * config/i386/prfchwintrin.h: Ditto.
19193 * config/i386/pmmintrin.h: Ditto.
19194 * config/i386/tmmintrin.h: Ditto.
19195 * config/i386/xmmintrin.h: Ditto.
19196 * config/i386/popcntintrin.h: Ditto.
19197 * config/i386/rdseedintrin.h: Ditto.
19198 * config/i386/ammintrin.h: Ditto.
19199 * config/i386/emmintrin.h: Ditto.
19200 * config/i386/immintrin.h: Remove all header include guards.
19201 * config/i386/fma4intrin.h: Ditto.
19202 * config/i386/lwpintrin.h: Ditto.
19203 * config/i386/xopintrin.h: Ditto.
19204 * config/i386/ia32intrin.h: Ditto.
19205 * config/i386/avxintrin.h: Ditto.
19206 * config/i386/rtmintrin.h: Ditto.
19207 * config/i386/fmaintrin.h: Ditto.
19208 * config/i386/mm3dnow.h: Ditto.
19210 2013-06-22 Sriraman Tallam <tmsriram@google.com>
19212 * common/config/i386/i386-common.c: Handle LZCNT.
19214 2013-06-22 Andi Kleen <ak@linux.intel.com>
19216 * doc/extend.texi: Use __atomic_store_n instead of
19217 __atomic_store in HLE example.
19219 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
19221 * config/sh/sh.c: Remove <cstdlib> workaround.
19223 2013-06-21 Andi Kleen <ak@linux.intel.com>
19225 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
19227 2013-06-21 Andi Kleen <ak@linux.intel.com>
19229 * doc/extend.texi: Document that __atomic_clear and
19230 __atomic_test_and_set should only be used with bool.
19232 2013-06-20 Jan Hubicka <jh@suse.cz>
19234 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
19235 types_same_for_odr.
19236 * tree.c (decls_same_for_odr): New function.
19237 (same_for_edr): New function.
19238 (types_same_for_odr): New function.
19239 (get_binfo_at_offset): Use it.
19240 * tree.h (types_same_for_odr): Declare.
19242 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
19243 Jason Merrill <jason@redhat.com>
19245 * system.h: Include <cstdlib> as well as <stdlib.h>.
19247 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
19250 * config/i386/i386.c (construct_container): Report error if
19251 long double is used with disabled x87 float returns.
19253 2013-06-20 Jan Hubicka <jh@suse.cz>
19255 * lto-cgraph.c (input_symtab): Do not set cgraph state.
19257 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
19259 PR rtl-optimization/57425
19260 PR rtl-optimization/57569
19261 * alias.c (write_dependence_p): Remove parameters mem_mode and
19262 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
19263 Changed all callers.
19264 (canon_anti_dependence): Get comments and semantics in sync.
19265 Add parameter mem_canonicalized. Changed all callers.
19266 * rtl.h (canon_anti_dependence): Update prototype.
19268 2013-06-20 Richard Biener <rguenther@suse.de>
19270 * data-streamer-in.c (streamer_read_uhwi): Optimize single
19271 byte case, inline streamer_read_uchar and defer section
19274 2013-06-20 Richard Biener <rguenther@suse.de>
19276 PR tree-optimization/57584
19277 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
19278 SSA names into the expanded expression that take part in
19279 abnormal coalescing.
19281 2013-06-19 Sharad Singhai <singhai@google.com>
19283 * gcov.c (print_usage): Handle new option.
19284 (process_args): Ditto.
19285 (get_gcov_intermediate_filename): New function.
19286 (output_intermediate_file): New function.
19287 (output_gcov_file): New function
19288 (generate_results): Handle new option.
19289 (release_function): Relase demangled name.
19290 (read_graph_file): Handle demangled name.
19291 (output_lines): Ditto.
19292 * doc/gcov.texi: Document gcov intermediate format.
19294 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
19297 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
19298 (lra_emit_add): Use the functions. Add comment about Y as an
19301 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
19304 * collect2.c (collect_atexit): New.
19305 (collect_exit): Delete.
19306 (main): Register collect_atexit with atexit.
19307 (collect_wait): Change collect_exit to exit.
19309 * collect2.h (collect_exit): Delete.
19310 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
19312 2013-06-19 Wei Mi <wmi@google.com>
19314 PR rtl-optimization/57518
19315 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
19316 if regno is used in paradoxical subreg.
19317 (update_equiv_regs): Check pdx_subregs[regno] before
19318 set a reg to be equivalent with a mem.
19320 2013-06-19 Matthias Klose <doko@ubuntu.com>
19323 * file-find.h (find_a_file): Add a mode parameter.
19324 * file-find.c (find_a_file): Likewise.
19325 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
19326 with X_OK for the executables.
19327 * collect2.c (main): Call find_a_file with X_OK.
19329 2013-06-19 Steve Ellcey <sellcey@mips.com>
19332 * config/mips/mips.md (casesi_internal_mips16_<mode>):
19333 Use NEXT_INSN instead of next_real_insn.
19335 2013-06-19 Jan Hubicka <jh@suse.cz>
19337 * cgraph.h (const_value_known_p): Replace by ...
19338 (ctor_for_folding): .. this one.
19339 * cgraphunit.c (process_function_and_variable_attributes): Use it.
19340 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
19341 * expr.c (expand_expr_real_1): Likewise.
19342 (string_constant): Likewise.
19343 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
19344 * ipa.c (process_references): Likewise.
19345 (symtab_remove_unreachable_nodes): Likewise.
19346 * ipa-inline-analysis.c (param_change_prob): Likewise.
19347 * gimple-fold.c (canonicalize_constructor_val): Likewise.
19348 (get_base_constructor): Likwise.
19349 * varpool.c (varpool_remove_node): Likewise.
19350 (varpool_remove_initializer): LIkewise.
19351 (dump_varpool_node): LIkwise.
19352 (const_value_known_p): Rewrite to ...
19353 (ctor_for_folding): ... this one.
19355 2013-06-19 Jakub Jelinek <jakub@redhat.com>
19358 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
19359 PERSONALITY in $PATH derived prefixes.
19361 2013-06-19 Jeff Law <law@redhat.com>
19363 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
19366 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
19367 (simplify_bitwise_binary): Use it to simpify certain binary ops on
19370 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
19372 * config/arm/vfp.md: Move VFP instruction classification documentation
19374 * config/arm/arm.md: ... here. Update instruction classification
19377 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
19379 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
19380 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
19381 pattern. Use more efficient sequences on ARMv5 and Thumb2.
19383 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
19386 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19387 with NEXT_INSN. Use tablejump_p to check for jump table data
19390 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
19393 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
19394 that now in C++ the value is correct per the C++ standards.
19396 2013-06-19 Richard Biener <rguenther@suse.de>
19398 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
19399 for global context.
19401 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19404 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19407 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19408 with next_active_insn.
19410 2013-06-18 Sriraman Tallam <tmsriram@google.com>
19412 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
19413 functions are inlined during failures to flag an error.
19414 * tree-inline.c (expand_call_inline): Allow the error to be flagged
19415 in early inline pass.
19417 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
19419 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
19422 2013-06-18 Julian Brown <julian@codesourcery.com>
19424 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
19425 Permit virtual register pre-reload if !strict.
19426 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
19428 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
19430 * config/arm/neon.md (movmisalign<mode>): Use
19431 neon_perm_struct_or_reg_operand instead of
19432 neon_struct_or_register_operand.
19433 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
19434 neon_permissive_struct_operand instead of neon_struct_operand.
19435 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
19436 neon_vector_mem_operand.
19437 * config/arm/predicates.md (neon_struct_operand): Adjust call to
19438 neon_vector_mem_operand.
19439 (neon_permissive_struct_operand): New.
19440 (neon_struct_or_register_operand): Rename to...
19441 (neon_perm_struct_or_reg_operand): This. Adjust call to
19442 neon_vector_mem_operand.
19444 2013-06-18 Richard Biener <rguenther@suse.de>
19446 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
19447 * lto-streamer.h: Include pointer-set.h.
19448 (struct lto_decl_slot): Remove.
19449 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
19450 Remove next_index entry.
19451 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
19452 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
19453 (lto_init_tree_ref_encoder): Adjust.
19454 (lto_destroy_tree_ref_encoder): Likewise.
19455 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
19456 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
19457 (lto_output_decl_index): Adjust.
19458 (lto_new_out_decl_state): Likewise.
19459 (lto_record_function_out_decl_state): Likewise.
19460 * lto-streamer-out.c (copy_function): Likewise.
19462 2013-06-18 Richard Biener <rguenther@suse.de>
19464 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
19465 * cgraphunit.c: Include cfgloop.h.
19466 (init_lowered_empty_function): Initialize the loop tree.
19467 (assemble_thunk): Insert new BBs into loops.
19469 2013-06-18 Richard Biener <rguenther@suse.de>
19471 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
19472 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
19473 the map from cache entry to cache index optional.
19474 (streamer_tree_cache_replace_tree): Adjust accordingly.
19475 (streamer_tree_cache_append): Likewise.
19476 (streamer_tree_cache_delete): Likewise.
19477 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
19478 streamer cache map from cache entry to cache index.
19479 * lto-streamer-out.c (create_output_block): Adjust.
19481 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
19483 * config/arm/arm.md (attribute "insn"): Move multiplication and
19484 division attributes to...
19485 (attribute "type"): ... here. Remove mult.
19486 (attribute "mul32"): New attribute.
19487 (attribute "mul64"): Add umaal.
19488 (*arm_mulsi3): Update attributes.
19489 (*arm_mulsi3_v6): Likewise.
19490 (*thumb_mulsi3): Likewise.
19491 (*thumb_mulsi3_v6): Likewise.
19492 (*mulsi3_compare0): Likewise.
19493 (*mulsi3_compare0_v6): Likewise.
19494 (*mulsi_compare0_scratch): Likewise.
19495 (*mulsi_compare0_scratch_v6): Likewise.
19496 (*mulsi3addsi): Likewise.
19497 (*mulsi3addsi_v6): Likewise.
19498 (*mulsi3addsi_compare0): Likewise.
19499 (*mulsi3addsi_compare0_v6): Likewise.
19500 (*mulsi3addsi_compare0_scratch): Likewise.
19501 (*mulsi3addsi_compare0_scratch_v6): Likewise.
19502 (*mulsi3subsi): Likewise.
19503 (*mulsidi3adddi): Likewise.
19504 (*mulsi3addsi_v6): Likewise.
19505 (*mulsidi3adddi_v6): Likewise.
19506 (*mulsidi3_nov6): Likewise.
19507 (*mulsidi3_v6): Likewise.
19508 (*umulsidi3_nov6): Likewise.
19509 (*umulsidi3_v6): Likewise.
19510 (*umulsidi3adddi): Likewise.
19511 (*umulsidi3adddi_v6): Likewise.
19512 (*smulsi3_highpart_nov6): Likewise.
19513 (*smulsi3_highpart_v6): Likewise.
19514 (*umulsi3_highpart_nov6): Likewise.
19515 (*umulsi3_highpart_v6): Likewise.
19516 (mulhisi3): Likewise.
19517 (*mulhisi3tb): Likewise.
19518 (*mulhisi3bt): Likewise.
19519 (*mulhisi3tt): Likewise.
19520 (maddhisi4): Likewise.
19521 (*maddhisi4tb): Likewise.
19522 (*maddhisi4tt): Likewise.
19523 (maddhidi4): Likewise.
19524 (*maddhidi4tb): Likewise.
19525 (*maddhidi4tt): Likewise.
19526 (divsi3): Likewise.
19527 (udivsi3): Likewise.
19528 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
19529 (thumb2_mulsi_short_compare0): Likewise.
19530 (thumb2_mulsi_short_compare0_scratch): Likewise.
19531 * config/arm/arm1020e.md (1020mult1): Update attribute change.
19532 (1020mult2): Likewise.
19533 (1020mult3): Likewise.
19534 (1020mult4): Likewise.
19535 (1020mult5): Likewise.
19536 (1020mult6): Likewise.
19537 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
19539 (cortex_a15_mult64): Likewise.
19540 (cortex_a15_sdiv): Likewise.
19541 (cortex_a15_udiv): Likewise.
19542 * config/arm/arm1026ejs.md (mult1): Update attribute change.
19548 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
19549 (pj4_ir_div): Likewise.
19550 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
19551 (11_mult2): Likewise.
19552 (11_mult3): Likewise.
19553 (11_mult4): Likewise.
19554 (11_mult5): Likewise.
19555 (11_mult6): Likewise.
19556 (11_mult7): Likewise.
19557 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
19558 (cortex_a8_mla): Likewise.
19559 (cortex_a8_mull): Likewise.
19560 (cortex_a8_smulwy): Likewise.
19561 (cortex_a8_smlald): Likewise.
19562 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
19563 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
19564 (cortex_r4_mul_3): Likewise.
19565 (cortex_r4_mla_4): Likewise.
19566 (cortex_r4_mla_3): Likewise.
19567 (cortex_r4_smlald): Likewise.
19568 (cortex_r4_mull): Likewise.
19569 (cortex_r4_sdiv): Likewise.
19570 (cortex_r4_udiv): Likewise.
19571 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
19572 (cortex_a7_idiv): Likewise.
19573 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
19574 (9_mult2): Likewise.
19575 (9_mult3): Likewise.
19576 (9_mult4): Likewise.
19577 (9_mult5): Likewise.
19578 (9_mult6): Likewise.
19579 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
19580 (cortex_a53_sdiv): Likewise.
19581 (cortex_a53_udiv): Likewise.
19582 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
19583 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
19584 (mp626_mult2): Likewise.
19585 (mp626_mult3): Likewise.
19586 (mp626_mult4): Likewise.
19587 * config/arm/fa526.md (526_mult1): Update attribute change.
19588 (526_mult2): Likewise.
19589 * config/arm/arm-generic.md (mult): Update attribute change.
19590 (mult_ldsched_strongarm): Likewise.
19591 (mult_ldsched): Likewise.
19592 (multi_cycle): Likewise.
19593 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
19594 * config/arm/fa606te.md (606te_mult1): Update attribute change.
19595 (606te_mult2): Likewise.
19596 (606te_mult3): Likewise.
19597 (606te_mult4): Likewise.
19598 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
19599 (cortex_a9_mac16): Likewise.
19600 (cortex_a9_multiply): Likewise.
19601 (cortex_a9_mac): Likewise.
19602 (cortex_a9_multiply_long): Likewise.
19603 * config/arm/fa626te.md (626te_mult1): Update attribute change.
19604 (626te_mult2): Likewise.
19605 (626te_mult3): Likewise.
19606 (626te_mult4): Likewise.
19608 2013-06-18 Richard Biener <rguenther@suse.de>
19611 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
19613 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19616 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19617 with next_active_insn.
19619 2013-06-18 Alan Modra <amodra@gmail.com>
19621 * config/rs6000/rs6000.h (enum data_align): New.
19622 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
19623 (DATA_ABI_ALIGNMENT): Define.
19624 (CONSTANT_ALIGNMENT): Correct comment.
19625 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
19626 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
19628 2013-06-17 David Malcolm <dmalcolm@redhat.com>
19630 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
19631 ATTRIBUTE_UNUSED marking.
19633 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
19635 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
19636 alternative and update.
19637 (aarch64_dup_lanedi): Delete.
19638 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
19639 * config/aarch64/aarch64-simd-builtins.def: Update.
19641 2013-06-17 Richard Biener <rguenther@suse.de>
19643 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
19644 (lto_input_scc): Declare.
19645 (lto_input_tree_1): Likewise.
19646 (struct lto_stats_d): Add num_tree_bodies_output and
19647 num_pickle_refs_output.
19648 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
19649 (lto_read_tree_1): Split out from ...
19650 (lto_read_tree): ... this.
19651 (lto_input_scc): New function.
19652 (lto_input_tree_1): Split out from ...
19653 (lto_input_tree): ... this. Handle LTO_tree_scc.
19654 (lto_data_in_create): Create the streamer cache without hashes.
19655 * lto-streamer-out.c (create_output_block): Create the streamer
19656 cache with hashes when not doing WPA.
19657 (lto_write_tree_1): Split out from ...
19658 (lto_write_tree): ... this.
19659 (get_symbol_initial_value): New function.
19660 (lto_output_tree_1): Split out from ...
19661 (lto_output_tree): ... this. Write trees as series of SCCs
19662 using a DFS walk via DFS_write_tree.
19663 (struct sccs, struct scc_entry): New types.
19664 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
19665 (DFS_write_tree_body): New function.
19666 (DFS_write_tree): Likewise.
19667 (hash_tree): Likewise.
19668 (scc_entry_compare): Likewise.
19669 (hash_scc): Likewise.
19670 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
19671 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
19672 TREE_CHAIN as regular reference.
19673 (streamer_read_integer_cst): Remove.
19674 (streamer_get_pickled_tree): Adjust.
19675 * tree-streamer-out.c (streamer_write_chain): Disable streaming
19676 of DECL_EXTERNALs in BLOCK_VARS for now.
19677 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
19679 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
19680 Add hash value argument and record that if hashes are recorded
19682 (streamer_tree_cache_insert_1): Adjust.
19683 (streamer_tree_cache_insert): Likewise.
19684 (streamer_tree_cache_insert_at): Rename to ...
19685 (streamer_tree_cache_replace_tree): ... this and adjust.
19686 (streamer_tree_cache_append): Adjust.
19687 (record_common_node): Likewise.
19688 (streamer_tree_cache_create): Add argument whether to
19689 record hash values together with trees.
19690 (streamer_tree_cache_delete): Adjust.
19691 * tree-streamer.h (struct streamer_tree_cache_d): Add
19693 (streamer_read_integer_cst): Remove.
19694 (streamer_tree_cache_insert): Adjust.
19695 (streamer_tree_cache_append): Likewise.
19696 (streamer_tree_cache_insert_at): Rename to ...
19697 (streamer_tree_cache_replace_tree): ... this and adjust.
19698 (streamer_tree_cache_create): Add argument whether to record hashes.
19699 (streamer_tree_cache_get): Rename to ...
19700 (streamer_tree_cache_get_tree): ... this.
19701 (streamer_tree_cache_get_hash): New function.
19702 * tree.c (cache_integer_cst): New function.
19703 * tree.h (cache_integer_cst): Declare.
19704 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
19705 * lto-symtab.c (lto_varpool_replace_node): Only release
19706 DECL_INITIAL of non-prevailing decls.
19707 * varpool.c (varpool_remove_initializer): Do not release
19708 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
19710 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
19712 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
19713 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
19714 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
19715 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
19716 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
19717 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
19718 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
19719 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
19720 instead of TARGET_64BIT.
19721 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
19722 Require ISA_HAS_<D>DIV.
19724 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
19726 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
19727 (mips*-*-linux*): Move default with_llsc setting to where other
19729 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
19731 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
19732 Likewise. Remove default with_tune setting. Move default float
19733 setting to its own block. Handle with_llsc in the same block as above.
19735 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
19737 PR rtl-optimization/57425
19738 PR rtl-optimization/57569
19739 * alias.c (write_dependence_p): Add new parameters mem_mode,
19740 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
19741 Changed all callers.
19742 (canon_anti_dependence): New function.
19743 * cse.c (check_dependence): Use canon_anti_dependence.
19744 * cselib.c (cselib_invalidate_mem): Likewise.
19745 * rtl.h (canon_anti_dependence): Declare.
19747 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
19749 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
19750 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
19751 ".set mips3" for 64-bit targets.
19753 2013-06-15 Dehao Chen <dehao@google.com>
19755 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
19756 * gimple-low.c (gimple_check_call_matching_types): Likewise.
19757 (gimple_check_call_args): Likewise.
19758 * value-prof.c (check_ic_target): Likewise.
19759 * ipa-inline.c (early_inliner): Likewise.
19760 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
19761 * cgraph.c (cgraph_create_edge_1): Likewise.
19762 (cgraph_make_edge_direct): Likewise.
19764 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
19767 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
19768 rs6000_output_move_128bit to handle emitting quad memory
19769 operations. Set attribute length to 8 bytes.
19771 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
19773 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
19775 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
19776 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
19777 (aarch64_<su>mlsl<mode>): Likewise.
19779 2013-06-14 Mike Stump <mikestump@comcast.net>
19781 * Makefile.in (TARGET_H): Add insn-codes.h.
19783 2013-06-14 Alan Modra <amodra@gmail.com>
19785 PR middle-end/57134
19786 PR middle-end/57586
19787 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
19788 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
19789 bitfield expansion when EXPAND_MEMORY.
19790 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
19792 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
19794 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
19795 test for clearing quad memory on 32-bit later.
19797 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19799 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
19800 (fold_negate_expr): Likewise.
19801 (fold_real_zero_addition_p): Handle vectors.
19802 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
19804 2013-06-14 Alan Modra <amodra@gmail.com>
19806 * varasm.c (force_const_mem): Revert 2013-06-07 change.
19808 2013-06-13 Jan Hubicka <jh@suse.cz>
19810 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
19811 Local comdats are not externally visible.
19812 * symtab.c (dump_symtab_base): Dump externally visible.
19813 (verify_symtab_base): Verify back links in the symtab hash.
19815 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19817 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
19818 CONVERT_EXPR as equal nodes.
19820 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19822 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
19824 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19826 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
19827 Generalize to complex and vector.
19828 * tree.c (build_all_ones_cst): New function.
19829 * tree.h (build_all_ones_cst): Declare it.
19831 2013-06-13 Alan Modra <amodra@gmail.com>
19833 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
19834 * config/rs6000/rs6000.md (signbittf2): New insn.
19835 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
19836 (abstf2_internal, cmptf_internal2): Likewise.
19837 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
19839 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
19840 Pat Haugen <pthaugen@us.ibm.com>
19841 Peter Bergner <bergner@vnet.ibm.com>
19843 * config/rs6000/rs6000.c (emit_load_locked): Add support for
19844 power8 byte, half-word, and quad-word atomic instructions.
19845 (emit_store_conditional): Likewise.
19846 (rs6000_expand_atomic_compare_and_swap): Likewise.
19847 (rs6000_expand_atomic_op): Likewise.
19849 * config/rs6000/sync.md (larx): Add new modes for power8.
19851 (AINT): New mode iterator to include TImode as well as normal
19852 integer modes on power8.
19853 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
19854 that VSX registers are not considered. Use AINT mode iterator
19855 instead of INT1 to allow inclusion of quad word atomic operations
19857 (load_locked<mode>): Likewise.
19858 (store_conditional<mode>): Likewise.
19859 (atomic_compare_and_swap<mode>): Likewise.
19860 (atomic_exchange<mode>): Likewise.
19861 (atomic_nand<mode>): Likewise.
19862 (atomic_fetch_<fetchop_name><mode>): Likewise.
19863 (atomic_nand_fetch<mode>): Likewise.
19864 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
19866 (ATOMIC): On power8, add QImode, HImode modes.
19867 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
19868 modes that promote to SImode.
19869 (load_lockedti): Convert TImode arguments to PTImode, so that we
19870 get a guaranteed even/odd register pair.
19871 (load_lockedpti): Likewise.
19872 (store_conditionalti): Likewise.
19873 (store_conditionalpti): Likewise.
19875 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
19876 atomic load/store instructions.
19879 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
19881 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
19883 (insn_count): New attribute, with most cases extracted from...
19884 (length): ...here. Redefine most cases in terms of insn_count.
19885 (single_insn): Delete.
19886 (can_delay): Use insn_count to check for single instructions.
19887 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
19888 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
19889 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
19890 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
19891 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
19892 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
19893 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
19894 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
19895 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
19896 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
19897 rather than "length".
19898 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
19899 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
19900 Use "insn_count" rather than "length".
19901 * config/mips/mips-dsp.md
19902 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
19903 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
19906 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
19908 PR tree-optimization/57361
19909 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
19911 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
19913 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
19915 (aarch64_simd_combine<mode>): New instruction expansion.
19916 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
19917 function prototype.
19918 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
19919 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
19921 2013-06-12 Jan Hubicka <jh@suse.cz>
19923 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
19924 decl has when in streaming stage.
19925 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
19926 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
19928 2013-06-12 Roland Stigge <stigge@antcom.de>
19931 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
19933 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19935 PR tree-optimization/57537
19936 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
19937 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
19939 2013-06-12 Richard Biener <rguenther@suse.de>
19941 * data-streamer.h (streamer_write_char_stream): CSE
19942 obs->current_pointer.
19943 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
19944 streamer_write_char_stream manually and optimize the resulting loop.
19945 (streamer_write_hwi_stream): Likewise.
19947 2013-06-12 Jan Hubicka <jh@suse.cz>
19949 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
19950 * cgraph.h (varpool_create_empty_node): Declare.
19951 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
19953 * symtab.c (symtab_unregister_node): Be lax about missin entries
19955 (symtab_get_node): Update comment.
19956 * varpool.c (varpool_create_empty_node): Break out from ...
19957 (varpool_node_for_decl): ... here.
19958 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
19960 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
19962 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
19963 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
19964 part. Use straight-line flow at the end.
19965 <COMPONENT_REF>: Remove superfluous else.
19966 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
19968 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19971 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
19972 target hook even for !TREE_PUBLIC decls. If no resolution info
19973 is available, return false for common and external decls.
19975 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19977 * config/rl78/constraints.md (U): New constraint.
19978 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
19981 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19984 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
19985 to allow returning address to AT_PLATFORM name.
19987 2013-06-11 Jan Hubicka <jh@suse.cz>
19989 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
19990 * cgraph.h (symtab_node_base): Add weakref flag.
19991 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
19992 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
19993 (output_weakrefs): Use weakref flag.
19994 * fold-const.c (simple_operand_p): Handle WEAK.
19995 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
19996 * ipa.c (varpool_externally_visible_p): Drop weakref.
19997 (function_and_variable_visibility): Update comment; fix weakref
19998 sanity checks; do not clear DECL_WEAK on them.
19999 * lto-cgraph.c (lto_output_node): update.
20000 (lto_output_varpool_node): Update.
20001 (input_overwrite_node): Update.
20002 (input_node): Update.
20003 (input_varpool_node): Update.
20004 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
20005 (lto_symtab_merge_symbols): Add sanity check.
20006 (lto_symtab_prevailing_decl): Do not special case weakrefs.
20007 * passes.c (rest_of_decl_compilation): Set static flag, too.
20008 * symtab.c (dump_symtab_base): Dump weakref.
20009 (verify_symtab_base): Sanity check weakrefs.
20010 (symtab_make_decl_local): Remove duplicated code.
20011 (symtab_alias_ultimate_target): Simplify.
20012 * varpool.c (varpool_create_variable_alias): Set weakref flag.
20014 2013-06-11 Tom de Vries <tom@codesourcery.com>
20016 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
20017 sequence_vect == NULL.
20019 2013-06-11 DJ Delorie <dj@redhat.com>
20021 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
20022 (rl78_unwind_word_mode): New.
20024 2013-06-11 David Malcolm <dmalcolm@redhat.com>
20026 * final.c (debug_prefix_maps): Make static.
20028 2013-06-11 David Malcolm <dmalcolm@redhat.com>
20030 * function.c (initial_trampoline): Remove stray copy.
20032 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
20034 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
20036 2013-06-11 Martin Jambor <mjambor@suse.cz>
20038 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
20039 within bounds at the beginning of the function.
20041 2013-06-11 Alan Modra <amodra@gmail.com>
20043 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
20045 (get_named_section): Don't NULL !DECL_P decl.
20047 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
20049 * doc/invoke.texi (core-avx2): Document.
20051 (atom): Updated with MOVBE.
20053 2013-06-11 Richard Biener <rguenther@suse.de>
20055 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
20057 2013-06-11 Anton Blanchard <anton@samba.org>
20059 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
20060 correct shift value in little-endian mode.
20062 2013-06-11 Jakub Jelinek <jakub@redhat.com>
20065 * varasm.c (get_variable_align): Move #endif to the right place.
20067 2013-06-10 Cary Coutant <ccoutant@google.com>
20069 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
20070 for hash so that hash table traversal order is deterministic.
20072 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
20073 Pat Haugen <pthaugen@us.ibm.com>
20074 Peter Bergner <bergner@vnet.ibm.com>
20076 * config/rs6000/vector.md (GPR move splitter): Do not split moves
20077 of vectors in GPRS if they are direct moves or quad word load or
20080 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
20082 (direct_move_p): Likewise.
20083 (quad_load_store_p): Likewise.
20085 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
20086 classes into bins based on the physical register type.
20087 (reg_class_to_reg_type): Likewise.
20088 (IS_STD_REG_TYPE): Likewise.
20089 (IS_FP_VECT_REG_TYPE): Likewise.
20090 (reload_fpr_gpr): Arrays to determine what insn to use if we can
20091 use direct move instructions.
20092 (reload_gpr_vsx): Likewise.
20093 (reload_vsx_gpr): Likewise.
20094 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
20095 information that is a simplification of register classes. Also
20096 precalculate direct move reload helpers.
20097 (direct_move_p): New function to return true if the operation can
20098 be done as a direct move instruciton.
20099 (quad_load_store_p): New function to return true if the operation
20100 is a quad memory operation.
20101 (rs6000_legitimize_address): If quad memory, only allow register
20102 indirect for TImode addresses.
20103 (rs6000_legitimate_address_p): Likewise.
20104 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
20105 (rs6000_reload_register_type): Likewise.
20106 (register_to_reg_type): Return register type.
20107 (rs6000_secondary_reload_simple_move): New helper function for
20108 secondary reload and secondary memory needed to identify anything
20109 that is a simple move, and does not need reloading.
20110 (rs6000_secondary_reload_direct_move): New helper function for
20111 secondary reload to identify cases that can be done with several
20112 instructions via the direct move instructions.
20113 (rs6000_secondary_reload_move): New helper function for secondary
20114 reload to identify moves between register types that can be done.
20115 (rs6000_secondary_reload): Add support for quad memory operations
20116 and for direct move.
20117 (rs6000_secondary_memory_needed): Likewise.
20118 (rs6000_debug_secondary_memory_needed): Change argument names.
20119 (rs6000_output_move_128bit): New function to return the move to
20120 use for 128-bit moves, including knowing about the various
20121 limitations of quad memory operations.
20123 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
20124 memory operations. call rs6000_output_move_128bit for the actual
20125 instruciton(s) to generate.
20126 (vsx_movti_64bit): Likewise.
20128 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
20129 (UNSPEC_P8V_MTVSRWZ): Likewise.
20130 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
20131 (UNSPEC_P8V_MTVSRD): Likewise.
20132 (UNSPEC_P8V_XXPERMDI): Likewise.
20133 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
20134 (UNSPEC_FUSION_GPR): Likewise.
20135 (FMOVE128_GPR): New iterator for direct move.
20136 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
20137 (f32_sv): Likewise.
20138 (f32_dm): Likewise.
20139 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
20140 loads and direct move instructions.
20141 (zero_extendsidi2_lfiwzx): Likewise.
20142 (extendsidi2_lfiwax): Likewise.
20143 (extendsidi2_nocell): Likewise.
20144 (floatsi<mode>2_lfiwax): Likewise.
20145 (lfiwax): Likewise.
20146 (floatunssi<mode>2_lfiwzx): Likewise.
20147 (lfiwzx): Likewise.
20148 (fix_trunc<mode>_stfiwx): Likewise.
20149 (fixuns_trunc<mode>_stfiwx): Likewise.
20150 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
20151 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
20152 (parity<mode>2_cmpb): Set length/type attr.
20153 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
20154 for 'mr.' to fast_compare.
20155 (bpermd_<mode>): Change type attr to popcnt.
20156 (p8_fmrgow_<mode>): New insns for power8 direct move support.
20157 (p8_mtvsrwz_1): Likewise.
20158 (p8_mtvsrwz_2): Likewise.
20159 (reload_fpr_from_gpr<mode>): Likewise.
20160 (p8_mtvsrd_1): Likewise.
20161 (p8_mtvsrd_2): Likewise.
20162 (p8_xxpermdi_<mode>): Likewise.
20163 (reload_vsx_from_gpr<mode>): Likewise.
20164 (reload_vsx_from_gprsf): Likewise.
20165 (p8_mfvsrd_3_<mode>): LIkewise.
20166 (reload_gpr_from_vsx<mode>): Likewise.
20167 (reload_gpr_from_vsxsf): Likewise.
20168 (p8_mfvsrd_4_disf): Likewise.
20169 (multi-word GPR splits): Do not split direct moves or quad memory
20172 2013-06-10 David Malcolm <dmalcolm@redhat.com>
20174 * tree-into-ssa.c (interesting_blocks): Make static.
20176 2013-06-10 Jakub Jelinek <jakub@redhat.com>
20179 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
20180 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
20181 Use DATA_ABI_ALIGNMENT for that case instead if defined.
20182 (get_variable_align): New function.
20183 (get_variable_section, emit_bss, emit_common,
20184 assemble_variable_contents, place_block_symbol): Use
20185 get_variable_align instead of DECL_ALIGN.
20186 (assemble_noswitch_variable): Add align argument, use it
20187 instead of DECL_ALIGN.
20188 (assemble_variable): Adjust caller. Use get_variable_align
20189 instead of DECL_ALIGN.
20190 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
20192 (DATA_ABI_ALIGNMENT): Define.
20193 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
20194 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
20195 opt is false, only return the psABI mandated alignment increase.
20196 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
20197 (DATA_ABI_ALIGNMENT): ... this.
20198 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
20199 (DATA_ABI_ALIGNMENT): ... this.
20200 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
20201 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
20202 (DATA_ABI_ALIGNMENT): ... this.
20203 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
20204 * doc/tm.texi: Regenerated.
20206 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
20208 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
20209 cmp_code to construct REG_EQUAL note.
20211 2013-06-09 Jakub Jelinek <jakub@redhat.com>
20214 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
20215 that operands[2] doesn't overlap with operands[0].
20217 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
20218 Jan Hubicka <jh@suse.cz>
20220 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
20221 hack to mark symbols as used.
20223 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
20225 PR rtl-optimization/57559
20226 * lra-constraints.c (process_alt_operands): Don't discourage
20227 memory with known offset for offsetable memory constraint.
20228 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
20230 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
20232 * varasm.c (struct oc_local_state): Reorder fields.
20233 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
20234 and adjust accordingly.
20235 (output_constructor): Reorder initialization code and adjust call to
20236 output_constructor_bitfield.
20238 2013-06-07 Jan Hubicka <jh@suse.cz>
20240 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
20242 2013-06-07 David Malcolm <dmalcolm@redhat.com>
20244 * tree-object-size.c (unknown): Make const.
20246 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20248 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
20249 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
20250 for last alternative in the cpu_facility attribute.
20252 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20255 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
20256 (xordi3): Change operand 2 constraint to arm_xordi_operand.
20257 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
20258 * config/arm/constraints.md (Dg): New constraint.
20259 * config/arm/neon.md (xordi3_neon): Remove.
20260 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
20261 * config/arm/predicates.md (arm_xordi_operand): New predicate.
20263 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20265 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
20266 Clean up alternatives.
20268 2013-06-07 Alan Modra <amodra@gmail.com>
20270 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
20273 2013-06-07 Alan Modra <amodra@gmail.com>
20275 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
20277 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20279 * config/arm/constraints.md (Df): New constraint.
20280 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
20281 Correct length attribute for last two alternatives.
20283 2013-06-07 Alan Modra <amodra@gmail.com>
20285 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
20286 override user -mfp-in-toc.
20287 (offsettable_ok_by_alignment): Consider just the current access
20288 rather than the whole object, unless BLKmode. Handle
20289 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
20290 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
20291 for -mcmodel=medium.
20292 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
20293 override user -mfp-in-toc or -msum-in-toc. Default to
20294 -mno-fp-in-toc for -mcmodel=medium.
20296 2013-06-06 DJ Delorie <dj@redhat.com>
20298 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
20299 TARGET_VALID_POINTER_MODE.
20301 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
20302 Pat Haugen <pthaugen@us.ibm.com>
20303 Peter Bergner <bergner@vnet.ibm.com>
20305 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
20306 Document new power8 builtins.
20308 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
20309 condition code register, to allow 128-bit logical operations to be
20310 done in the VSX or GPR registers.
20311 (nor<mode>3): Use the canonical form for nor.
20312 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
20313 vclz*, and vpopcnt* vector instructions.
20314 (nand<mode>3): Likewise.
20315 (orc<mode>3): Likewise.
20316 (clz<mode>2): LIkewise.
20317 (popcount<mode>2): Likewise.
20319 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
20320 that only the GPRs are recognized.
20322 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
20323 support for new power8 builtins.
20325 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
20327 (xscvdpspn): Likewise.
20333 (vpopcnt): Likewise.
20334 (vpopcntb): Likewise.
20335 (vpopcnth): Likewise.
20336 (vpopcntw): Likewise.
20337 (vpopcntd): Likewise.
20339 (vmrgew): Likewise.
20340 (vmrgow): Likewise.
20342 (eqv_v16qi3): Likewise.
20343 (eqv_v8hi3): Likewise.
20344 (eqv_v4si3): Likewise.
20345 (eqv_v2di3): Likewise.
20346 (eqv_v4sf3): Likewise.
20347 (eqv_v2df3): Likewise.
20349 (nand_v16qi3): Likewise.
20350 (nand_v8hi3): Likewise.
20351 (nand_v4si3): Likewise.
20352 (nand_v2di3): Likewise.
20353 (nand_v4sf3): Likewise.
20354 (nand_v2df3): Likewise.
20356 (orc_v16qi3): Likewise.
20357 (orc_v8hi3): Likewise.
20358 (orc_v4si3): Likewise.
20359 (orc_v2di3): Likewise.
20360 (orc_v4sf3): Likewise.
20361 (orc_v2df3): Likewise.
20363 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
20364 allow power8 quad mode in 64-bit.
20365 (rs6000_builtin_vectorized_function): Add support to vectorize
20366 ISA 2.07 count leading zeros, population count builtins.
20367 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
20368 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
20369 (builtin_function_type): Add vgbbd builtin function which takes an
20371 (altivec_expand_vec_perm_const): Add support for new power8 merge
20374 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
20375 that does not include TImdoe for use with 32-bit.
20376 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
20378 (UNSPEC_VSX_CVDPSPN): Likewise.
20379 (vsx_xscvdpspn): Likewise.
20380 (vsx_xscvspdpn): Likewise.
20381 (vsx_xscvdpspn_scalar): Likewise.
20382 (vsx_xscvspdpn_directmove): Likewise.
20383 (vsx_and<mode>3): Split logical operations into 32-bit and
20384 64-bit. Add support to do logical operations on TImode as well as
20385 VSX vector types. Allow logical operations to be done in either
20386 VSX registers or in general purpose registers in 64-bit mode. Add
20387 splitters if GPRs were used. For AND, add clobber of CCmode to
20388 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
20390 (vsx_and<mode>3_32bit): Likewise.
20391 (vsx_and<mode>3_64bit): Likewise.
20392 (vsx_ior<mode>3): Likewise.
20393 (vsx_ior<mode>3_32bit): Likewise.
20394 (vsx_ior<mode>3_64bit): Likewise.
20395 (vsx_xor<mode>3): Likewise.
20396 (vsx_xor<mode>3_32bit): Likewise.
20397 (vsx_xor<mode>3_64bit): Likewise.
20398 (vsx_one_cmpl<mode>2): Likewise.
20399 (vsx_one_cmpl<mode>2_32bit): Likewise.
20400 (vsx_one_cmpl<mode>2_64bit): Likewise.
20401 (vsx_nor<mode>3): Likewise.
20402 (vsx_nor<mode>3_32bit): Likewise.
20403 (vsx_nor<mode>3_64bit): Likewise.
20404 (vsx_andc<mode>3): Likewise.
20405 (vsx_andc<mode>3_32bit): Likewise.
20406 (vsx_andc<mode>3_64bit): Likewise.
20407 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
20408 and xxlorc instructions.
20409 (vsx_eqv<mode>3_64bit): Likewise.
20410 (vsx_nand<mode>3_32bit): Likewise.
20411 (vsx_nand<mode>3_64bit): Likewise.
20412 (vsx_orc<mode>3_32bit): Likewise.
20413 (vsx_orc<mode>3_64bit): Likewise.
20415 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
20417 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
20419 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
20420 (p8_vmrgow): Likewise.
20421 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
20422 GPRs to be split under VSX.
20423 (p8v_clz<mode>2): Add power8 count leading zero support.
20424 (p8v_popcount<mode>2): Add power8 population count support.
20425 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
20428 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
20431 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
20433 (vec_nand): Likewise.
20434 (vec_vclz): Likewise.
20435 (vec_vclzb): Likewise.
20436 (vec_vclzd): Likewise.
20437 (vec_vclzh): Likewise.
20438 (vec_vclzw): Likewise.
20439 (vec_vgbbd): Likewise.
20440 (vec_vmrgew): Likewise.
20441 (vec_vmrgow): Likewise.
20442 (vec_vpopcnt): Likewise.
20443 (vec_vpopcntb): Likewise.
20444 (vec_vpopcntd): Likewise.
20445 (vec_vpopcnth): Likewise.
20446 (vec_vpopcntw): Likewise.
20448 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20450 PR rtl-optimization/57468
20451 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
20454 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20456 PR rtl-optimization/57459
20457 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
20458 type when setting live regs.
20460 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20462 * config/s390/s390.opt (mlra): New option.
20463 * config/s390/s390.c (s390_decompose_address): Check displacement
20464 for all registers for LRA.
20465 (s390_secondary_reload): Don't used secondary reloads for LRA.
20466 (s390_lra_p): New function.
20467 (TARGET_LRA_P): Define.
20468 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
20469 of attribute cpu_facility to zarch for the last alternative.
20470 (*cmpmem_short): Ditto.
20472 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
20474 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
20475 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
20476 (arm_expand_prologue): Likewise.
20478 2013-06-06 Teresa Johnson <tejohnson@google.com>
20481 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
20482 as this is now done by redirect_edge_and_branch_force.
20483 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
20484 barriers, and fix interaction with splitting.
20485 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
20486 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
20487 reflect changes made in the routine.
20488 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
20489 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
20490 since this is called in cfglayout mode, and replace partition fixup
20491 with assert as that is now done by force_nonfallthru_and_redirect.
20492 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
20493 already be marked with region crossing note.
20494 (insert_section_boundary_note): Make non-static, gate on flag
20495 has_bb_partition, rewrite to also check for multiple partitions.
20496 (rest_of_handle_reorder_blocks): Remove call to
20497 insert_section_boundary_note, now done later during free_cfg.
20498 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
20499 * bb-reorder.h (insert_section_boundary_note): Declare.
20500 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
20501 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
20502 invoke insert_section_boundary_note.
20503 (try_redirect_by_replacing_jump): Remove unnecessary
20504 check for region crossing note.
20505 (fixup_partition_crossing): New function.
20506 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
20507 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
20508 in non-cfglayout mode.
20509 (force_nonfallthru_and_redirect): Fixup partition boundaries,
20510 remove old code that tried to do this. Emit barrier correctly
20511 when we are in cfglayout mode.
20512 (last_bb_in_partition): New function.
20513 (rtl_split_edge): Correctly fixup partition boundaries.
20514 (commit_one_edge_insertion): Remove old code that tried to
20515 fixup region crossing edge since this is now handled in
20516 split_block, and set up insertion point correctly since
20517 block may now end in a jump.
20518 (verify_hot_cold_block_grouping): Guard against checking when not in
20519 linearized RTL mode.
20520 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
20522 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
20523 rtl_verify_flow_info, so not called in cfglayout mode.
20524 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
20525 (fixup_reorder_chain): Remove old code that attempted to fixup region
20526 crossing note as this is now handled in force_nonfallthru_and_redirect.
20527 (duplicate_insn_chain): Don't duplicate switch section notes.
20528 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
20530 * basic-block.h (emit_barrier_after_bb): Declare.
20532 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20534 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
20535 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
20536 arm_usatsihi): Adjust alternatives for arm_restrict_it.
20538 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20540 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
20542 * config/arm/ldmstm.md: Regenerate.
20544 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20546 * config/arm/sync.md (atomic_loaddi_1):
20547 Disable predication for arm_restrict_it.
20548 (arm_load_exclusive<mode>): Likewise.
20549 (arm_load_exclusivesi): Likewise.
20550 (arm_load_exclusivedi): Likewise.
20551 (arm_load_acquire_exclusive<mode>): Likewise.
20552 (arm_load_acquire_exclusivesi): Likewise.
20553 (arm_load_acquire_exclusivedi): Likewise.
20554 (arm_store_exclusive<mode>): Likewise.
20555 (arm_store_exclusive<mode>): Likewise.
20556 (arm_store_release_exclusivedi): Likewise.
20557 (arm_store_release_exclusive<mode>): Likewise.
20559 2013-06-06 Richard Biener <rguenther@suse.de>
20561 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
20563 (lto_tag_is_tree_code_p): Adjust.
20564 (lto_tag_is_gimple_code_p): Likewise.
20565 (lto_gimple_code_to_tag): Likewise.
20566 (lto_tag_to_gimple_code): Likewise.
20567 (lto_tree_code_to_tag): Likewise.
20568 (lto_tag_to_tree_code): Likewise.
20569 * data-streamer.h (streamer_write_hwi_in_range): Use
20570 uhwi streaming to stream the normalized range.
20571 (streamer_read_hwi_in_range): Likewise.
20573 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20575 * config/arm/arm.md (enabled_for_depr_it): New attribute.
20576 (predicable_short_it): Likewise.
20577 (predicated): Likewise.
20578 (enabled): Handle above.
20579 (define_cond_exec): Set predicated attribute to yes.
20581 2013-06-05 Mike Stump <mikestump@comcast.net>
20583 * gdbinit.in (__FUNCTION__): Add.
20585 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
20587 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
20588 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
20590 2013-06-05 Jan Hubicka <jh@suse.cz>
20592 * varasm.c (mark_decl_referenced): Revert the removal until targets
20595 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
20597 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
20598 instead of mark_decl_referenced.
20600 2013-06-05 Jan Hubicka <jh@suse.cz>
20602 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
20603 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
20604 and symtab_used_from_object_file_p.
20605 (cgraph_make_node_local_1): Clear forced_by_abi.
20606 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
20607 * cgraph.h (symtab_node_base): Add forced_by_abi.
20608 (decide_is_variable_needed): Remove.
20609 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
20610 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
20611 (decide_is_symbol_needed): ... this one; handle symbols in general;
20612 always analyze virtuals; honnor forced_by_abi.
20613 (cgraph_finalize_function): Update.
20614 (varpool_finalize_decl): Update.
20615 (symbol_defined_and_needed): Remove.
20616 (analyze_functions): Update.
20617 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20618 output_refs, input_overwrite_node): Handle forced_by_abi.
20619 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
20620 (address_taken_from_non_vtable_p): ... this one.
20621 (comdat_can_be_unshared_p_1): New function.
20622 (cgraph_comdat_can_be_unshared_p): Rename to ...
20623 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
20624 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
20625 (function_and_variable_visibility): Clear forced_by_abi as needed.
20626 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
20627 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
20628 * symtab.c (dump_symtab_base): Dump forced_by_abi.
20629 * varpool.c (decide_is_variable_needed): Remove.
20631 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20633 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
20634 (arm_option_override): Override arm_restrict_it where appropriate.
20635 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
20636 * config/arm/arm.opt (mrestrict-it): New command-line option.
20637 * doc/invoke.texi: Document -mrestrict-it.
20639 2013-06-05 David Malcolm <dmalcolm@redhat.com>
20641 * tsan.c (tsan_atomic_table): Make const.
20643 2013-06-05 Richard Biener <rguenther@suse.de>
20645 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
20646 index associated with the tree we are supposed to replace.
20647 * tree-streamer-out.c (pack_ts_base_value_fields): Output
20648 TREE_ASM_WRITTEN as zero for everything but SSA names.
20650 2013-06-05 David Malcolm <dmalcolm@redhat.com>
20652 * tree-ssa-structalias.c (call_stmt_vars): Make static.
20654 2013-06-04 Jan Hubicka <jh@suse.cz>
20656 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
20657 (input_node, input_varpool_node): Handle correctly external same
20659 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
20660 nodes at ltrans stage.
20662 2013-06-04 Jan Hubicka <jh@suse.cz>
20664 * ipa-inline.c (update_caller_keys): Fix availability test.
20665 (update_callee_keys): Likewise.
20666 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
20667 to follow ELF standard.
20669 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
20671 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
20672 (mips64r5900el-*-elf*): New configurations.
20673 * config/mips/mips-cpus.def (r5900): New processor.
20674 * config/mips/mips-tables.opt: Regenerate.
20675 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
20676 (mips_issue_rate): Handle PROCESSOR_R5900.
20677 (mips_reorg_process_insns): Force reorder mode for the R5900.
20678 * config/mips/mips.h (TARGET_MIPS5900): Define.
20679 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
20681 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
20683 * config/mips/mips.md (processor): Add r5900.
20684 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
20686 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20688 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
20689 into function to generate MOVI instruction.
20690 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
20691 (aarch64_preferred_simd_mode): Turn into wrapper.
20692 (aarch64_output_scalar_simd_mov_immediate): New function.
20693 * config/aarch64/aarch64-protos.h: Add prototype for above.
20695 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20697 * config/aarch64/aarch64.c (simd_immediate_info): Remove
20698 element_char member.
20699 (sizetochar): Return signed char.
20700 (aarch64_simd_valid_immediate): Remove elchar and other
20701 unnecessary variables.
20702 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
20703 Calculate element_char as required.
20704 * config/aarch64/aarch64-protos.h: Update and move prototype
20705 for aarch64_output_simd_mov_immediate.
20706 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
20709 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20711 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
20712 information completed by aarch64_simd_valid_immediate.
20713 (aarch64_legitimate_constant_p): Update arguments.
20714 (aarch64_simd_valid_immediate): Work with struct rather than many
20716 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
20717 (aarch64_simd_make_constant): Update arguments.
20718 (aarch64_output_simd_mov_immediate): Work with struct rather than
20719 many pointers. Output immediate directly rather than as operand.
20720 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
20722 * config/aarch64/constraints.md (Dn): Update arguments.
20724 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20726 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
20728 (aarch64_simd_immediate_valid_for_move): Remove.
20729 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
20730 (aarch64_simd_make_constant): Update call.
20731 (aarch64_output_simd_mov_immediate): Update call.
20732 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
20734 * config/aarch64/constraints.md (Dn): Update call.
20736 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20738 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
20739 return type to bool for prototype.
20740 (aarch64_legitimate_constant_p): Check for true instead of not -1.
20741 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
20742 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
20744 2013-06-04 Catherine Moore <clm@codesourcery.com>
20746 * config/mips/mips.opt (meva): New.
20747 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
20748 (ASM_SPEC): Handle -meva.
20749 * doc/invoke.texi (meva): Document.
20751 2013-06-04 Alan Modra <amodra@gmail.com>
20753 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
20756 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20758 * rtl.def: Add extra fourth optional field to define_cond_exec.
20759 * gensupport.c (process_one_cond_exec): Process attributes from
20761 * doc/md.texi: Document fourth field in define_cond_exec.
20763 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
20765 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
20766 out the processing order as in store_bit_field_1.
20768 2013-06-04 Jan Hubicka <jh@suse.cz>
20770 PR middle-end/57500
20771 * cgraphunit.c (cgraph_process_same_body_aliases): Create
20772 non-VAR_DECL node if it does not exist yet.
20774 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
20776 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
20777 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
20778 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
20779 target_cpu_default setting.
20781 2013-06-03 Teresa Johnson <tejohnson@google.com>
20783 * dumpfile.c (opt_info_switch_p): Change -fopt-info
20784 default to -fopt-info=optimized instead of all.
20785 * doc/invoke.texi: Ditto.
20786 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
20787 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
20788 (execute_vect_slp): Emit BB vectorization success under
20789 MSG_OPTIMIZED_LOCATIONS.
20790 * tree-vect-slp.c (vect_slp_transform_bb): Change
20791 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
20792 * tree-vect-loop.c (vect_transform_loop): Ditto.
20794 2013-06-03 Jason Merrill <jason@redhat.com>
20797 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
20798 Use TARGET_EXPR for C++.
20800 2013-06-03 Jakub Jelinek <jakub@redhat.com>
20802 PR rtl-optimization/57268
20803 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
20804 if DEBUG_INSN_P (insn).
20807 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20809 PR rtl-optimization/57268
20810 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20811 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20813 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20815 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
20816 (ix86_avoid_lea_for_addr): Likewise.
20817 (exact_dependency_1): Likewise.
20818 (ix86_adjust_cost): Likewise.
20819 (swap_top_of_ready_list): Fix formatting and !reload_completed check
20821 (do_reorder_for_imul): Fix typo, formatting and
20822 !reload_completed check removed.
20823 (ix86_sched_reorder): Fix typo and formatting.
20824 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
20827 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
20829 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
20831 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
20833 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
20834 <STRING_CST>: Likewise.
20835 <VECTOR_CST>: Likewise.
20837 2013-06-01 Janus Weil <janus@gcc.gnu.org>
20838 Mikael Morin <mikael@gcc.gnu.org>
20840 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
20841 * config.in: Regenerated.
20842 * configure: Regenerated.
20844 2013-06-01 Jan Hubicka <jh@suse.cz>
20846 PR middle-end/57366
20847 * cgraphunit.c (compile): When weakref is not supported,
20848 set up transparent aliases before final output pass.
20849 * varasm.c (assemble_alias): Do not try to do it here.
20851 2013-06-01 Jan Hubicka <jh@suse.cz>
20853 PR middle-end/57467
20854 * passes.c (for_per_function): Skip unanalyzed functions.
20856 2013-06-01 Jan Hubicka <jh@suse.cz>
20858 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
20859 (lto_symtab_merge_symbols_1): ... this one.
20860 (lto_symtab_merge_cgraph_nodes): Rename to ...
20861 (lto_symtab_merge_symbols): ... this one; simplify.
20862 * cgraph.c (same_body_aliases_done): Rename to ...
20863 (cpp_implicit_aliases_done): ... this one.
20864 (cgraph_create_function_alias): Update.
20865 (cgraph_same_body_alias): Update.
20866 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
20867 (verify_edge_corresponds_to_fndecl): Simplify.
20868 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
20869 (cgraph_node): Remove same_body_alias.
20870 (varpool_node): Remove alias_of and extra_name_alias.
20871 (same_body_aliases_done): Rename to ..
20872 (cpp_implicit_aliases_done): ... this one.
20873 (symtab_alias_ultimate_target): Add default parameter.
20874 (symtab_resolve_alias): New function.
20875 (fixup_same_cpp_alias_visibility): Declare.
20876 (cgraph_function_node): Add default parameter.
20877 (cgraph_node_asm_name): Likewise.
20878 (cgraph_function_or_thunk_node): Add default parameter; do
20879 not ICE when it is NULL.
20880 (varpool_variable_node): Likewise.
20881 * tree-emutls.c (create_emultls_var): Update.
20882 (ipa_lower_emutls): Update.
20883 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
20884 (cgraph_reset_node): Reset alias info.
20885 (cgraph_finalize_function): Update.
20886 (fixup_same_cpp_alias_visibility): Move to symtab.c.
20887 (analyze_function): Simplify.
20888 (cgraph_process_same_body_aliases): Simplify.
20889 (analyze_functions): Fixup same body aliases.
20890 (handle_alias_pairs): Simplify.
20891 (assemble_thunk): Update.
20892 (assemble_thunks_and_aliases): Update.
20893 (output_weakrefs): Rewrite.
20894 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
20895 (lto_output_varpool_node): Likewise.
20896 (compute_ltrans_boundary): Remve assert.
20897 (get_alias_symbol): New functoin.
20898 (input_node): Rewrite alias handling.
20899 (input_varpool_node): Likewise.
20900 * ipa-pure-const.c (propagate_pure_const): Fix formating.
20901 * ipa.c (process_references): Handle weakrefs correctly.
20902 (symtab_remove_unreachable_nodes): Likewise.
20903 * trans-mem.c (get_cg_data): Update.
20904 (ipa_tm_create_version_alias): Update.
20905 (ipa_tm_execute): Update.
20906 * symtab.c (dump_symtab_base): Dump aliases.
20907 (verify_symtab_base): Verify aliases.
20908 (symtab_node_availability): New function.
20909 (symtab_alias_ultimate_target): Simplify.
20910 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
20911 handle all the fixup cases.
20912 (symtab_resolve_alias): New function.
20913 * passes.c (ipa_write_summaries): Handle weakrefs.
20914 * varpool.c (varpool_analyze_node): Simplify.
20915 (assemble_aliases): Update.
20916 (varpool_create_variable_alias): Simplify.
20917 (varpool_extra_name_alias): Simplify.
20918 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
20919 (lto_symtab_merge_symbols): ... this one.
20921 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
20924 PR rtl-optimization/57268
20925 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20926 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20928 2013-06-01 Tobias Burnus <burnus@net-b.de>
20930 Partially reverted:
20931 2013-05-31 Tobias Burnus <burnus@net-b.de>
20933 PR middle-end/57073
20934 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
20937 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20939 PR rtl-optimization/57268
20940 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20941 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20943 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20945 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
20946 unordered comparison operators when -fno-trapping-math is in effect
20948 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
20949 and implement unordered comparison operators properly on the e500.
20951 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20953 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
20954 for constant scalar integers.
20955 (simplify_relational_operation_1): Likewise.
20957 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
20959 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
20960 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
20963 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20964 Igor Zamyatin <igor.zamyatin@intel.com>
20966 Silvermont (SLM) architecture performance tuning.
20967 * config/i386/i386.h (enum ix86_tune_indices): Add
20968 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
20969 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
20971 * config/i386/i386.c (initial_ix86_tune_features)
20972 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
20973 (ix86_lea_outperforms): Handle Silvermont tuning.
20974 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
20976 (ix86_use_lea_for_mov): Likewise.
20977 (ix86_avoid_lea_for_addr): Likewise.
20978 (ix86_lea_for_add_ok): Likewise.
20979 (exact_dependency_1): New function.
20980 (exact_store_load_dependency): Likewise.
20981 (ix86_adjust_cost): Handle Silvermont tuning.
20982 (do_reoder_for_imul): Likewise.
20983 (swap_top_of_ready_list): New function.
20984 (ix86_sched_reorder): Changed to handle Silvermont tuning.
20986 * config/i386/i386.md (peepholes that split memory operand in fp
20989 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20991 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
20992 Remove un-necessary braces.
20994 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20996 * config/aarch64/aarch64.c (aarch64_classify_symbol):
20997 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
20999 2013-05-31 Tobias Burnus <burnus@net-b.de>
21001 PR middle-end/57073
21002 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
21004 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21007 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
21008 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
21009 * config/arm/neon.md (iordi3_neon): Remove.
21010 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
21011 * config/arm/predicates.md (imm_for_neon_logic_operand):
21012 Move to earlier in the file.
21013 (neon_logic_op2): Likewise.
21014 (arm_iordi_operand_neon): New predicate.
21016 2013-05-31 Richard Biener <rguenther@suse.de>
21018 PR tree-optimization/57478
21019 PR tree-optimization/57453
21020 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
21023 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
21025 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
21026 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
21028 2013-05-30 Tobias Burnus <burnus@net-b.de>
21029 Thomas Koenig <tkoenig@gcc.gnu.org>
21031 PR middle-end/57073
21032 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
21033 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
21035 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
21037 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
21039 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
21041 * target.def (register_usage_leveling_p): New hook.
21042 * targhooks.c (default_register_usage_leveling_p): New.
21043 * targhooks.h (default_register_usage_leveling_p): New prototype.
21044 * lra-assigns.c (register_usage_leveling_p): Use the hook.
21045 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
21046 * doc/tm.texi: Update.
21047 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
21049 2013-05-30 Ian Bolton <ian.bolton@arm.com>
21051 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
21052 (*insv_reg<mode>): New define_insn.
21054 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
21056 PR rtl-optimization/57439
21057 * postreload.c (move2add_valid_value_p): Check that we have
21058 a zero subreg_regno_offset when accessing the register in
21059 the requested mode.
21061 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
21062 Igor Zamyatin <igor.zamyatin@intel.com>
21064 Silvermont (SLM) architecture pipeline model, tuning and
21066 * config.gcc: Add slm config options and target.
21068 * config/i386/slm.md: New.
21070 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
21072 * config/i386/i386-c.c (ix86_target_macros_internal): New case
21074 (ix86_target_macros_internal): Likewise.
21076 * config/i386/i386.c (slm_cost): New cost.
21077 (m_SLM): New macro flag.
21078 (initial_ix86_tune_features): Set m_SLM.
21079 (x86_accumulate_outgoing_args): Likewise.
21080 (x86_arch_always_fancy_math_387): Likewise.
21081 (processor_target_table): Add slm cost.
21082 (cpu_names): Add slm cpu name.
21083 (x86_option_override_internal): Set SLM ISA.
21084 (ix86_issue_rate): New case PROCESSOR_SLM.
21085 (ia32_multipass_dfa_lookahead): Likewise.
21086 (fold_builtin_cpu): Add slm.
21088 * config/i386/i386.h (TARGET_SLM): New target macro.
21089 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
21090 (processor_type): Add PROCESSOR_SLM.
21092 * config/i386/i386.md (cpu): Add new value "slm".
21093 (slm.md): Include slm.md.
21095 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
21096 Zhenqiang Chen <zhenqiang.chen@linaro.org>
21098 * config/arm/arm-protos.h: Add and update function protos.
21099 * config/arm/arm.c (use_simple_return_p): New added.
21100 (thumb2_expand_return): Check simple_return flag.
21101 * config/arm/arm.md: Add simple_return and conditional simple_return.
21102 * config/arm/iterators.md: Add iterator for return and simple_return.
21104 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
21106 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
21107 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
21108 (arm_emit_vfp_multi_reg_pop): Likewise.
21109 (thumb2_emit_ldrd_pop): Likewise.
21110 (arm_expand_epilogue): Add misc REG_CFA notes.
21111 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
21113 2013-05-29 Lawrence Crowl <crowl@google.com>
21115 * config/arm/t-arm: Update for below.
21117 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
21118 Change type to hash_table. Update dependent calls and types.
21120 * config/i386/t-cygming: Update for below.
21122 * config/i386/t-interix: Update for below.
21124 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
21125 Change type to hash_table. Update dependent calls and types.
21126 (i386_find_on_wrapper_list::wrappers): Likewise.
21128 * config/ia64/t-ia64: Update for below.
21130 * config/ia64/ia64.c (bundle_state_table):
21131 Change type to hash_table. Update dependent calls and types.
21133 * config/mips/mips.c (mips_reorg_process_insns::htab):
21134 Change type to hash_table. Update dependent calls and types.
21136 * config/sol2.c (solaris_comdat_htab):
21137 Change type to hash_table. Update dependent calls and types.
21139 * config/t-sol2: Update for above.
21141 2013-05-29 Teresa Johnson <tejohnson@google.com>
21143 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
21144 functions are not yet marked as defined.
21146 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
21147 Pat Haugen <pthaugen@us.ibm.com>
21148 Peter Bergner <bergner@vnet.ibm.com>
21150 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
21154 (vrotl<mode>3): Likewise.
21155 (vashl<mode>3): Likewise.
21156 (vlshr<mode>3): Likewise.
21157 (vashr<mode>3): Likewise.
21159 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
21160 support for power8 V2DI builtins.
21162 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
21163 power8 V2DI builtins.
21164 (vupkhsw): Likewise.
21165 (vupklsw): Likewise.
21166 (vaddudm): Likewise.
21167 (vminsd): Likewise.
21168 (vmaxsd): Likewise.
21169 (vminud): Likewise.
21170 (vmaxud): Likewise.
21171 (vpkudum): Likewise.
21172 (vpksdss): Likewise.
21173 (vpkudus): Likewise.
21174 (vpksdus): Likewise.
21179 (vsubudm): Likewise.
21180 (vcmpequd): Likewise.
21181 (vcmpgtsd): Likewise.
21182 (vcmpgtud): Likewise.
21183 (vcmpequd_p): Likewise.
21184 (vcmpgtsd_p): Likewise.
21185 (vcmpgtud_p): Likewise.
21186 (vupkhsw): Likewise.
21187 (vupklsw): Likewise.
21188 (vaddudm): Likewise.
21189 (vmaxsd): Likewise.
21190 (vmaxud): Likewise.
21191 (vminsd): Likewise.
21192 (vminud): Likewise.
21193 (vpksdss): Likewise.
21194 (vpksdus): Likewise.
21195 (vpkudum): Likewise.
21196 (vpkudus): Likewise.
21201 (vsubudm): Likewise.
21203 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
21204 support for power8 V2DI instructions.
21206 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
21207 power8 V2DI instructions. Combine pack and unpack insns to use an
21208 iterator for each mode. Check whether a particular mode supports
21209 Altivec instructions instead of just checking TARGET_ALTIVEC.
21210 (UNSPEC_VPKUWUM): Likewise.
21211 (UNSPEC_VPKSHSS): Likewise.
21212 (UNSPEC_VPKSWSS): Likewise.
21213 (UNSPEC_VPKUHUS): Likewise.
21214 (UNSPEC_VPKSHUS): Likewise.
21215 (UNSPEC_VPKUWUS): Likewise.
21216 (UNSPEC_VPKSWUS): Likewise.
21217 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
21218 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
21219 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
21220 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
21221 (UNSPEC_VUPKHSB): Likewise.
21222 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
21223 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
21224 (UNSPEC_VUPKHSH): Likewise.
21225 (UNSPEC_VUPKLSB): Likewise.
21226 (UNSPEC_VUPKLSH): Likewise.
21228 (VI_char): Likewise.
21229 (VI_scalar): Likewise.
21230 (VI_unit): Likewise.
21232 (VP_small): Likewise.
21233 (VP_small_lc): Likewise.
21234 (VU_char): Likewise.
21235 (add<mode>3): Likewise.
21236 (altivec_vaddcuw): Likewise.
21237 (altivec_vaddu<VI_char>s): Likewise.
21238 (altivec_vadds<VI_char>s): Likewise.
21239 (sub<mode>3): Likewise.
21240 (altivec_vsubcuw): Likewise.
21241 (altivec_vsubu<VI_char>s): Likewise.
21242 (altivec_vsubs<VI_char>s): Likewise.
21243 (altivec_vavgs<VI_char>): Likewise.
21244 (altivec_vcmpbfp): Likewise.
21245 (altivec_eq<mode>): Likewise.
21246 (altivec_gt<mode>): Likewise.
21247 (altivec_gtu<mode>): Likewise.
21248 (umax<mode>3): Likewise.
21249 (smax<mode>3): Likewise.
21250 (umin<mode>3): Likewise.
21251 (smin<mode>3): Likewise.
21252 (altivec_vpkuhum): Likewise.
21253 (altivec_vpkuwum): Likewise.
21254 (altivec_vpkshss): Likewise.
21255 (altivec_vpkswss): Likewise.
21256 (altivec_vpkuhus): Likewise.
21257 (altivec_vpkshus): Likewise.
21258 (altivec_vpkuwus): Likewise.
21259 (altivec_vpkswus): Likewise.
21260 (altivec_vpks<VI_char>ss): Likewise.
21261 (altivec_vpks<VI_char>us): Likewise.
21262 (altivec_vpku<VI_char>us): Likewise.
21263 (altivec_vpku<VI_char>um): Likewise.
21264 (altivec_vrl<VI_char>): Likewise.
21265 (altivec_vsl<VI_char>): Likewise.
21266 (altivec_vsr<VI_char>): Likewise.
21267 (altivec_vsra<VI_char>): Likewise.
21268 (altivec_vsldoi_<mode>): Likewise.
21269 (altivec_vupkhsb): Likewise.
21270 (altivec_vupkhs<VU_char>): Likewise.
21271 (altivec_vupkls<VU_char>): Likewise.
21272 (altivec_vupkhsh): Likewise.
21273 (altivec_vupklsb): Likewise.
21274 (altivec_vupklsh): Likewise.
21275 (altivec_vcmpequ<VI_char>_p): Likewise.
21276 (altivec_vcmpgts<VI_char>_p): Likewise.
21277 (altivec_vcmpgtu<VI_char>_p): Likewise.
21278 (abs<mode>2): Likewise.
21279 (vec_unpacks_hi_v16qi): Likewise.
21280 (vec_unpacks_hi_v8hi): Likewise.
21281 (vec_unpacks_lo_v16qi): Likewise.
21282 (vec_unpacks_hi_<VP_small_lc>): Likewise.
21283 (vec_unpacks_lo_v8hi): Likewise.
21284 (vec_unpacks_lo_<VP_small_lc>): Likewise.
21285 (vec_pack_trunc_v8h): Likewise.
21286 (vec_pack_trunc_v4si): Likewise.
21287 (vec_pack_trunc_<mode>): Likewise.
21289 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
21291 (vec_vmaxsd): Likewise.
21292 (vec_vmaxud): Likewise.
21293 (vec_vminsd): Likewise.
21294 (vec_vminud): Likewise.
21295 (vec_vpksdss): Likewise.
21296 (vec_vpksdus): Likewise.
21297 (vec_vpkudum): Likewise.
21298 (vec_vpkudus): Likewise.
21299 (vec_vrld): Likewise.
21300 (vec_vsld): Likewise.
21301 (vec_vsrad): Likewise.
21302 (vec_vsrd): Likewise.
21303 (vec_vsubudm): Likewise.
21304 (vec_vupkhsw): Likewise.
21305 (vec_vupklsw): Likewise.
21307 2013-05-29 Jan Hubicka <jh@suse.cz>
21309 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
21310 flags; reorder rest of fields in more consistent way.
21311 (varpool_node): Remove analyzed, finalized and alias.
21312 (cgraph_ndoe): Likewise.
21313 (symtab_alias_ultimate_target): New function.
21314 (cgraph_function_node): Move offline.
21315 (cgraph_reset_node): Declare.
21316 (cgraph_comdat_can_be_unshared_p): Remove.
21317 (varpool_remove_initializer): Declare.
21318 (varpool_first_defined_variable, varpool_next_defined_variable
21319 cgraph_first_defined_function, cgraph_next_defined_function): Update.
21320 (cgraph_function_with_gimple_body_p): Update.
21321 (varpool_all_refs_explicit_p): Update.
21322 (symtab_alias_target): New function.
21323 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
21324 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
21325 (cgraph_function_or_thunk_node): Simplify using
21326 symtab_alias_ultimate_target.
21327 (varpool_variable_node): Likewise.
21328 * cgraph.c (cgraph_create_function_alias): Update.
21329 (cgraph_add_thunk): Update.
21330 (cgraph_remove_node): Update.
21331 (dump_cgraph_node): Do not dump removed flags.
21332 (cgraph_function_body_availability): Update.
21333 (cgraph_propagate_frequency): Update.
21334 (verify_cgraph_node): Check sanity of local flag.
21335 (cgraph_function_node): Move here from cgraph.h; revamp for
21336 cgraph_function_or_thunk_node.
21337 * lto-symtab.c (lto_varpool_replace_node): Update.
21338 (lto_symtab_resolve_can_prevail_p): Update.
21339 (lto_symtab_merge_cgraph_nodes): Update.
21340 * ipa-cp.c (determine_versionability, initialize_node_lattices,
21341 propagate_constants_accross_call, devirtualization_time_bonus,
21342 ipcp_propagate_stage): Update.
21343 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
21344 * ipa-inline-transform.c (clone_inlined_nodes,
21345 preserve_function_body_p): Update.
21346 * ipa-reference.c (propagate): Update.
21347 (write_node_summary_p): Update.
21348 * toplev.c (wrapup_global_declaration_2): Update.
21349 * cgraphunit.c (cgraph_analyze_function): Rename to ...
21350 (analyze_function) ... this one.
21351 (cgraph_process_new_functions): Update.
21352 (cgraph_reset_node): Export.
21353 (cgraph_finalize_function): Update.
21354 (cgraph_add_new_function): Update.
21355 (process_function_and_variable_attributes): Update.
21356 (varpool_finalize_decl): Update.
21357 (symbol_finalized): Remove.
21358 (symbol_finalized_and_needed): Rename to ...
21359 (symbol_defined_and_needed): ... update.
21360 (cgraph_analyze_functions): Update.
21361 (handle_alias_pairs): Update.
21362 (mark_functions_to_output): Update.
21363 (assemble_thunk): Update.
21364 (output_in_order): Update.
21365 (output_weakrefs): Update.
21366 (finalize_compilation_unit): Update.
21367 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
21368 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
21369 input_node, input_varpool_node): Update.
21370 * dbxout.c (dbxout_expand_expr): Update.
21371 * cgraphclones.c (cgraph_clone_node): Update.
21372 (cgraph_copy_node_for_versioning): Update.
21373 (cgraph_materialize_clone): Update.
21374 (cgraph_materialize_all_clones): Update.
21375 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
21376 propagate_pure_const, propagate_nothrow): Update.
21377 * lto-streamer-out.c (lto_output, write_symbol): Update.
21378 * ipa-utils.c (ipa_reverse_postorder): Update.
21379 * ipa-inline.c (can_inline_edge_p): Update.
21380 (update_caller_keys, ipa_inline): Update.
21381 * dwarf2out.c (reference_to_unused,
21382 premark_types_used_by_global_vars_helper): Update.
21383 * tree-eh.c (tree_could_trap_p): Update.
21384 * ipa-split.c (consider_split, execute_split_functions): Update.
21385 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
21386 has_addr_references_p): Update; move ahead in file for better
21388 (process_references): Simplify.
21389 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
21390 bodies are removed.
21391 (cgraph_comdat_can_be_unshared_p): Make static.
21392 (cgraph_externally_visible_p): Update.
21393 (varpool_externally_visible_p): Update.
21394 (function_and_variable_visibility): Update.
21395 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
21396 ipa_tm_mark_force_output_node): Update.
21397 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
21398 estimate_edge_devirt_benefit, inline_generate_summary,
21399 inline_write_summary): Update.
21400 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
21401 * ipa-prop.c (ipa_compute_jump_functions): Update.
21402 (ipa_print_node_params, ipa_prop_read_section,
21403 ipa_update_after_lto_read, read_replacements_section): Update.
21404 * varasm.c (mark_decl_referenced): Update.
21405 (assemble_alias, dump_tm_clone_pairs): Update.
21406 * tree-inline.c (copy_bb): Update.
21407 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
21409 * symtab.c (dump_symtab_base): Print new flags.
21410 (verify_symtab_base): Verify new flags.
21411 (symtab_alias_ultimate_target): New function.
21412 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
21413 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
21415 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
21417 * i386.c (ix86_get_function_versions_dispatcher,
21418 ix86_generate_version_dispatcher_body): Update.
21419 (fold_builtin_cpu): Use varpool_add_new_variable.
21420 * varpool.c (varpool_remove_initializer): Break out from ...
21421 (varpool_remove_node): ... this one.
21422 (dump_varpool_node, varpool_node_for_asm,
21423 cgraph_variable_initializer_availability, varpool_analyze_node,
21424 varpool_assemble_decl, varpool_remove_unreferenced_decls,
21425 varpool_finalize_named_section_flags, varpool_create_variable_alias):
21428 2013-05-29 Jan Hubicka <jh@suse.cz>
21430 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
21432 2013-05-29 Easwaran Raman <eraman@google.com>
21434 PR tree-optimization/57442
21435 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
21436 when control exits the main loop.
21438 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
21440 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
21442 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
21443 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
21444 * rx/t-rx: Add rx100 under multi library matches option for nofpu
21447 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21449 PR tree-optimization/57441
21450 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
21451 Don't limit size of incr_vec to number of candidates.
21453 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
21455 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
21456 and mips16 directories.
21457 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
21458 (MULTILIB_DIRNAMES): Ditto.
21459 (MULTILIB_EXCEPTIONS): Add new exceptions.
21460 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
21461 (MULTILIB_DIRNAMES): Ditto.
21462 (MULTILIB_EXCEPTIONS): Add new exceptions.
21464 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21465 Marcus Shawcroft <marcus.shawcroft@arm.com>
21467 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
21468 SYMBOL_TINY_ABSOLUTE.
21469 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
21470 SYMBOL_TINY_ABSOLUTE.
21471 (aarch64_expand_mov_immediate): Likewise.
21472 (aarch64_classify_symbol): Likewise.
21473 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
21474 Permit SYMBOL_TINY_ABSOLUTE.
21475 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
21477 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21478 Marcus Shawcroft <marcus.shawcroft@arm.com>
21480 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
21481 Refactor if/switch. Replace gcc_assert with if.
21483 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21485 * config/i386/i386.c (initial_ix86_tune_features): Enable
21486 FP Reassociation for AMD bdver1 and bdver2.
21488 2013-05-29 Martin Jambor <mjambor@suse.cz>
21490 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
21491 and IMAGPART_EXPR do not occur within other handled_components.
21493 2013-05-29 Richard Biener <rguenther@suse.de>
21495 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
21496 access on whether the use is in the BB we currently try to
21498 (vect_bb_vectorization_profitable_p): Pass the BB we currently
21499 vectorize to vect_bb_slp_scalar_cost.
21501 2013-05-29 Richard Biener <rguenther@suse.de>
21503 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
21504 computing scalar cost offsetted by stmts that are kept live
21506 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
21507 for computation of scalar cost.
21509 2013-05-28 Steve Ellcey <sellcey@mips.com>
21511 * config/mips/mips-cpus.def (mips32r2): Change processor type.
21513 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
21515 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
21516 array notation built-in reduction functions.
21517 * doc/passes.texi (Passes): Added documentation about changes done
21519 * doc/invoke.texi (C Dialect Options): Added documentation about
21520 the -fcilkplus flag.
21521 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
21522 (BUILTINS_DEF): Depend on cilkplus.def.
21523 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
21524 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
21525 * cilkplus.def: New file.
21527 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
21529 PR rtl-optimization/57439
21530 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
21532 2013-05-28 Easwaran Raman <eraman@google.com>
21534 PR tree-optimization/57337
21535 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
21536 (find_insert_point): Correctly identify the insertion point
21537 when two statements with the same UID is compared.
21539 2013-05-28 Richard Biener <rguenther@suse.de>
21541 PR tree-optimization/56787
21542 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
21543 from the list of data references.
21544 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
21546 (vect_analyze_loop_operations): Likewise.
21547 (vect_transform_loop): Remove clobbers.
21549 2013-05-28 Martin Jambor <mjambor@suse.cz>
21551 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
21552 and REALPART_EXPRs have scalar type.
21554 2013-05-28 Richard Biener <rguenther@suse.de>
21556 PR tree-optimization/57411
21557 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
21559 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
21560 virtual operand propagation.
21562 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
21564 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
21565 destination register for bmasksi_vis.
21566 (vector_init_bshuffle): Likewise.
21567 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
21569 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
21571 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
21572 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
21573 mode if the instruction isn't available in the original mode.
21574 * config/sparc/sparc.opt (mfix-ut699): New option.
21575 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
21576 (divdf3): Turn into expander.
21577 (divdf3_nofix): New insn.
21578 (divdf3_fix): Likewise.
21579 (divsf3): Disable if -mfix-ut699.
21580 (sqrtdf2): Turn into expander.
21581 (sqrtdf2_nofix): New insn.
21582 (sqrtdf2_fix): Likewise.
21583 (sqrtsf2): Disable if -mfix-ut699.
21585 2013-05-27 Richard Biener <rguenther@suse.de>
21587 PR middle-end/57412
21588 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
21589 block for the new loop.
21591 2013-05-27 Richard Biener <rguenther@suse.de>
21593 PR tree-optimization/57343
21594 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
21595 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
21596 (number_of_iterations_cond): Do not build the folded tree.
21598 2013-05-27 Richard Biener <rguenther@suse.de>
21601 PR middle-end/57381
21602 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
21603 OEP_CONSTANT_ADDRESS_OF retained.
21605 PR tree-optimization/57417
21606 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
21607 for unchanged base.
21608 (set_ssa_val_to): Compare addresses using
21609 get_addr_base_and_unit_offset.
21611 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
21613 PR rtl-optimization/56833
21614 * postreload.c (move2add_record_mode): New function.
21615 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
21616 (move2add_use_add2_insn): Use move2add_record_sym_value.
21617 (move2add_use_add3_insn): Likewise.
21618 (reload_cse_move2add): Use move2add_valid_value_p and
21619 move2add_record_mode. Invalidate call-clobbered and REG_INC
21620 affected regs by setting reg_mode to VOIDmode.
21621 (move2add_note_store): Don't pretend the inside of a SUBREG is
21622 the actual destination. Invalidate single/leading registers by
21623 setting reg_mode to VOIDmode.
21624 Use move2add_record_sym_value, move2add_valid_value_p and
21625 move2add_record_mode.
21627 2013-05-27 Richard Biener <rguenther@suse.de>
21629 PR tree-optimization/57396
21630 * tree-affine.c (double_int_constant_multiple_p): Properly
21631 return false for val == 0 and div != 0.
21633 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21635 * config/mips/mips.h: Use #elif in preprocessor conditions.
21637 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21640 * config/mips/constraints.md (kl): New constraint.
21641 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
21642 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
21643 constraint for operand 0. Split after CSE for MIPS16. Emit a move
21644 from LO for MIPS16.
21645 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
21647 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21650 * config/mips/mips.c (mips_can_inline_p): New function.
21651 (TARGET_CAN_INLINE_P): Define.
21653 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
21655 * sched-int.h (ds_t, dw_t): Make unsigned int.
21656 Fix documentation that describes how all the ds_t bits are used.
21657 Reserve the last bit for delayed-branch scheduling.
21658 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
21659 (BITS_PER_DEP_WEAK): Fix definition and documentation.
21660 (gen_dep_weak_1): Remove prototype.
21661 * sched-deps.c (get_dep_weak_1): Make static.
21662 * target.def (speculate_insn, needs_block_p, gen_spec_check,
21663 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
21664 * doc/tm.texi: Regenerate.
21665 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
21667 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
21670 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
21672 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
21673 Sandra Loosemore <sandra@codesourcery.com>
21675 * config.gcc (powerpc-*): Allow native for with-cpu.
21677 2013-05-24 Jeff Law <law@redhat.com>
21679 PR tree-optimization/57124
21680 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
21681 conversion feeding a condition if the range has an overflow
21682 if -fstrict-overflow. Add warnings for when we do make the
21685 2013-05-24 Dehao Chen <dehao@google.com>
21687 * tree-cfg.c (locus_discrim_map): Fix the typo.
21688 (locus_discrim_hasher): Likewise.
21689 (locus_discrim_hasher::hash): Likewise.
21690 (locus_discrim_hasher::equal): Likewise.
21692 2013-05-24 Martin Jambor <mjambor@suse.cz>
21694 PR tree-optimization/57294
21695 * cgraph.h (ipa_record_stmt_references): Declare.
21696 * cgraphbuild.c (ipa_record_stmt_references): New function.
21697 (build_cgraph_edges): Use ipa_record_stmt_references.
21698 (rebuild_cgraph_edges): Likewise.
21699 (cgraph_rebuild_references): Likewise.
21700 * ipa-prop.c (ipa_modify_call_arguments): Discard references
21701 associated with the old statement and build references from the
21702 newly built statements.
21703 * ipa-ref.c (ipa_remove_stmt_references): New function.
21704 * ipa-ref.h (ipa_remove_stmt_references): Declare.
21706 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
21708 * lra-constraints.c (emit_spill_move): Use smaller mode for
21710 (check_and_process_move): Consider mem-reg moves for secondary
21712 (curr_insn_transform): Don't lose insns emitted before for
21713 secondary memory moves.
21714 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
21715 reg set up in the current insn.
21717 2013-05-24 Dehao Chen <dehao@google.com>
21719 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
21721 (locus_descrim_hasher::equal): Likewise.
21722 (build_gimple_cfg): New discriminator assignment algorithm.
21723 (make_edges): Likewise.
21724 (next_discriminator_for_locus): Likewise.
21725 (same_line_p): Likewise.
21726 (assign_discriminators): Likewise.
21727 (make_cond_expr_edges): Likewise.
21728 (make_gimple_switch_edges): Likewise.
21729 (make_goto_expr_edges): Likewise.
21730 (make_gimple_asm_edges): Likewise.
21732 2013-05-24 Ian Bolton <ian.bolton@arm.com>
21734 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
21735 X format specifier to only display bottom 16 bits.
21736 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
21737 immediate to match for operand 2, since it will be masked.
21739 2013-05-24 Richard Biener <rguenther@suse.de>
21741 PR tree-optimization/57287
21742 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
21743 all SSA names that occur in abnormal PHIs.
21745 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
21748 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
21749 that index is not negative.
21751 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
21753 PR rtl-optimization/55177
21754 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
21755 (simplify_byte_swapping_operation): New.
21756 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
21757 (simplify_relational_operation_1): Deal with BSWAP.
21759 2013-05-23 Richard Henderson <rth@redhat.com>
21762 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
21763 (ix86_reorg): Call it.
21765 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
21768 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
21769 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
21770 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
21772 2013-05-23 Christian Bruel <christian.bruel@st.com>
21775 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
21777 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21778 Marcus Shawcroft <marcus.shawcroft@arm.com>
21780 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
21781 * config/aarch64/constraints.md (Usa): Remove.
21782 * doc/md.texi (AArch64 Usa): Remove.
21784 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21785 Marcus Shawcroft <marcus.shawcroft@arm.com>
21787 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
21788 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
21789 * config/aarch64/predicates.md (aarch64_const_address): Remove.
21790 (aarch64_mov_operand): Use aarch64_mov_operand_p.
21792 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
21794 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
21795 instruction (AdvSIMD).
21796 * config/aarch64/aarch64-builtins.c
21797 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
21798 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
21800 2013-05-23 Martin Jambor <mjambor@suse.cz>
21802 PR middle-end/57347
21803 * tree.h (contains_bitfld_component_ref_p): Declare.
21804 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
21805 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
21807 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
21808 not access a bit-field. Assert all final offsets are byte-aligned.
21810 2013-05-23 Richard Biener <rguenther@suse.de>
21812 PR tree-optimization/57380
21813 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
21814 least one invariant or re-used load.
21815 * passes.c (init_optimization_passes): Move pass_phiprop before
21818 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
21820 * config/aarch64/aarch64-simd.md
21821 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
21823 2013-05-23 Richard Biener <rguenther@suse.de>
21825 PR middle-end/57381
21826 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
21827 OEP_CONSTANT_ADDRESS_OF retained.
21829 2013-05-23 Jakub Jelinek <jakub@redhat.com>
21831 PR middle-end/57344
21832 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
21833 don't lower unit. Handle unit not being always BITS_PER_WORD.
21835 2013-05-23 Richard Biener <rguenther@suse.de>
21837 PR rtl-optimization/57341
21838 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
21839 instead of true_dependence.
21841 2013-05-22 David Malcolm <dmalcolm@redhat.com>
21843 * bb-reorder.c (branch_threshold): Make const.
21844 (exec_threshold): Ditto.
21846 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21847 Pat Haugen <pthaugen@us.ibm.com>
21848 Peter Bergner <bergner@vnet.ibm.com>
21850 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
21851 documentation for the power8 crypto builtins.
21853 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
21855 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
21856 macros for defining power8 builtin functions.
21857 (BU_P8V_AV_2): Likewise.
21858 (BU_P8V_AV_P): Likewise.
21859 (BU_P8V_VSX_1): Likewise.
21860 (BU_P8V_OVERLOAD_1): Likewise.
21861 (BU_P8V_OVERLOAD_2): Likewise.
21862 (BU_CRYPTO_1): Likewise.
21863 (BU_CRYPTO_2): Likewise.
21864 (BU_CRYPTO_3): Likewise.
21865 (BU_CRYPTO_OVERLOAD_1): Likewise.
21866 (BU_CRYPTO_OVERLOAD_2): Likewise.
21867 (XSCVSPDP): Fix typo, point to the correct instruction.
21868 (VCIPHER): Add power8 crypto builtins.
21869 (VCIPHERLAST): Likewise.
21870 (VNCIPHER): Likewise.
21871 (VNCIPHERLAST): Likewise.
21872 (VPMSUMB): Likewise.
21873 (VPMSUMH): Likewise.
21874 (VPMSUMW): Likewise.
21875 (VPERMXOR_V2DI): Likewise.
21876 (VPERMXOR_V4SI: Likewise.
21877 (VPERMXOR_V8HI: Likewise.
21878 (VPERMXOR_V16QI: Likewise.
21879 (VSHASIGMAW): Likewise.
21880 (VSHASIGMAD): Likewise.
21881 (VPMSUM): Likewise.
21882 (VPERMXOR): Likewise.
21883 (VSHASIGMA): Likewise.
21885 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
21886 __CRYPTO__ if the crypto instructions are available.
21887 (altivec_overloaded_builtins): Add support for overloaded power8
21890 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
21891 support for power8 crypto builtins.
21892 (builtin_function_type): Likewise.
21893 (altivec_init_builtins): Add support for builtins that take vector
21894 long long (V2DI) arguments.
21896 * config/rs6000/crypto.md: New file, define power8 crypto
21899 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21900 Pat Haugen <pthaugen@us.ibm.com>
21901 Peter Bergner <bergner@vnet.ibm.com>
21903 * doc/invoke.texi (Option Summary): Add power8 options.
21904 (RS/6000 and PowerPC Options): Likewise.
21906 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
21907 constraints.md instead of rs6000.h. Reorder w* constraints. Add
21908 wm, wn, wr documentation.
21910 * config/rs6000/constraints.md (wm): New constraint for VSX
21911 registers if direct move instructions are enabled.
21912 (wn): New constraint for no registers.
21913 (wq): New constraint for quad word even GPR registers.
21914 (wr): New constraint if 64-bit instructions are enabled.
21915 (wv): New constraint if power8 vector instructions are enabled.
21916 (wQ): New constraint for quad word memory locations.
21918 * config/rs6000/predicates.md (const_0_to_15_operand): New
21919 constraint for 0..15 for crypto instructions.
21920 (gpc_reg_operand): If VSX allow registers in VSX registers as well
21921 as GPR and floating point registers.
21922 (int_reg_operand): New predicate to match only GPR registers.
21923 (base_reg_operand): New predicate to match base registers.
21924 (quad_int_reg_operand): New predicate to match even GPR registers
21925 for quad memory operations.
21926 (vsx_reg_or_cint_operand): New predicate to allow vector logical
21927 operations in both GPR and VSX registers.
21928 (quad_memory_operand): New predicate for quad memory operations.
21929 (reg_or_indexed_operand): New predicate for direct move support.
21931 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
21932 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
21933 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
21934 (POWERPC_MASKS): Add power8 options.
21935 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
21938 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
21939 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
21941 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
21942 (-mpower8-fusion): New power8 options.
21943 (-mpower8-fusion-sign): Likewise.
21944 (-mpower8-vector): Likewise.
21945 (-mcrypto): Likewise.
21946 (-mdirect-move): Likewise.
21947 (-mquad-memory): Likewise.
21949 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
21950 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
21952 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
21953 (rs6000_debug_vector_unit): Add p8_vector.
21954 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
21955 definitions. Also print fusion state.
21956 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
21957 (rs6000_builtin_mask_calculate): Add power8 builtin support.
21958 (rs6000_option_override_internal): Add support for power8.
21959 (rs6000_common_init_builtins): Add debugging for skipped builtins
21960 if -mdebug=builtin.
21961 (rs6000_adjust_cost): Add power8 support.
21962 (rs6000_issue_rate): Likewise.
21963 (insn_must_be_first_in_group): Likewise.
21964 (insn_must_be_last_in_group): Likewise.
21965 (force_new_group): Likewise.
21966 (rs6000_register_move_cost): Likewise.
21967 (rs6000_opt_masks): Likewise.
21969 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
21970 power8 capable assembler, default to power7 options.
21971 (TARGET_DIRECT_MOVE): Likewise.
21972 (TARGET_CRYPTO): Likewise.
21973 (TARGET_P8_VECTOR): Likewise.
21974 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
21975 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
21976 (VECTOR_MEM_P8_VECTOR_P): Likewise.
21977 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
21978 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
21979 (TARGET_XSCVDPSPN): Likewise.
21980 (TARGET_XSCVSPDPN): Likewsie.
21981 (TARGET_SYNC_HI_QI): Likewise.
21982 (TARGET_SYNC_TI): Likewise.
21983 (MASK_CRYPTO): Likewise.
21984 (MASK_DIRECT_MOVE): Likewise.
21985 (MASK_P8_FUSION): Likewise.
21986 (MASK_P8_VECTOR): Likewise.
21987 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
21988 temporary used by some of the direct move instructions to get two FP
21989 temporary registers does not force creation of a stack frame.
21990 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
21991 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
21992 that any VSX registers are tieable, even if they are also an
21993 Altivec vector mode.
21994 (r6000_reg_class_enum): Add wm, wr, wv constraints.
21995 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
21996 (RS6000_BTM_CRYPTO): Likewise.
21997 (RS6000_BTM_COMMON): Likewise.
21999 * config/rs6000/rs6000.md (cpu attribute): Add power8.
22000 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
22001 (enum rs6000_vector): Add power8 vector support.
22003 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22007 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
22008 (any_sibcall_could_use_r3): this and handle indirect calls.
22009 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
22011 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22013 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
22015 2013-05-22 Richard Biener <rguenther@suse.de>
22017 PR middle-end/57349
22018 * profile.c (branch_prob): Do not split blocks that are
22019 abnormally receiving from ECF_RETURNS_TWICE functions.
22021 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
22023 * recog.c (offsettable_address_addr_space_p): Fix calculation of
22024 address mode. Move pointer mode initialization to the same place.
22026 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
22028 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
22029 while it has any effect.
22031 2013-05-21 Easwaran Raman <eraman@google.com>
22033 PR tree-optimization/57322
22034 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
22035 UID of the statement added to the BB to be 1.
22037 2013-05-21 Jakub Jelinek <jakub@redhat.com>
22039 PR tree-optimization/57331
22040 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
22041 of conversion from pointer type to integral type with integer.
22043 2013-05-21 Martin Jambor <mjambor@suse.cz>
22046 * ipa-prop.c (ipa_read_node_info): Process param_used and
22047 controlled_uses in the same order as when writing.
22049 2013-05-21 Magnus Granberg <baldrick@free.fr>
22052 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
22054 2013-05-21 Richard Biener <rguenther@suse.de>
22056 PR tree-optimization/57318
22057 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
22058 estimate stmts with side-effects as likely eliminated.
22060 2013-05-21 Richard Biener <rguenther@suse.de>
22062 PR tree-optimization/57330
22063 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
22064 preserve the call stmts fntype.
22066 2013-05-21 Richard Biener <rguenther@suse.de>
22068 PR tree-optimization/57303
22069 * tree-ssa-sink.c (statement_sink_location): Improve killing
22070 stmt detection and properly handle self-assignments.
22072 2013-05-21 Christian Bruel <christian.bruel@st.com>
22074 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
22075 spanning registers. LEAF_REG_REMAP is supported only for contiguous
22076 registers. Set register size out of the PARALLEL loop.
22078 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
22081 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
22082 (*fmasf4, *fmasf4_media): New insns.
22084 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
22086 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
22087 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
22088 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
22089 (mips_idiv_insns): Update the comments to say that the returned
22090 instruction counts are in units of BASE_INSN_LENGTH.
22091 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
22092 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
22093 using 2 rather than 4 as the length of indirect MIPS16 and
22094 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
22095 length of a NOP. Don't divide MIPS16 lengths by 2.
22096 (mips16_split_long_branches): Assume a branch is long if the
22097 length is greater than 4 rather than 8.
22098 * config/mips/mips.md (length): Give MIPS16 lengths directly,
22099 rather than multiplying them by 2. Multiply instruction counts
22100 by BASE_INSN_LENGTH rather than 4.
22101 (*jump_mips16, tls_get_tp_mips16_<mode>)
22102 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
22104 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
22106 * config/mips/mips.md (extended_mips16): Remove branch case.
22107 (length): Remove duplicated extended_mips16 test.
22109 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
22111 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
22113 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
22115 * recog.h (Recog_data): Rename to...
22116 (recog_data_d): ...this.
22117 (recog_data): Update accordingly.
22118 * recog.c (recog_data): Likewise.
22119 * reload.c (save_recog_data): Likewise.
22120 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
22121 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
22123 2013-05-17 Julian Brown <julian@codesourcery.com>
22125 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
22126 found in a REG_EQUAL note, invalidate it.
22128 2013-05-17 Easwaran Raman <eraman@google.com>
22130 * tree-ssa-reassoc.c (find_insert_point): New function.
22131 (insert_stmt_after): Likewise.
22132 (get_def_stmt): Likewise.
22133 (ensure_ops_are_available): Likewise.
22134 (not_dominated_by): Likewise.
22135 (rewrite_expr_tree): Do not move statements beyond what is
22136 necessary. Remove call to swap_ops_for_binary_stmt...
22137 (reassociate_bb): ... and move it here.
22138 (build_and_add_sum): Assign UIDs for new statements.
22139 (linearize_expr): Likewise.
22140 (do_reassoc): Renumber gimple statement UIDs.
22142 2013-05-17 Jan Hubicka <jh@suse.cz>
22144 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
22146 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
22147 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
22149 (output_weakrefs): Update.
22151 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
22152 Martin Jambor <mjambor@suse.cz>
22154 PR middle-end/57276
22155 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
22156 value that corresponds to the given aggval is found in values vector.
22158 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
22160 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
22161 sse, sse2, sse3, ssse3 and sse4a flags to options.
22163 2013-05-17 David Malcolm <dmalcolm@redhat.com>
22165 * gengtype-state.c: (s_expr_writer): New class, to handle
22166 prettifying of output layout of s-expressions.
22167 (state_writer): New class, to write out gtype.state.
22168 (state_written_type_count): Move this variable into member data of
22170 (s_expr_writer::s_expr_writer): New code: constructor for new class
22171 (state_writer::state_writer(): ditto
22172 (s_expr_writer::write_new_line): New function
22173 (s_expr_writer::write_any_indent): ditto
22174 (s_expr_writer::begin_s_expr): ditto
22175 (s_expr_writer::end_s_expr): ditto
22176 (write_state_fileloc): convert to method of state_writer...
22177 (state_writer:: write_state_fileloc): ...and use methods of
22178 s_expr_writer to write indentation into the gtype.state output file
22179 to visually represent the hierarchical structure of the list
22181 (write_state_fields): ditto, renaming to...
22182 (state_writer::write_state_fields)
22183 (write_state_a_string): ditto, renaming to...
22184 (state_writer::write_state_a_string)
22185 (write_state_string_option): ditto, renaming to...
22186 (state_writer::write_state_string_option)
22187 (write_state_type_option): ditto, renaming to...
22188 (state_writer::write_state_type_option)
22189 (write_state_nested_option): ditto, renaming to...
22190 (state_writer::write_state_nested_option)
22191 (write_state_option): ditto, renaming to...
22192 (state_writer::write_state_option)
22193 (write_state_options): ditto, renaming to...
22194 (state_writer::write_state_options)
22195 (write_state_lang_bitmap): ditto, renaming to...
22196 (state_writer::write_state_lang_bitmap)
22197 (write_state_version): ditto, renaming to...
22198 (state_writer::write_state_version)
22199 (write_state_scalar_type): ditto, renaming to...
22200 (state_writer::write_state_scalar_type)
22201 (write_state_string_type): ditto, renaming to...
22202 (state_writer::write_state_string_type)
22203 (write_state_undefined_type): ditto, renaming to...
22204 (state_writer::write_state_undefined_type)
22205 (write_state_struct_union_type): ditto, renaming to...
22206 (state_writer::write_state_struct_union_type)
22207 (write_state_struct_type): ditto, renaming to...
22208 (state_writer::write_state_struct_type)
22209 (write_state_user_struct_type): ditto, renaming to...
22210 (state_writer::write_state_user_struct_type)
22211 (write_state_lang_struct_type): ditto, renaming to...
22212 (state_writer::write_state_lang_struct_type)
22213 (write_state_param_struct_type): ditto, renaming to...
22214 (state_writer::write_state_param_struct_type)
22215 (write_state_pointer_type): ditto, renaming to...
22216 (state_writer::write_state_pointer_type)
22217 (write_state_array_type): ditto, renaming to...
22218 (state_writer::write_state_array_type)
22219 (write_state_gc_used): ditto, renaming to...
22220 (state_writer::write_state_gc_used)
22221 (write_state_common_type_content): ditto, renaming to...
22222 (state_writer::write_state_common_type_content)
22223 (write_state_type): ditto, renaming to...
22224 (state_writer::write_state_type)
22225 (write_state_pair_list): ditto, renaming to...
22226 (state_writer::write_state_pair_list)
22227 (write_state_pair): ditto, renaming to...
22228 (state_writer::write_state_pair)
22229 (write_state_typedefs): ditto, renaming to...
22230 (state_writer::write_state_typedefs)
22231 (write_state_structures): ditto, renaming to...
22232 (state_writer::write_state_structures)
22233 (write_state_param_structs): ditto, renaming to...
22234 (state_writer::write_state_param_structs)
22235 (write_state_variables): ditto, renaming to...
22236 (state_writer::write_state_variables)
22237 (write_state_srcdir): ditto, renaming to...
22238 (state_writer::write_state_srcdir)
22239 (write_state_files_list): ditto, renaming to...
22240 (state_writer::write_state_files_list)
22241 (write_state_languages): ditto, renaming to...
22242 (state_writer::write_state_languages)
22243 (write_state): create a state_writer instance and use it when
22244 writing out the state file
22246 2013-05-17 Mike Stump <mikestump@comcast.net>
22248 PR rtl-optimization/57304
22249 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
22250 accessing DF_REF_REAL_LOC.
22252 2013-05-17 Jakub Jelinek <jakub@redhat.com>
22254 PR rtl-optimization/57281
22255 PR rtl-optimization/57300
22256 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
22257 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
22258 what the other splitter did if the registers are dead.
22260 2013-05-17 Richard Biener <rguenther@suse.de>
22262 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
22265 2013-05-17 Jakub Jelinek <jakub@redhat.com>
22267 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
22270 2013-05-17 Marek Polacek <polacek@redhat.com>
22272 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
22273 length when doing non-zero store of storing '\0' to '\0'.
22275 2013-05-17 Jakub Jelinek <jakub@redhat.com>
22277 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
22278 vect_external_def oprnd1 with loop_vinfo, try to emit
22279 optional cast, negation and and stmts on the loop preheader
22280 edge instead of into the pattern def seq.
22282 PR tree-optimization/57051
22283 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
22284 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
22286 2013-05-16 Nick Clifton <nickc@redhat.com>
22288 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
22289 (rl78_is_naked_func): New function.
22290 (rl78_expand_prologue): Skip prologue generation for naked functions.
22291 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
22292 * doc/extend.texi (naked): Add RL78 to the list of processors
22293 that supports this attribute.
22295 2013-05-16 Jeff Law <law@redhat.com>
22297 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
22299 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
22301 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
22302 cache parameters using detect_caches_amd also for CYRIX,
22303 NSC and TM2 signatures.
22305 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
22306 Dzianis Kahanovich <mahatma@eu.by>
22310 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
22311 VIA/Centaur processors and determine their cache parameters
22312 using detect_caches_amd.
22314 2013-05-16 Teresa Johnson <tejohnson@google.com>
22316 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
22317 (rtl_verify_edges): New function.
22318 (rtl_verify_bb_insns): Ditto.
22319 (rtl_verify_bb_pointers): Ditto.
22320 (rtl_verify_bb_insn_chain): Ditto.
22321 (rtl_verify_fallthru): Ditto.
22322 (rtl_verify_bb_layout): Ditto.
22323 (rtl_verify_flow_info_1): Outline checks into new functions.
22324 (rtl_verify_flow_info): Ditto.
22326 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
22328 * cfghooks.c (copy_bbs): Add update_dominance argument.
22329 * cfghooks.h (copy_bbs): Update prototype.
22330 * tree-cfg.c (gimple_duplicate_sese_region):
22331 Add update_dominance argument.
22332 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
22333 * tree-ssa-loop-ch.c (copy_loop_headers): Update
22334 gimple_duplicate_sese_region call.
22335 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
22336 Update copy_bbs call.
22337 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
22338 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
22340 2013-05-16 Jakub Jelinek <jakub@redhat.com>
22342 * tree-vectorizer.h (NUM_PATTERNS): Increment.
22343 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
22344 vect_recog_rotate_pattern.
22345 (vect_recog_rotate_pattern): New function.
22347 2013-05-16 Jason Merrill <jason@redhat.com>
22349 * Makefile.in (LLINKER): New variable.
22350 (mostlyclean): Remove link mutex.
22351 * configure.ac: Handle --enable-link-mutex.
22352 * lock-and-run.sh: New script.
22354 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22357 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
22360 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22362 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
22364 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
22366 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
22367 * config/arm/arm.c (next_consecutive_mem): New function.
22368 (gen_movmem_ldrd_strd): Likewise.
22369 * config/arm/arm.md (movmemqi): Update condition and code.
22370 (unaligned_loaddi, unaligned_storedi): New patterns.
22372 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22374 * config.gcc: Obsolete *-*-solaris2.9*.
22375 * doc/install.texi (Specific, *-*-solaris2*): Document it.
22377 2013-05-16 Richard Biener <rguenther@suse.de>
22379 * passes.c (init_optimization_passes): Move pass_parallelize_loops
22380 earlier, after GRAPHITE transforms and IV canonicalization.
22382 2013-05-16 Jakub Jelinek <jakub@redhat.com>
22384 * omp-low.c (extract_omp_for_data): For collapsed loops,
22385 if at least one of the loops is known at compile time to
22386 iterate zero times, set count to 0.
22387 (expand_omp_regimplify_p): New function.
22388 (expand_omp_for_generic): For collapsed loops, if at least
22389 one of the loops isn't known to iterate at least once,
22390 add runtime check with setting count to 0.
22391 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
22392 For unsigned types if it isn't known at compile time that
22393 the loop will iterate at least once, add runtime check to bypass
22394 the whole loop if initial condition isn't true.
22396 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
22398 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
22400 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
22402 PR middle-end/57286
22403 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
22404 transformations to avoid an infinite loop.
22406 2013-05-16 Marek Polacek <polacek@redhat.com>
22408 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
22410 2013-05-15 Leif Ekblad <leif@rdos.net>
22412 * config/i386/i386.c (ix86_decompose_address): Use
22413 DEFAULT_TLS_SEG_REG to access TLS segment register.
22414 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
22415 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
22416 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
22418 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
22421 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
22422 sibling calls to functions that would normally be lazily bound,
22423 unless $gp is call-clobbered.
22425 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
22427 * config/i386/i386.c (ix86_option_override_internal): Update
22428 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
22429 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
22430 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
22431 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
22433 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
22435 2013-05-15 Andreas Schwab <schwab@suse.de>
22437 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
22438 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
22441 2013-05-15 Teresa Johnson <tejohnson@google.com>
22443 * loop-unroll.c (report_unroll_peel): Check decision before
22444 emitting unroll/peel message.
22446 2013-05-15 Teresa Johnson <tejohnson@google.com>
22448 * function.h (has_bb_partition): New rtl_data flag.
22449 (bb_reorder_complete): Ditto.
22450 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
22451 instead of flag_reorder_blocks_and_partition.
22452 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
22453 with some enhancements.
22454 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
22455 * bb-reorder.c (connect_traces): Check for has_bb_partition
22456 instead of flag_reorder_blocks_and_partition.
22457 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
22458 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
22459 verify_hot_cold_block_grouping.
22460 (partition_hot_cold_basic_blocks): Set has_bb_partition.
22462 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22465 * config/arm/predicates.md (call_insn_operand): New predicate.
22466 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
22467 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
22468 if insn is not a tail call.
22469 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
22471 * config/arm/arm.h (enum reg_class): New caller save register class.
22472 (REG_CLASS_NAMES): Likewise.
22473 (REG_CLASS_CONTENTS): Likewise.
22474 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
22477 2013-05-15 Richard Biener <rguenther@suse.de>
22479 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
22480 of MSG_OPTIMIZED_LOCATIONS.
22481 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
22482 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
22484 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
22485 of MSG_OPTIMIZED_LOCATIONS.
22486 (execute_vect_slp): Likewise.
22487 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
22488 (vect_create_cond_for_alias_checks): Likewise.
22489 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
22490 (vect_recog_widen_mult_pattern): Likewise.
22491 (vect_recog_widen_sum_pattern): Likewise.
22492 (vect_recog_over_widening_pattern): Likewise.
22493 (vect_recog_widen_shift_pattern): Likewise.
22494 (vect_recog_vector_vector_shift_pattern): Likewise.
22495 (vect_recog_divmod_pattern): Likewise.
22496 (vect_recog_mixed_size_cond_pattern): Likewise.
22497 (vect_recog_bool_pattern): Likewise.
22498 (vect_pattern_recog_1): Likewise.
22500 2013-05-15 Martin Jambor <mjambor@suse.cz>
22502 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
22503 non-functions to builtin_unreachable.
22504 * ipa-inline-transform.c (inline_call): Do not assert estimates were
22505 correct when new direct edges were discovered.
22507 2013-05-15 Martin Jambor <mjambor@suse.cz>
22509 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
22510 header, print symbol order instead of node uid, print more information
22511 about indirect edge targets.
22512 (ipa_make_edge_direct_to_target): Print symbol order instead of node
22514 (ipa_make_edge_direct_to_target): Likewise.
22515 (remove_described_reference): Likewise.
22516 (propagate_controlled_uses): Likewise.
22517 (ipa_print_node_params): Also print symbol order.
22518 (ipcp_transform_function): Print symbol order instead of node uids.
22519 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
22520 (cgraph_get_create_real_symbol_node): Likewise.
22521 * ipa-cp.c (print_lattice): Likewise.
22522 (print_all_lattices): Likewise.
22523 (determine_versionability): Likewise.
22524 (initialize_node_lattices): Likewise.
22525 (estimate_local_effects): Likewise.
22526 (update_profiling_info): Likewise.
22527 (create_specialized_node): Likewise.
22528 (perhaps_add_new_callers): Likewise.
22529 (decide_about_value): Likewise.
22530 (decide_whether_version_node): Likewise.
22531 (identify_dead_nodes): Likewise.
22532 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
22533 (dump_inline_summary): Likewise.
22534 (estimate_node_size_and_time): Likewise.
22535 (inline_analyze_function): Likewise.
22536 * ipa-inline.c (report_inline_failed_reason): Likewise.
22537 (want_early_inline_function_p): Likewise.
22538 (edge_badness): Likewise.
22539 (update_edge_key): Likewise.
22540 (inline_small_functions): Likewise. Add dumping of order to two other
22542 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
22543 instead of node uids.
22544 (propagate_pure_const): Likewise.
22545 (propagate_pure_const): Likewise.
22546 * ipa-utils.c (dump_cgraph_node_set): Likewise.
22547 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
22548 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
22550 * tree-pretty-print.c (dump_function_header): Likewise.
22551 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
22552 Print symbol order instead of node uids.
22554 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22556 * config/s390/s390.c (s390_register_move_cost): Don't impose the
22557 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
22559 2013-05-15 Richard Biener <rguenther@suse.de>
22561 PR tree-optimization/57275
22562 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
22563 return value for fail to do runtime alias checks for gather loads.
22565 2013-05-15 Jan Hubicka <jh@suse.cz>
22569 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
22570 weakrefs are not external.
22571 (lto_symtab_merge_decls): Fix thinko when dealing with
22572 non-lto_symtab decls.
22573 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
22574 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
22575 * varpool.c (dump_varpool_node): Dump more flags.
22577 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22579 * config/i386/i386.c (processor_alias_table): Add instruction
22580 FSGSBASE for AMD bdver3 architecture.
22582 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22584 * tree.c (warn_deprecated_use): Print file:line using locus color.
22585 * diagnostic.c (diagnostic_report_current_module): Print file:line
22586 and file:line:column using locus color.
22588 2013-05-14 Mike Stump <mikestump@comcast.net>
22590 * gdbinit.in: Add __null.
22592 2013-05-14 Mike Stump <mikestump@comcast.net>
22594 * recog.h: Rename struct recog_data to Recog_data.
22595 * recog.c: Likewise.
22596 * reload.c (can_reload_into): Likewise.
22597 * config/picochip/picochip.c: Likewise.
22599 2013-05-14 Mike Stump <mikestump@comcast.net>
22601 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
22603 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
22605 * resource.h (struct resources): Remove unch_memory member.
22606 (CLEAR_RESOURCE): Don't clear unch_memory.
22607 * resource.c (mark_referenced_resources): Don't set it.
22608 (mark_set_resources): Likewise.
22609 (mark_target_live_regs): Don't clear it.
22610 (init_resource_info): Likewise.
22611 * reorg.c (resource_conflicts_p): Don't compare it.
22612 (redundant_insn): Don't set it.
22614 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
22616 * emit-rtl.c (next_label): Remove unused function.
22617 (skip_consecutive_labels, link_cc0_insns): Move to ...
22618 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
22619 only place where these functions are used, and make them static.
22621 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
22623 * fold-const.c (fold_negate_expr): Handle vectors.
22624 (fold_truth_not_expr): Make it static.
22625 (fold_invert_truthvalue): New static function.
22626 (invert_truthvalue_loc): Handle vectors. Do not call
22627 fold_truth_not_expr directly.
22628 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
22629 <TRUTH_NOT_EXPR>: Do not cast to boolean.
22630 (fold_comparison): Handle vector constants.
22631 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
22632 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
22633 * tree.h (fold_truth_not_expr): Remove declaration.
22635 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
22637 * config/aarch64/aarch64-simd.md
22638 (aarch64_vcond_internal<mode>): Rename to...
22639 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
22640 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
22641 float modes. Clarify all iterator modes.
22642 (vcond<mode><mode>): Use new name for vcond expanders.
22643 (vcond<v_cmp_result><mode>): Likewise.
22644 (vcondu<mode><mode>: Likewise.
22645 * config/aarch64/iterators.md (VDQF_COND): New.
22647 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
22650 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
22651 variable for the shift amount. Check that we shift by non-negative
22654 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
22657 * config/arm/arm.h (EPILOGUE_USES): Only return true
22658 for LR_REGNUM after epilogue_completed.
22660 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
22662 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
22663 is error_mark_node.
22665 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22668 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
22669 and Solaris 11+/x86 with gld.
22670 * configure: Regenerate.
22672 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22674 * expmed.c (expand_shift_1): Canonicalize rotates by
22675 constant bitsize / 2 to bitsize - 1.
22676 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
22677 case ROTATERT>: Likewise.
22680 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22682 * config/i386/i386.md (rotateinv): New code attr.
22683 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22684 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22685 roll $31, %eax, etc.
22687 2013-05-14 Richard Biener <rguenther@suse.de>
22689 PR middle-end/57235
22690 * tree-eh.c (sink_clobbers): Give up for successors with
22691 multiple predecessors and no virtual uses.
22693 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
22695 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
22696 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
22698 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22700 PR middle-end/57251
22701 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
22702 the case when both op0 and op1 have VOIDmode.
22704 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
22706 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
22707 in multiply-accumulate mode.
22709 2013-05-13 Guozhi Wei <carrot@google.com>
22711 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
22713 2013-05-13 Kai Tietz <ktietz@redhat.com>
22716 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
22717 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
22718 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
22719 * config/i386/i386.c (ix86_option_override_internal): Likewise.
22720 (ix86_expand_prologue): Likewise.
22721 (ix86_expand_split_stack_prologue): Likewise.
22722 (legitimate_pic_address_disp_p): Likewise.
22723 (legitimize_pic_address): Likewise.
22724 (legitimize_tls_address): Likewise.
22725 (legitimize_pe_coff_symbol): Likewise.
22726 (output_pic_addr_const): Likewise.
22727 (construct_plt_address): Likewise.
22728 (ix86_expand_call): Likewise.
22729 (x86_output_mi_thunk): Likewise.
22730 (x86_function_profiler): Likewise.
22732 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
22734 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
22735 similar switch cases.
22736 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
22737 (aarch64_simd_mov_to_<mode>low): Delete.
22738 (aarch64_simd_mov_to_<mode>high): Delete.
22739 (move_lo_quad_<mode>): Add w<-r alternative.
22740 (aarch64_simd_move_hi_quad_<mode>): Likewise.
22741 (aarch64_simd_mov_from_*): Update type attribute.
22742 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
22745 2013-05-13 Jan Hubicka <jh@suse.cz>
22747 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
22748 * config/i386/i386.c (ix86_compute_frame_layout,
22749 ix86_expand_epilogue, emit_i387_cw_initialization,
22750 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
22751 ix86_local_alignment): Fix use of size/speed predicates.
22753 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22755 PR tree-optimization/45216
22756 PR tree-optimization/57157
22757 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
22758 the (-Y) & (B - 1) variant if OP is |.
22759 * expmed.c (expand_shift_1): For rotations by const0_rtx just
22760 return shifted. Use (-op1) & (prec - 1) as other_amount
22761 instead of prec - op1.
22763 2013-05-13 Martin Jambor <mjambor@suse.cz>
22765 PR middle-end/42371
22766 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
22767 (ipa_constant_data): New type.
22768 (ipa_jump_func): Use ipa_constant_data to hold information about
22769 constant jump functions.
22770 (ipa_get_jf_constant): Adjust to jump function type changes.
22771 (ipa_get_jf_constant_rdesc): New function.
22772 (ipa_param_descriptor): New field controlled_uses.
22773 (ipa_get_controlled_uses): New function.
22774 (ipa_set_controlled_uses): Likewise.
22775 * ipa-ref.h (ipa_find_reference): Declare.
22776 * ipa-prop.c (ipa_cst_ref_desc): New type.
22777 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
22779 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
22780 New parameter cs. Adjust all callers.
22781 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
22782 (remove_described_reference): New function.
22783 (jfunc_rdesc_usable): Likewise.
22784 (try_make_edge_direct_simple_call): Decrement controlled use count,
22785 attempt to remove reference if it hits zero.
22786 (combine_controlled_uses_counters): New function.
22787 (propagate_controlled_uses): Likewise.
22788 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
22789 (ipa_edge_duplication_hook): Duplicate reference descriptions.
22790 (ipa_print_node_params): Print described use counter.
22791 (ipa_write_jump_function): Adjust to jump function type changes.
22792 (ipa_read_jump_function): New parameter CS, pass it to
22793 ipa_set_jf_constant. Adjust caller.
22794 (ipa_write_node_info): Stream controlled use count
22795 (ipa_read_node_info): Likewise.
22796 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
22798 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
22799 count. Remove cloning-added reference if it reaches zero.
22800 * ipa-ref.c (ipa_find_reference): New function.
22802 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22804 * config/i386/i386.c (processor_target_table): Modified default
22805 alignment values for AMD BD and BT architectures.
22807 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
22809 * tree-vect-generic.c (uniform_vector_p): Move ...
22810 * tree.c (uniform_vector_p): ... here.
22811 * tree.h (uniform_vector_p): Declare it.
22812 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
22815 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22817 PR tree-optimization/57230
22818 * tree-ssa-strlen.c (handle_char_store): Record length for
22819 array store from STRING_CST.
22821 PR tree-optimization/57230
22822 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
22825 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
22827 * config/epiphany/epiphany.c (epiphany_init): Check size of
22828 NUM_MODES_FOR_MODE_SWITCHING.
22829 (epiphany_expand_prologue):
22830 Remove CONFIG_REGNUM initial value handling code.
22831 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
22832 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
22833 (emit_set_fp_mode, epiphany_mode_after): Likewise.
22834 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
22835 Don't return 1 for FP_MODE_NONE.
22836 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
22837 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
22838 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
22839 * config/epiphany/epiphany.md (save_config): New pattern.
22841 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
22843 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
22845 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
22847 * config/i386/i386.md (memory): Handle sseishft1.
22848 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
22849 (*vec_extractv2di_1): Ditto.
22851 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
22853 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
22856 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
22858 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
22859 Add mthumb/march=armv7-a multilib.
22860 Add mthumb/march=armv7-r multilib.
22861 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
22863 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
22865 * config/v850/t-rtems: Add more multilibs.
22867 2013-05-10 Richard Biener <rguenther@suse.de>
22869 PR tree-optimization/57214
22870 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
22871 not propagate from SSA names that occur in abnormal PHI nodes.
22873 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
22875 * stor-layout.c (element_precision): New function.
22876 * machmode.h (element_precision): Declare it.
22877 * tree.c (build_minus_one_cst): New function.
22878 (element_precision): Likewise.
22879 * tree.h (build_minus_one_cst): Declare new function.
22880 (element_precision): Likewise.
22881 * fold-const.c (operand_equal_p): Use element_precision.
22882 (fold_binary_loc): Handle vector types.
22883 * convert.c (convert_to_integer): Use element_precision.
22884 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
22887 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
22889 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
22890 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
22891 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
22892 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22893 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
22894 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
22895 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
22896 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22897 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
22898 (Uuw8): New constraints.
22899 (Usb4): Move into alphabetical order.
22900 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
22901 (sd8_operand, ub8_operand, uw8_operand): New predicates.
22902 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
22903 previously unnamed patterns.
22904 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
22905 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
22906 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
22907 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
22908 of set_attr_alternative/if_then_else. Use extended_mips16 instead
22909 of specific lengths.
22911 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22913 * config/i386/i386.md (rotateinv): New code attr.
22914 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22915 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22916 roll $31, %eax, etc.
22918 PR tree-optimization/45216
22919 PR tree-optimization/57157
22920 * tree-ssa-forwprop.c (simplify_rotate): New function.
22921 (ssa_forward_propagate_and_combine): Call it.
22923 2013-05-10 Richard Biener <rguenther@suse.de>
22925 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
22926 disable peeling when we version for aliasing.
22927 (vector_alignment_reachable_p): Honor explicit user alignment.
22928 (vect_supportable_dr_alignment): Likewise.
22929 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
22930 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
22931 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
22932 then peeling to arrange for the cost-model check to come first.
22934 2013-05-10 Alan Modra <amodra@gmail.com>
22936 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
22937 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
22938 * configure: Regenerate.
22940 2013-05-10 Alan Modra <amodra@gmail.com>
22943 * varasm.c (default_elf_select_section): Move !DECL_P check..
22944 (get_named_section): ..to here before calling get_section_name.
22946 (default_section_type_flags): Add DECL_P check.
22947 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
22948 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
22950 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
22952 * config/epiphany/epiphany.c (epiphany_expand_prologue):
22953 When using gen_stack_adjust_str with a register offset, add a
22954 REG_FRAME_RELATED_EXPR note.
22956 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
22958 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
22959 (*vec_extractv4si_zext_mem): Ditto.
22960 (*vec_extractv2di): Add 0->x and x->x alternatives.
22961 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
22962 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
22964 2013-05-09 Jason Merrill <jason@redhat.com>
22966 N3639 C++1y VLA support
22967 * gimplify.c (gimplify_vla_decl): Don't touch an existing
22970 * tree.c (build_constructor_va): New.
22971 * tree.h: Declare it.
22973 2013-05-09 Martin Jambor <mjambor@suse.cz>
22976 * gimple-fold.c (canonicalize_constructor_val): Call
22977 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
22979 2013-05-09 Jan Hubicka <jh@suse.cz>
22980 Richard Biener <rguenther@suse.de>
22983 * symtab.c (symtab_make_decl_local): Do not add private names.
22985 2013-05-09 Jan Hubicka <jh@suse.cz>
22988 * symtab.c (insert_to_assembler_name_hash): Handle clones.
22989 (unlink_from_assembler_name_hash): Likewise.
22990 (symtab_prevail_in_asm_name_hash, symtab_register_node,
22991 symtab_unregister_node, symtab_initialize_asm_name_hash,
22992 change_decl_assembler_name): Update.
22994 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
22996 * config/aarch64/aarch64.md: New movtf split.
22997 (*movtf_aarch64): Update.
22998 (aarch64_movdi_tilow): Handle TF modes and rename to
22999 aarch64_movdi_<mode>low.
23000 (aarch64_movdi_tihigh): Handle TF modes and rename to
23001 aarch64_movdi_<mode>high
23002 (aarch64_movtihigh_di): Handle TF modes and rename to
23003 aarch64_mov<mode>high_di
23004 (aarch64_movtilow_di): Handle TF modes and rename to
23005 aarch64_mov<mode>low_di
23006 (aarch64_movtilow_tilow): Remove spurious whitespace.
23007 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
23009 (aarch64_print_operand): Update.
23011 2013-05-09 Alan Modra <amodra@gmail.com>
23013 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
23015 * configure: Regenerate.
23017 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
23019 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
23020 splitter preparation statements.
23021 * config/i386/sse.md (*vec_extract* splitters): Ditto.
23022 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
23025 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23027 * gimple-ssa-strength-reduction.c (count_candidates): Change
23028 return value to int.
23029 (analyze_candidates_and_replace): Change type of length to int.
23031 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
23033 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
23034 (*vec_extract<mode>): Use VI12_128 mode iterator.
23035 (*vec_extract<mode>_mem): Ditto.
23036 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
23039 2013-05-08 Diego Novillo <dnovillo@google.com>
23044 2012-08-17 Diego Novillo <dnovillo@google.com>
23047 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
23048 * config.in: Regenerate.
23049 * configure: Regenerate.
23050 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
23052 2013-05-08 Jan Hubicka <jh@suse.cz>
23055 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
23056 * cgraph.h (symtab_node_base): Add unique_name.
23057 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
23058 input_overwrite_node, input_varpool_node): Stream unique_name.
23059 * cgraphclones.c (cgraph_create_virtual_clone,
23060 cgraph_function_versioning): Set unique_name.
23061 * ipa.c (function_and_variable_visibility): Set unique_name.
23063 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23065 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
23066 (alloc_cand_and_find_basis): Restrict conditional candidate
23067 processing to CAND_MULTs.
23069 2013-05-08 Jan Hubicka <jh@suse.cz>
23072 lto-symtab.c (lto_symtab_symbol_p): New function.
23073 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
23074 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
23075 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
23076 Skip static symbols.
23078 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
23080 PR tree-optimization/57200
23081 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
23082 Only call inform if the preceding warning_at returns true.
23084 2013-05-07 Han Shen <shenhan@google.com>
23086 * cfgexpand.c (record_or_union_type_has_array_p): New function.
23087 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
23088 * common.opt (fstack-protector-strong): New option.
23089 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
23090 * doc/invoke.texi (Optimization Options): Document
23091 "-fstack-protector-strong".
23092 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
23094 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
23096 * config/mips/mips.c (mips_machine_reorg2): Return 0.
23098 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
23100 * ira.c (update_equiv_regs): Add insn having equiv memory even if
23101 it is not lhs of the insn.
23102 (setup_reg_equiv): Remove insn having equiv memory which it is not
23104 * lra-constraints.c (process_address): Try to improve generation
23105 code for address base + disp.
23106 (lra_constraints): Make correct the code for checking insn setting
23107 up backward equivalence. Remove insn only if it is in the init
23109 * lra-eliminations.c (update_reg_eliminate): Change return value.
23110 (lra_eliminate): Use the result.
23112 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
23114 * config/i386/sse.md (ssescalarnummask): New mode attribute.
23115 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
23116 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
23117 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
23118 register target operands.
23119 (*vec_extractv8hi_sse2): New pattern.
23120 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
23121 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
23122 (*vec_extract<mode>_mem): New insn and split pattern.
23124 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
23126 * config/arm/arm.c (arm_asan_shadow_offset): New function.
23127 (TARGET_ASAN_SHADOW_OFFSET): Define.
23128 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
23129 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
23131 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23133 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
23134 (incr_vec_index): Return -1 if increment not found.
23135 (create_add_on_incoming_edge): Assert if increment not found.
23136 (record_increment): Limit number of increments recorded.
23137 (all_phi_incrs_profitable): Return false if an increment not found.
23138 (replace_profitable_candidates): Don't process increments that were
23140 (analyze_candidates_and_replace): Limit size of incr_vec.
23142 2013-05-07 Richard Biener <rguenther@suse.de>
23144 * calls.c (special_function_p): setjmp-like functions are leaf.
23145 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
23146 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
23148 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
23150 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
23151 (aarch64_simd_mov<mode>): New expander.
23152 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
23153 (aarch64_simd_mov_to_<mode>high): Likewise.
23154 (aarch64_simd_mov_from_<mode>low): Likewise.
23155 (aarch64_simd_mov_from_<mode>high): Likewise.
23156 (aarch64_dup_lane<mode>): Update.
23157 (aarch64_dup_lanedi): New instruction pattern.
23158 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
23159 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
23161 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23163 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
23164 (replace_mult_candidate): Remove unnecessary argument; remove
23165 unnecessary parameter from call to introduce_cast_before_cand.
23166 (replace_unconditional_candidate): Remove unnecessary parameter
23167 from call to replace_mult_candidate.
23168 (replace_conditional_candidate): Likewise.
23169 (insert_initializers): Use make_temp_ssa_name.
23170 (introduce_cast_before_cand): Remove unnecessary argument; use
23171 make_temp_ssa_name.
23172 (replace_one_candidate): Remove unnecessary argument; remove
23173 unnecessary parameter from calls to introduce_cast_before_cand.
23174 (replace_profitable_candidates): Remove unnecessary parameters
23175 from calls to replace_one_candidate.
23177 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23179 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
23180 phi def as possibly hiding a basis for a CAND_ADD whose operands
23181 have been commuted in the analysis.
23182 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
23184 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23186 * config/aarch64/aarch64.md
23187 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
23188 shift value between 0-4.
23190 2013-05-07 Richard Biener <rguenther@suse.de>
23192 * double-int.h (rshift): New overload.
23193 * double-int.c (rshift): New function.
23194 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
23195 (create_reference_ops_from_ref): Remove.
23196 (vn_reference_insert): Use shared ops for constructing the
23197 reference and copy it.
23199 2013-05-07 Richard Biener <rguenther@suse.de>
23201 PR middle-end/57190
23202 * tree-eh.c (sink_clobbers): Properly propagate
23203 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
23205 2013-05-07 Jakub Jelinek <jakub@redhat.com>
23207 PR tree-optimization/57149
23208 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
23209 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
23210 collect_phi_def_edges, execute_late_warn_uninitialized): Use
23211 uninit_undefined_value_p instead of ssa_undefined_value_p.
23214 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
23215 for modifier == EXPAND_INITIALIZER.
23217 2013-05-07 Anton Blanchard <anton@samba.org>
23219 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
23220 for powerpc64 little endian.
23221 * configure: Regenerate.
23223 2013-05-06 Graham Stott <grahams@btinternet.com>
23225 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
23226 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
23227 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
23228 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
23230 2013-05-06 Graham Stott <grahams@btinternet.com>
23232 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
23233 codes which allow non-lvalues.
23235 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
23237 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
23238 components are all 1s.
23239 (integer_minus_onep): New function.
23240 * tree.h (integer_minus_onep): Declare it.
23241 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
23242 integer_minus_onep instead of integer_all_onesp.
23244 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
23247 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
23248 variations of these patterns.
23250 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
23252 * config/i386/i386.md (isa): Add x64_sse4 member.
23253 (enabled): Handle x64_sse4.
23254 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
23255 instruction for 64bit SSE4_1 targets. Update insn attributes.
23256 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
23257 instruction for SSE4_1 targets. Update insn attributes.
23258 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
23259 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
23260 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
23262 (*vec_extractv4si): Rename from *sse4_1_pextrd.
23263 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
23264 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
23266 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
23269 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
23271 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
23273 * final.c (do_assembler_dialects): Don't handle curly braces and
23274 vertical bar escaped by % as dialect delimiters.
23275 (output_asm_insn): Print curly braces and vertical bar if escaped
23276 by % and ASSEMBLER_DIALECT defined.
23277 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
23278 * doc/tm.texi: Regenerated.
23280 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
23282 * config/mips/mips.c: Include tree-pass.h.
23283 (mips_reorg): Split in pre- and post-dbr_schedule parts.
23284 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
23285 (pass_mips_machine_reorg2): New machine specific pass.
23286 (insert_pass_mips_machine_reorg2): New pass plugin definition.
23287 (mips_option_override): Register the new pass.
23288 * rtl.h (cleanup_barriers): Remove prototype.
23289 (dbr_schedule): Likewise.
23290 * jump.c (cleanup_barriers): Make static.
23291 * reorg.c (dbr_schedule): Likewise.
23293 2013-05-06 Richard Biener <rguenther@suse.de>
23295 PR tree-optimization/57185
23296 * tree-parloops.c (add_field_for_reduction): Handle anonymous
23297 SSA names properly.
23299 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
23302 * config/i386/i386.c (add_parameter_dependencies): Add dependence
23303 between "first_arg" and "insn", not "last" and "insn".
23305 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23307 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
23308 (find_candidates_in_block): Re-enable slsr_process_phi.
23309 (create_phi_basis): Fix double counting of candidate adjustment.
23311 2013-05-06 Richard Biener <rguenther@suse.de>
23313 PR middle-end/57147
23314 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
23315 the edge is also fallthru, preserve it and just clear the
23317 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
23318 also complex, preserve that and just clear the fallthru flag.
23319 * tree-inline.c (update_ssa_across_abnormal_edges): Also
23320 update virtual operands.
23322 2013-05-06 Alan Modra <amodra@gmail.com>
23324 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
23325 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
23326 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
23327 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
23328 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
23329 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
23331 2013-05-06 Alan Modra <amodra@gmail.com>
23333 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
23334 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
23335 (DEFAULT_ASM_ENDIAN): Define.
23336 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
23337 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
23338 Update -K PIC clause from sysv4.h.
23339 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
23340 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
23342 2013-05-06 Alan Modra <amodra@gmail.com>
23344 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
23345 twice for little-endian.
23346 (ashrdi3_no_power, ashrdi3): Support little-endian.
23348 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
23351 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
23352 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
23354 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
23355 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
23356 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
23359 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
23360 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23362 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
23363 * config/pa/pa.opt: Make mbig-switch a no-op.
23364 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
23365 (CASE_VECTOR_MODE): Always return SImode.
23366 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
23367 for the !TARGET_BIG_SWITCH case.
23368 * config/pa/pa-linux.h: Likewise.
23369 * config/pa/pa-openbsd.h: Likewise.
23370 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
23371 * config/pa/pa.md (short_jump): Remove define_insn.
23372 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
23373 (casesi0): Remove define_insn.
23374 (type): Remove btable_branch.
23375 (pa_combine_type): Likewise.
23376 (in_nullified_branch_delay): Likewise.
23377 (in_call_delay): Likewise.
23378 (define_delay): Likewise.
23379 (define_insn_reservation "Z3"): Likewise.
23380 (define_insn_reservation "Z4"): Likewise.
23381 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
23382 (pa_adjust_insn_length): Remove adjustment for btable branches.
23383 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
23386 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
23388 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
23389 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
23390 Add m->r,x alternatives.
23391 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
23392 splitters using SWI48x mode iterator.
23393 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
23394 TARGET_64BIT. Add m->x alternative.
23395 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
23396 Add o->x alternative. Enable for TARGET_SSE.
23397 (sse_storeq): Remove expander.
23398 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
23399 with memory input operand.
23400 (*vec_extractv2di_1 splitter): New.
23401 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
23402 * config/i386/i386.md (ssevecmodelower): New mode attribute.
23404 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
23406 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
23407 (INT_LOWPART): Delete.
23408 (extract_MB): Adjust.
23409 (extract_ME): Adjust.
23410 (print_operand): Adjust.
23412 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
23414 * config/rs6000/predicates.md (reg_or_add_cint_operand,
23415 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
23416 (reg_or_logical_cint_operand, easy_fp_constant,
23417 logical_const_operand): Delete "CONST_DOUBLE" case.
23418 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
23419 "HOST_BITS_PER_WIDE_INT == 64" test.
23420 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
23421 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
23422 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
23423 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
23425 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
23426 CONST_DOUBLE DImode/VOIDmode case.
23427 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
23428 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
23429 CONST_DOUBLE VOIDmode case.
23430 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
23431 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
23432 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
23433 Delete CONST_DOUBLE case.
23434 (splitters for mov FMOVE64 const_double): Delete
23435 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
23436 "HOST_BITS_PER_WIDE_INT >= 64" test.
23437 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
23439 (mov DI const_double): Delete.
23441 2013-05-04 Jakub Jelinek <jakub@redhat.com>
23443 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
23444 on op shows all bits zero in mode of a lowpart subreg, return zero.
23446 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
23449 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
23450 to save TFmode registers and DImode to save TImode registers for
23451 caller save operations.
23452 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
23453 mark being partially clobbered since they only use the first
23456 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
23457 and TDmode only use the upper 64-bits of each VSX register.
23459 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23461 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
23462 (find_candidates_in_block): Disable slsr_process_phi.
23464 2013-05-03 Guozhi Wei <carrot@google.com>
23466 * coverage.c (coverage_obj_init): Move the construction of gcov
23468 (build_init_ctor): ... here.
23470 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23472 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
23473 (slsr_cand_d): Redefine def_phi.
23474 (stride_status, phi_adjust_status, count_phis_status): New enums.
23475 (find_phi_def): New.
23476 (find_basis_for_base_expr): New.
23477 (find_basis_for_candidate): Handle hidden bases.
23478 (alloc_cand_and_find_basis): Handle phi candidates.
23479 (slsr_process_phi): New.
23480 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
23481 (create_mul_imm_cand): Likewise.
23482 (create_add_ssa_cand): Exclude phi base candidates.
23483 (create_add_imm_cand): Likewise.
23484 (slsr_process_cast): Likewise.
23485 (slsr_process_copy): Likewise.
23486 (find_candidates_in_block): Handle phi candidates.
23487 (dump_candidate): Likewise.
23488 (unconditional_cands): Delete.
23489 (unconditional_cands_with_known_stride_p): Delete.
23490 (phi_dependent_cand_p): New.
23491 (cand_increment): Handle phi-dependent candidates.
23492 (replace_dependent): Delete.
23493 (replace_mult_candidate): New.
23494 (replace_unconditional_candidate): New.
23495 (incr_vec_index): Move to avoid forward reference.
23496 (create_add_on_incoming_edge): New.
23497 (create_phi_basis): New.
23498 (replace_dependents): Delete.
23499 (replace_conditional_candidate): New.
23500 (phi_add_costs): New.
23501 (replace_uncond_cands_and_profitable_phis): New.
23502 (record_increment): Handle phi adjustments.
23503 (record_phi_increments): New.
23504 (record_increments): Handle phi adjustments.
23505 (phi_incr_cost): New.
23506 (lowest_cost_path): Handle phis.
23507 (total_savings): Likewise.
23508 (analyze_increments): Likewise.
23509 (ncd_with_phi): New.
23510 (ncd_of_cand_and_phis): New.
23511 (nearest_common_dominator_for_cands): Handle phi increments.
23512 (all_phi_incrs_profitable): New.
23513 (replace_profitable_candidates): Handle phi-dependent candidates.
23514 (analyze_candidates_and_replace): Likewise.
23516 2013-05-03 Teresa Johnson <tejohnson@google.com>
23519 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
23520 do not exceed REG_BR_PROB_BASE.
23522 2013-05-03 Jeff Law <law@redhat.com>
23524 PR tree-optimization/57144
23525 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
23526 operand of the condition will bit into the new type when eliminating
23527 a cast feeding a condition.
23529 2013-05-03 Jakub Jelinek <jakub@redhat.com>
23531 PR rtl-optimization/57130
23532 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
23533 of COMPARE as in_code to the recursive call if needed.
23535 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
23537 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
23538 (enabled): Handle new members.
23539 * config/i386/sse.md (*vec_concatv2si): Merge from
23540 *vec_concatv2si_sse2 and vec_concatv2si_sse.
23541 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
23543 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
23545 PR tree-optimization/57027
23546 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
23547 for fnms opportunity, check we got the prerequisite kind
23548 of tree / gimple before using accessor functions.
23550 2013-05-03 Richard Biener <rguenther@suse.de>
23552 * double-int.h (lshift): New overload without precision
23553 and arith argument.
23554 (operator *=, operator +=, operator -=): Move ...
23555 * double-int.c (operator *=, operator +=, operator -=): ... here
23556 and implement more efficiently.
23557 (mul_double_with_sign): Remove.
23558 (lshift_double): Adjust to take unsinged shift argument, push
23559 dispatching code to callers.
23560 (mul_double_wide_with_sign): Add early out for callers that
23561 are not interested in high parts or overflow.
23562 (lshift): New function.
23563 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
23564 dispatch code here.
23565 (lrotate, rrotate): Use logical shifts.
23566 * expr.c (get_inner_reference): Use lshift.
23567 * fixed-value.c (do_fixed_divide): Likewise.
23568 * tree-dfa.c (get_ref_base_and_extent): Likewise.
23569 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
23570 (indirect_refs_may_alias_p): Likewise.
23571 (stmt_kills_ref_p_1): Likewise.
23573 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
23575 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
23577 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
23579 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
23580 scalar form of FABD instruction.
23582 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
23584 * lra-constraints.c (process_alt_operands): Add checking alt
23585 number to choose the best alternative.
23587 2013-05-02 Richard Biener <rguenther@suse.de>
23589 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
23590 bitmap and its handling.
23591 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
23593 2013-05-02 Richard Biener <rguenther@suse.de>
23595 PR middle-end/57140
23596 * tree-inline.c (copy_loops): Properly handle removed loops.
23597 (copy_cfg_body): Mark destination loops for fixup if source
23598 loops needed fixup.
23600 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
23603 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
23604 generating simple_return for naked functions.
23606 2013-05-02 Martin Jambor <mjambor@suse.cz>
23608 PR middle-end/56988
23609 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
23610 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
23612 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
23613 ipa_agg_replacement_value structures.
23614 (known_aggs_to_agg_replacement_list): Likewise.
23615 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
23616 (read_agg_replacement_chain): Likewise.
23617 (ipcp_transform_function): Also check that by_ref flags match.
23619 2013-05-02 Richard Biener <rguenther@suse.de>
23621 * graphds.h (struct graph): Add obstack member.
23622 * graphds.c (new_graph): Initialize obstack and allocate
23624 (add_edge): Allocate edge from the obstack.
23625 (free_graph): Free the obstack instead of all edges and vertices.
23627 2013-05-02 Teresa Johnson <tejohnson@google.com>
23629 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
23631 * cfg.c (update_bb_profile_for_threading): Ditto.
23632 * tree-inline.c (copy_bb): Ditto.
23633 (copy_edges_for_bb): Ditto.
23634 (initialize_cfun): Ditto.
23635 (copy_cfg_body): Ditto.
23636 (expand_call_inline): Ditto.
23637 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
23638 (estimate_node_size_and_time): Ditto.
23639 (inline_merge_summary): Ditto.
23640 * cgraphclones.c (cgraph_clone_edge): Ditto.
23641 (cgraph_clone_node): Ditto.
23642 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23643 (compute_trg_info): Ditto.
23645 2013-05-02 Ian Bolton <ian.bolton@arm.com>
23647 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
23648 S reg when fp attribute set.
23649 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
23651 2013-05-02 Ian Bolton <ian.bolton@arm.com>
23653 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
23655 (*and_one_cmplsi3_compare0_uxtw): Likewise.
23656 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
23657 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
23659 2013-05-02 Richard Biener <rguenther@suse.de>
23661 * tree-scalar-evolution.c (scev_info_hasher): Remove.
23662 (struct instantiate_cache_entry): New type.
23663 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
23664 (struct instantiate_cache_type): New type.
23665 (set_instantiated_value, get_instantiated_value): Remove.
23666 (get_instantiated_value_entry): New function.
23667 (instantiate_scev_name): Use the new cache and adjust.
23668 (instantiate_scev_poly): Adjust.
23669 (instantiate_scev_binary): Likewise.
23670 (instantiate_array_ref): Likewise.
23671 (instantiate_scev_convert): Likewise.
23672 (instantiate_scev_not): Likewise.
23673 (instantiate_scev_3): Likewise.
23674 (instantiate_scev_2): Likewise.
23675 (instantiate_scev_r): Likewise.
23676 (instantiate_scev): Likewise.
23677 (resolve_mixers): Likewise.
23679 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
23682 * lra-constraints.c (best_small_class_operands_num): Remove.
23683 (process_alt_operands): Remove small_class_operands_num. Take
23684 small classes operands into losers and only if the operand is not
23685 matched. Modify debugging output.
23686 (curr_insn_transform): Remove best_small_class_operands_num.
23689 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23691 * config/aarch64/aarch64-builtins.c
23692 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
23693 * config/aarch64/aarch64-simd-builtins.def
23694 (reduc_splus_): Add new modes.
23695 (reduc_uplus_): New.
23696 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
23697 (reduc_uplus_v4sf): Likewise.
23698 (reduc_splus_v4sf): Likewise.
23699 (aarch64_addv<mode>): Likewise.
23700 (reduc_uplus_<mode>): Likewise.
23701 (reduc_splus_<mode>): Likewise.
23702 (aarch64_addvv2di): Likewise.
23703 (reduc_uplus_v2di): Likewise.
23704 (reduc_splus_v2di): Likewise.
23705 (aarch64_addvv2si): Likewise.
23706 (reduc_uplus_v2si): Likewise.
23707 (reduc_splus_v2si): Likewise.
23708 (reduc_<sur>plus_<mode>): New.
23709 (reduc_<sur>plus_v2di): Likewise.
23710 (reduc_<sur>plus_v2si): Likewise.
23711 (reduc_<sur>plus_v4sf): Likewise.
23712 (aarch64_addpv4sf): Likewise.
23713 * config/aarch64/arm_neon.h
23714 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
23715 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
23716 add UNSPEC_SADDV, UNSPEC_UADDV.
23718 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
23720 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23722 * config/aarch64/arm_neon.h
23723 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
23725 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23727 * config/aarch64/aarch64-builtins
23728 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
23730 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23732 * config/aarch64/aarch64-simd-builtins.def
23733 (reduc_smax_): New.
23734 (reduc_smin_): Likewise.
23735 (reduc_umax_): Likewise.
23736 (reduc_umin_): Likewise.
23737 (reduc_smax_nan_): Likewise.
23738 (reduc_smin_nan_): Likewise.
23741 (smax): Update for V2SF, V4SF and V2DF modes.
23744 (smin_nan): Likewise.
23745 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
23746 (<su><maxmin><mode>3): ...This, refactor.
23747 (s<maxmin><mode>3): New.
23748 (<maxmin_uns><mode>3): Likewise.
23749 (reduc_<maxmin_uns>_<mode>): Refactor.
23750 (reduc_<maxmin_uns>_v4sf): Likewise.
23751 (reduc_<maxmin_uns>_v2si): Likewise.
23752 (aarch64_<fmaxmin><mode>: Remove.
23753 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
23755 (vmin<q>_f<32,64>): Likewise.
23756 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23758 (su): Add mappings for smax, smin, umax, umin.
23760 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23761 (FMAXMIN): Rename as...
23762 (FMAXMIN_UNS): ...This.
23764 (fmaxminv): Likewise.
23765 (fmaxmin): Likewise.
23767 (maxmin_uns_op): Likewise.
23769 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23771 * config/aarch64/arm_neon.h
23772 (vac<ge, gt><sd>_f<32, 64>): Rename to...
23773 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
23774 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
23776 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23778 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
23779 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
23781 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23783 * config/aarch64/aarch64-simd.md
23784 (vcond<mode>_internal): Handle special cases for constant masks.
23785 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
23786 (vcondu<mode><mode>): Likewise.
23787 (vcond<v_cmp_result><mode>): New.
23789 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23791 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
23792 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
23793 * config/aarch64/aarch64-simd-builtins.def
23794 (cmeq): Update to BUILTIN_VALLDI.
23799 * config/aarch64/arm_neon.h
23800 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
23801 to builtins or C as appropriate.
23803 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23805 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
23807 (cmhi): Rename to...
23809 * config/aarch64/aarch64-simd.md
23810 (simd_mode): Add SF.
23811 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
23812 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
23813 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
23814 (cstore<mode>_neg): ...This.
23815 * config/aarch64/iterators.md
23817 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
23818 (COMPARISONS): New.
23819 (UCOMPARISONS): Likewise.
23820 (optab): Add missing comparisons.
23826 (VCMP_S): Likewise.
23827 (VCMP_U): Likewise.
23828 (V_cmp_result): Add DF, SF modes.
23829 (v_cmp_result): Likewise.
23831 (vmtype): Likewise.
23832 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
23834 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
23836 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
23837 define_insn to define_insn_and_split.
23838 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
23839 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
23840 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
23841 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
23842 (thumb2_negscc): Likewise.
23844 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23846 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
23848 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23850 * config/arm/thumb2.md: Remove trailing whitespaces.
23852 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
23854 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
23855 Use gen_int_mode rather than GEN_INT.
23857 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
23859 * value-prof.c (stream_in_histogram_value): Remove the strayed
23862 2013-04-30 Richard Biener <rguenther@suse.de>
23864 PR middle-end/57122
23865 * cfghooks.c (split_edge): Properly check for the loop latch edge.
23867 2013-04-30 Richard Biener <rguenther@suse.de>
23869 PR middle-end/57107
23870 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
23872 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
23874 PR rtl-optimization/56957
23875 PR rtl-optimization/57105
23876 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
23877 variable. Use just INSN_UID for determining whether an insn
23878 should be only disconnected from the insn stream.
23879 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
23881 2013-04-30 Jakub Jelinek <jakub@redhat.com>
23883 PR tree-optimization/57104
23884 * tsan.c (instrument_expr): Don't instrument accesses to
23885 DECL_HARD_REGISTER VAR_DECLs.
23887 2013-04-30 Richard Biener <rguenther@suse.de>
23889 * function.h (loops_for_fn): New inline function.
23890 (set_loops_for_fn): Likewise.
23891 * cfgloop.h (place_new_loop): Add struct function parameter.
23892 (get_loop): Likewise.
23893 (get_loops): Likewise.
23894 (number_of_loops): Likewise.
23895 (fel_next): Adjust.
23896 (fel_init): Likewise.
23897 * cfg.c (get_loop_copy): Adjust.
23898 * cfgloop.c (flow_loops_dump): Likewise.
23899 (record_loop_exits): Likewise.
23900 (verify_loop_structure): Likewise.
23901 * cfgloopanal.c (mark_irreducible_loops): Likewise.
23902 (estimate_reg_pressure_cost): Likewise.
23903 (mark_loop_exit_edges): Likewise.
23904 * cfgloopmanip.c (place_new_loop): Likewise.
23905 (add_loop): Likewise.
23906 (duplicate_loop): Likewise.
23907 * graph.c (draw_cfg_nodes): Likewise.
23908 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
23909 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
23910 (extract_affine_chrec): Likewise.
23911 (build_scop_iteration_domain): Likewise.
23912 * graphite.c (graphite_initialize): Likewise.
23913 * ira-build.c (create_loop_tree_nodes): Likewise.
23914 (more_one_region_p): Likewise.
23915 (rebuild_regno_allocno_maps): Likewise.
23916 (mark_loops_for_removal): Likewise.
23917 (mark_all_loops_for_removal): Likewise.
23918 (remove_unnecessary_regions): Likewise.
23919 (ira_build): Likewise.
23920 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
23921 * loop-init.c (fix_loop_structure): Likewise.
23922 (gate_rtl_move_loop_invariants): Likewise.
23923 (gate_rtl_unswitch): Likewise.
23924 (gate_rtl_unroll_and_peel_loops): Likewise.
23925 (rtl_doloop): Likewise.
23926 * lto-streamer-in.c (input_cfg): Likewise.
23927 * lto-streamer-out.c (output_cfg): Likewise.
23928 * modulo-sched.c (sms_schedule): Likewise.
23929 * predict.c (tree_estimate_probability): Likewise.
23930 (tree_estimate_probability_driver): Likewise.
23931 (estimate_loops): Likewise.
23932 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
23933 (move_sese_region_to_fn): Likewise.
23934 (debug_loop_num): Likewise.
23935 * tree-chrec.c (chrec_evaluate): Likewise.
23936 (hide_evolution_in_other_loops_than_loop): Likewise.
23937 (chrec_component_in_loop_num): Likewise.
23938 (reset_evolution_in_loop): Likewise.
23939 (evolution_function_is_invariant_rec_p): Likewise.
23940 * tree-if-conv.c (main_tree_if_conversion): Likewise.
23941 * tree-inline.c (copy_loops): Likewise.
23942 (copy_cfg_body): Likewise.
23943 (tree_function_versioning): Likewise.
23944 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
23945 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
23947 (add_to_evolution_1): Likewise.
23948 (scev_const_prop): Likewise.
23949 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
23950 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
23951 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
23952 (tree_ssa_lim_initialize): Likewise.
23953 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
23954 (verify_loop_closed_ssa): Likewise.
23955 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
23956 (tree_ssa_loop_im): Likewise.
23957 (tree_ssa_loop_unswitch): Likewise.
23958 (tree_vectorize): Likewise.
23959 (check_data_deps): Likewise.
23960 (tree_ssa_loop_ivcanon): Likewise.
23961 (tree_ssa_loop_bounds): Likewise.
23962 (tree_complete_unroll): Likewise.
23963 (tree_complete_unroll_inner): Likewise.
23964 (tree_parallelize_loops): Likewise.
23965 (tree_ssa_loop_prefetch): Likewise.
23966 (tree_ssa_loop_ivopts): Likewise.
23967 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
23968 * tree-vectorizer.c (vectorize_loops): Likewise.
23970 2013-04-29 Mike Frysinger <vapier@gentoo.org>
23972 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
23973 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
23974 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
23975 with EABI_LINK_SPEC.
23977 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23980 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
23983 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
23986 * lra-constraints.c (process_alt_operands): Discourage a bit more
23987 using memory for pseudos. Print cost dump for alternatives.
23988 Modify cost values for conflicts with early clobbers.
23989 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
23991 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23994 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
23996 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23998 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
23999 from/to S register.
24000 (movdi_aarch64): Support LDR/STR from/to D register.
24002 2013-04-29 Ian Bolton <ian.bolton@arm.com>
24004 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
24005 or higher by default.
24007 2013-04-29 Richard Biener <rguenther@suse.de>
24009 PR middle-end/57075
24010 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
24011 even if not adding abnormal edges for calls that can make
24014 2013-04-29 Richard Biener <rguenther@suse.de>
24016 PR middle-end/57103
24017 * tree-cfg.c (move_stmt_op): Fix condition under which to update
24019 (move_stmt_r): Remove redundant checking.
24021 2013-04-29 Teresa Johnson <tejohnson@google.com>
24024 * basic-block.h (apply_scale): New function.
24025 (apply_probability): Use apply_scale.
24026 * gimple-streamer-in.c (input_bb): Ditto.
24027 * lto-streamer-in.c (input_cfg): Ditto.
24028 * lto-cgraph.c (merge_profile_summaries): Ditto.
24029 * tree-optimize.c (execute_fixup_cfg): Ditto.
24030 * tree-inline.c (copy_bb): Update comment to use apply_scale.
24031 (copy_edges_for_bb): Ditto.
24032 (copy_cfg_body): Ditto.
24034 2013-04-29 Tom de Vries <tom@codesourcery.com>
24036 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
24037 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
24038 (tail_merge_optimize): Handle current_loops == NULL.
24040 2013-04-26 Jeff Law <law@redhat.com>
24042 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
24043 (simplify_cond_using_ranges): Generalize code to simplify
24044 COND_EXPRs where one argument is a constant and the other
24045 is an SSA_NAME created by an integral type conversion.
24047 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24049 * config/arm/arm.md (store_minmaxsi): Use only when
24050 optimize_insn_for_size_p.
24052 2013-04-29 Christian Bruel <christian.bruel@st.com>
24055 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
24057 2013-04-29 Richard Biener <rguenther@suse.de>
24059 PR middle-end/57089
24060 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
24061 loop tree make sure to schedule a fixup for the child as well.
24062 (expand_omp_for_generic): Properly add loops.
24063 (expand_omp_for_static_nochunk): Likewise.
24064 (expand_omp_for_static_chunk): Likewise.
24065 (expand_omp_for): For the degenerate case fixup loops.
24066 (expand_omp_sections): Fix default bb placement in loops.
24067 (expand_omp_atomic_pipeline): Properly add loops.
24069 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24071 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
24073 2013-04-29 Tom de Vries <tom@codesourcery.com>
24075 * tree-ssa-tail-merge.c: Update header comment.
24077 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24079 * config/aarch64/arm_neon.h
24080 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
24081 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
24082 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
24083 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
24084 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
24085 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
24086 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
24087 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
24089 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24091 * config/aarch64/aarch64-simd.md
24092 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
24093 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
24094 fix_trunc, fixuns_trunc.
24095 (ftrunc<VDQF:mode>2): New.
24096 * config/aarch64/iterators.md (optab): Add fix, fixuns.
24097 (fix_trunc_optab): New.
24099 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24101 * config/aarch64/aarch64-builtins.c
24102 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
24103 iceilf, lround, iroundf.
24105 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
24108 * config/i386/i386.h (enum ix86_tune_indices)
24109 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
24110 New, split from X86_TUNE_INTER_UNIT_MOVES.
24111 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
24112 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
24113 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
24114 (TARGET_INTER_UNIT_MOVES): Remove.
24115 * config/i386/i386.c (initial_ix86_tune_features): Update.
24116 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
24117 (ix86_expand_convert_uns_didf_sse): Use
24118 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
24119 (ix86_expand_vector_init_one_nonzero): Ditto.
24120 (ix86_expand_vector_init_interleave): Ditto.
24121 (inline_secondary_memory_needed): Return true for moves from SSE class
24122 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
24123 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
24124 * config/i386/constraints.md (Yi, Ym): Depend on
24125 TARGET_INTER_UNIT_MOVES_TO_VEC.
24126 (Yj, Yn): New constraints.
24127 * config/i386/i386.md (*movdi_internal): Change constraints of
24128 operand 1 from Yi to Yj and from Ym to Yn.
24129 (*movsi_internal): Ditto.
24130 (*movdf_internal): Ditto.
24131 (*movsf_internal): Ditto.
24132 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
24133 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
24134 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
24135 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
24136 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
24137 * config/i386/sse.md (movdi_to_sse): Ditto.
24138 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
24139 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
24140 TARGET_INTER_UNIT_MOVES.
24141 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
24142 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
24143 instead of TARGET_INTER_UNIT_MOVES.
24144 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
24145 operand 1 from Yi to Yj and from Ym to Yn.
24147 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24149 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
24150 (float_truncate_hi_): Likewise.
24151 (float_extend_lo_): Likewise.
24152 (float_truncate_lo_): Likewise.
24153 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
24154 (aarch64_float_extend_lo_v2df): Likewise.
24155 (vec_unpacks_hi_v4sf): Likewise.
24156 (aarch64_float_truncate_lo_v2sf): Likewise.
24157 (aarch64_float_truncate_hi_v4sf): Likewise.
24158 (vec_pack_trunc_v2df): Likewise.
24159 (vec_pack_trunc_df): Likewise.
24161 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24163 * config/aarch64/aarch64-builtins.c
24164 (aarch64_fold_builtin): Fold float conversions.
24165 * config/aarch64/aarch64-simd-builtins.def
24166 (floatv2si, floatv4si, floatv2di): New.
24167 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
24168 * config/aarch64/aarch64-simd.md
24169 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
24170 * config/aarch64/iterators.md (FLOATUORS): New.
24171 (optab): Add float, floatuns.
24172 (su_optab): Likewise.
24174 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24176 * config/aarch64/aarch64-builtins.c
24177 (aarch64_builtin_vectorized_function): Use new names for
24179 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
24180 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
24181 (fcvtzu): Split as...
24182 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
24183 (fcvtas): Split as...
24184 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
24185 (fcvtau): Split as...
24186 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
24187 (fcvtps): Split as...
24188 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
24189 (fcvtpu): Split as...
24190 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
24191 (fcvtms): Split as...
24192 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
24193 (fcvtmu): Split as...
24194 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
24195 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
24196 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
24197 (lfrintnusf, lfrintnudf): Likewise.
24198 * config/aarch64/aarch64-simd.md
24199 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
24201 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
24202 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
24203 (fcvt_pattern): Likewise.
24205 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24207 * config/aarch64/aarch64-simd.md
24208 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
24209 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
24211 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24213 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
24214 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
24215 (vrnd<a,m,n,p>_f32): Implement using builtins.
24216 (vrnd<i,x><q>_f<32, 64>): New.
24218 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
24220 * config/aarch64/aarch64-builtins.c
24221 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
24222 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
24223 (frintz): Rename to...
24225 (frintp): Rename to...
24227 (frintm): Rename to...
24229 (frinti): Rename to...
24230 (nearbyint): ...this.
24231 (frintx): Rename to...
24233 (frinta): Rename to...
24235 * config/aarch64/aarch64-simd.md
24236 (aarch64_frint<frint_suffix><mode>): Delete.
24237 (<frint_pattern><mode>2): Convert to insn.
24238 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
24239 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
24240 (frint_pattern): Likewise.
24241 (frint_suffix): Likewise.
24243 2013-04-29 Richard Biener <rguenther@suse.de>
24245 PR tree-optimization/57081
24246 * loop-init.c: Include tree-flow.h.
24247 (loop_optimizer_finalize): Free number of iteration estimates.
24248 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
24250 2013-04-29 Jakub Jelinek <jakub@redhat.com>
24252 PR tree-optimization/57083
24253 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
24254 non-singleton shift count range, zero extend low_bound for uns case.
24256 * config/i386/predicates.md (general_vector_operand): New predicate.
24257 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
24258 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
24259 if they aren't nonimmediate operands. If their original values
24260 satisfy const_vector_equal_evenodd_p, don't shift them.
24261 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
24262 predicates. For the SSE4.1 case force operands[{1,2}] into registers
24263 if not nonimmediate_operand.
24264 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
24265 instead of register_operand.
24266 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
24268 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
24270 * stor-layout.c (finalize_size_functions): Allocate a structure and
24271 reset cfun before dumping the functions.
24273 2013-04-27 Jakub Jelinek <jakub@redhat.com>
24275 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
24278 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
24279 use xop_pmacsdqh if uns_p.
24280 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
24281 the immediate rotate count.
24283 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
24285 * rtl.h (struct rtx_def): Add comment for field jump.
24286 (LRA_SUBREG_P): New macro.
24287 * recog.c (register_operand): Check LRA_SUBREG_P.
24288 * lra.c (lra): Add note at the end of RTL code. Align non-empty
24290 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
24291 (lra_final_code_change): Skip subreg change for operators.
24292 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
24293 if there are no operand changes.
24294 * lra-constraints.c (curr_insn_set): New.
24295 (match_reload): Set LRA_SUBREG_P.
24296 (emit_spill_move): Ditto.
24297 (check_and_process_move): Use curr_insn_set. Process only single
24298 set insns. Don't initialize sec_mem_p and change_p.
24299 (simplify_operand_subreg): Use LRA_SUBREG_P.
24300 (reg_in_class_p): New function.
24301 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
24302 of #ifdef. Add code to remove cycling.
24303 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
24304 non-null disp. Reload inner instead of disp when base and index
24305 are null. Try to put lo_sum into register.
24306 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
24307 (check_and_process_move): Move code for move cost check to
24308 simple_move_p. Remove equiv_substitution.
24309 (simple_move_p): New function.
24310 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
24311 curr_insn_set. Call check_and_process_move only for single set
24312 insns. Use the new function. Move call of check_and_process_move
24313 after operand equiv substitution and address process.
24315 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24318 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
24319 with nonlocal goto receivers or returns twice calls, ignore
24320 unininitialized values from abnormal edges to nl goto receiver
24321 or returns twice call.
24323 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24325 PR tree-optimization/57051
24326 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
24327 and VEC_RSHIFT_EXPR if shift count is a multiple of element
24330 2013-04-26 Richard Biener <rguenther@suse.de>
24332 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
24333 (expand_omp_taskreg): Likewise. Mark loops for fixup.
24334 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
24335 (fixup_loop_arrays_after_move): New function.
24336 (move_sese_region_to_fn): Properly outline the loop tree parts
24337 of the SESE region.
24339 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
24341 * config/i386/i386.md (type, unit): Fix long lines.
24343 2013-04-26 Richard Biener <rguenther@suse.de>
24345 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
24346 (lto-streamer-out.o): Likewise.
24347 * cfgloop.c (init_loops_structure): Export, add struct function
24348 argument and adjust.
24349 (flow_loops_find): Adjust.
24350 * cfgloop.h (enum loop_estimation): Add EST_LAST.
24351 (init_loops_structure): Declare.
24352 * lto-streamer-in.c: Include cfgloop.h.
24353 (input_cfg): Input the loop tree.
24354 * lto-streamer-out.c: Include cfgloop.h.
24355 (output_cfg): Output the loop tree.
24356 (output_struct_function_base): Do not drop PROP_loops.
24358 2013-03-26 Richard Biener <rguenther@suse.de>
24360 * tree-cfg.c (execute_build_cfg): Build the loop tree.
24361 (pass_build_cfg): Provide PROP_loops.
24362 (move_sese_region_to_fn): Remove loops that are outlined into fn
24364 * tree-inline.c: Include cfgloop.h.
24365 (initialize_cfun): Do not drop PROP_loops.
24366 (copy_loops): New function.
24367 (copy_cfg_body): Copy loop structure.
24368 (tree_function_versioning): Initialize destination loop tree.
24369 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
24370 (pass_parallelize_loops): Do IL verification.
24371 * loop-init.c (loop_optimizer_init): Fixup loops if required.
24372 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
24373 the CFG make sure we fixup loops as well.
24374 * tree-ssa-tail-merge.c: Include cfgloop.h.
24375 (replace_block_by): When merging loop latches mark loops for fixup.
24376 * lto-streamer-out.c (output_struct_function_base): Drop
24377 PROP_loops for now.
24378 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
24379 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
24380 * ipa-split.c: Include cfgloop.h.
24381 (split_function): Add the new return block to the loop tree root.
24382 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
24383 whether we have removed the forwarder block.
24384 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
24385 * cfgloop.h (place_new_loop): Declare.
24386 * cfgloopmanip.c (place_new_loop): Export.
24387 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
24388 (tree-switch-conversion.o): Likewise.
24389 (tree-complex.o): Likewise.
24390 (tree-inline.o): Likewise.
24391 (tree-ssa-tailmerge.o): Likewise.
24392 (ipa-split.o): Likewise.
24393 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
24394 (tree-ssa-copy.o): Likewise.
24395 * tree-switch-conversion.c: Include cfgloop.h
24396 (process_switch): If we emit a bit-test cascade, schedule loops
24398 * tree-complex.c: Include cfgloop.h.
24399 (expand_complex_div_wide): Properly add new basic-blocks to loops.
24400 * asan.c: Include cfgloop.h.
24401 (create_cond_insert_point): Properly add new basic-blocks to
24402 loops, schedule loop fixup.
24403 * cfgloop.c (verify_loop_structure): Check that looks are not
24405 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
24407 (expand_omp_for_generic): Likewise.
24408 (expand_omp_sections): Likewise.
24409 (expand_omp_atomic_pipeline): Schedule loops for fixup.
24410 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
24411 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
24412 is initialized, not when loops are present.
24413 * tree-parloops.c (parallelize_loops): Remove checking here.
24414 * passes.c (init_optimization_passes): Schedule a copy-propagation
24415 pass before complete unrolling of inner loops.
24417 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24419 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
24420 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
24421 (colorize_init): Add argument to _WIN32 version.
24422 * toplev.c: Include diagnostic-color.h.
24423 (process_options): Default to -fdiagnostics-color=auto if
24424 GCC_COLORS env var is in the environment.
24425 * common.opt (fdiagnostics-color=): Add Var and Init.
24426 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
24427 env var is in the environment, the default is auto rather than never.
24429 * diagnostic.h (file_name_as_prefix): Add context argument.
24430 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
24431 the string as locus.
24432 * langhooks.c (lhd_print_error_function): Adjust caller.
24434 2013-04-25 Lawrence Crowl <crowl@google.com>
24436 * var-tracking.c (shared_hash_def::htab):
24437 Change type to hash_table. Update dependent calls and types.
24439 2013-04-25 Lawrence Crowl <crowl@google.com>
24441 * Makefile.in: Update as needed below.
24443 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
24444 Move declaration to after the type's method definitons.
24446 * attribs.c (htab_t scoped_attributes::attribute_hash):
24447 Change type to hash_table. Update dependent calls and types.
24449 * bitmap.c (htab_t bitmap_desc_hash):
24450 Change type to hash_table. Update dependent calls and types.
24452 * cselib.c (htab_t cselib_hash_table):
24453 Change type to hash_table. Update dependent calls and types.
24455 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
24456 (hash_string_slot_node): Move implementation into lto-streamer.h
24457 struct string_slot_hasher.
24458 (eq_string_slot_node): Likewise.
24460 * data-streamer-out.c: Update output_block::string_hash_table
24461 dependent calls and types.
24463 * dwarf2cfi.c (htab_t trace_index):
24464 Change type to hash_table. Update dependent calls and types.
24466 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
24467 Change type to hash_table. Update dependent calls and types.
24468 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
24469 (htab_t optimize_external_refs::map): Likewise.
24470 (htab_t output_comp_unit::extern_map): Likewise.
24471 (htab_t output_comdat_type_unit::extern_map): Likewise.
24472 (htab_t output_macinfo::macinfo_htab): Likewise.
24473 (htab_t optimize_location_lists::htab): Likewise.
24474 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
24476 * except.c (htab_t ehspec_hash_type):
24477 Change type to hash_table. Update dependent calls and types.
24478 (assign_filter_values::ttypes): Likewise.
24479 (assign_filter_values::ehspec): Likewise.
24480 (sjlj_assign_call_site_values::ar_hash): Likewise.
24481 (convert_to_eh_region_ranges::ar_hash): Likewise.
24483 * gcse.c (htab_t pre_ldst_table):
24484 Change type to hash_table. Update dependent calls and types.
24486 * ggc-common.c (htab_t saving_htab):
24487 Change type to hash_table. Update dependent calls and types.
24488 (htab_t loc_hash): Likewise.
24489 (htab_t ptr_hash): Likewise.
24490 (call_count): Rename ggc_call_count.
24491 (call_alloc): Rename ggc_call_alloc.
24492 (loc_descriptor): Rename make_loc_descriptor.
24493 (add_statistics): Rename ggc_add_statistics.
24495 * ggc-common.c (saving_htab):
24496 Change type to hash_table. Update dependent calls and types.
24498 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
24499 (push_gimplify_context): Likewise.
24500 (pop_gimplify_context): Likewise.
24501 (struct gimple_temp_hash_elt): Added.
24502 (struct gimplify_hasher): Likewise.
24503 (struct gimplify_ctx.temp_htab):
24504 Change type to hash_table. Update dependent calls and types.
24506 * gimple-fold.c: Include gimplify-ctx.h.
24508 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
24509 Change type to hash_table. Update dependent calls and types.
24510 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
24511 avoid potential global name collision.
24513 * gimplify.c: Include gimplify-ctx.h.
24514 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
24515 (htab_t gimplify_ctx::temp_htab):
24516 Update dependent calls and types for new type hash_table.
24517 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
24518 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
24520 * gimplify-ctx.h: New.
24521 (struct gimple_temp_hash_elt): Move from gimplify.c.
24522 (class gimplify_hasher): New.
24523 (struct gimplify_ctx): Move from gimple.h.
24524 (htab_t gimplify_ctx::temp_htab):
24525 Change type to hash_table. Update dependent calls and types.
24527 * graphite-clast-to-gimple.c: Include graphite-htab.h.
24528 (htab_t ivs_params::newivs_index):
24529 Change type to hash_table. Update dependent calls and types.
24530 (htab_t ivs_params::params_index): Likewise.
24531 (htab_t print_generated_program::params_index): Likewise.
24532 (htab_t gloog::newivs_index): Likewise.
24533 (htab_t gloog::params_index): Likewise.
24535 * graphite.c: Include graphite-htab.h.
24536 4htab_t graphite_transform_loops::bb_pbb_mapping):
24537 Change type to hash_table. Update dependent calls and types.
24539 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
24540 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
24541 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
24543 * graphite-dependences.c: Include graphite-htab.h.
24544 (loop_is_parallel_p): Change hash table type of parameter.
24546 * graphite-htab.h: New.
24547 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
24548 (extern find_pbb_via_hash): Move from graphite-poly.h.
24549 (extern loop_is_parallel_p): Move from graphite-poly.h.
24550 (extern get_loop_body_pbbs): Move from graphite-poly.h.
24552 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
24553 (extern loop_is_parallel_p): Move to graphite-htab.h.
24554 (extern get_loop_body_pbbs): Move to graphite-htab.h.
24556 * haifa-sched.c (htab_t delay_htab):
24557 Change type to hash_table. Update dependent calls and types.
24558 (htab_t delay_htab_i2): Likewise.
24560 * ira-color.c (htab_t allocno_hard_regs_htab):
24561 Change type to hash_table. Update dependent calls and types.
24563 * ira-costs.c (htab_t cost_classes_htab):
24564 Change type to hash_table. Update dependent calls and types.
24566 * loop-invariant.c (htab_t merge_identical_invariants::eq):
24567 Change type to hash_table. Update dependent calls and types.
24569 * loop-iv.c (htab_t bivs):
24570 Change type to hash_table. Update dependent calls and types.
24572 * loop-unroll.c (htab_t opt_info::insns_to_split):
24573 Change type to hash_table. Update dependent calls and types.
24574 (htab_t opt_info::insns_with_var_to_expand): Likewise.
24576 * lto-streamer.h (struct string_slot): Move from data-streamer.h
24577 (struct string_slot_hasher): New.
24578 (htab_t output_block::string_hash_table):
24579 Change type to hash_table. Update dependent calls and types.
24581 * lto-streamer-in.c (freeing_string_slot_hasher): New.
24582 (htab_t file_name_hash_table):
24583 Change type to hash_table. Update dependent calls and types.
24585 * lto-streamer-out.c: Update output_block::string_hash_table dependent
24588 * lto-streamer.c (htab_t tree_htab):
24589 Change type to hash_table. Update dependent calls and types.
24591 * omp-low.c: Include gimplify-ctx.h.
24593 * passes.c (htab_t name_to_pass_map):
24594 Change type to hash_table. Update dependent calls and types.
24595 (pass_traverse): Rename to passes_pass_traverse.
24597 * plugin.c (htab_t event_tab):
24598 Change type to hash_table. Update dependent calls and types.
24600 * postreload-gcse.c (htab_t expr_table):
24601 Change type to hash_table. Update dependent calls and types.
24602 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
24604 * sese.c (debug_rename_map_1): Make extern.
24605 (htab_t copy_bb_and_scalar_dependences::rename_map):
24606 Change type to hash_table. Update dependent calls and types.
24608 * sese.h (extern debug_rename_map): Move to .c file.
24610 * store-motion.c (htab_t store_motion_mems_table):
24611 Change type to hash_table. Update dependent calls and types.
24613 * trans-mem.c (htab_t tm_new_mem_hash):
24614 Change type to hash_table. Update dependent calls and types.
24616 * tree-browser.c (htab_t TB_up_ht):
24617 Change type to hash_table. Update dependent calls and types.
24619 * tree-cfg.c (htab_t discriminator_per_locus):
24620 Change type to hash_table. Update dependent calls and types.
24622 * tree-complex.c: Include tree-hasher.h
24623 (htab_t complex_variable_components):
24624 Change type to hash_table. Update dependent calls and types.
24626 * tree-eh.c (htab_t finally_tree):
24627 Change type to hash_table. Update dependent calls and types.
24629 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
24630 struct int_tree_hasher.
24631 (extern int_tree_map_eq): Likewise.
24632 (uid_decl_map_hash): Removed.
24633 (extern decl_tree_map_eq): Likewise.
24635 * tree-hasher.h: New.
24636 (struct int_tree_hasher): New.
24637 (typedef int_tree_htab_type): New.
24639 * tree-inline.c: Include gimplify-ctx.h.
24641 * tree-mudflap.c: Include gimplify-ctx.h.
24643 * tree-parloops.c: Include tree-hasher.h.
24644 (htab_t eliminate_local_variables_stmt::decl_address):
24645 Change type to hash_table. Update dependent calls and types.
24646 (htab_t separate_decls_in_region::decl_copies): Likewise.
24648 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
24649 Change type to hash_table. Update dependent calls and types.
24651 * tree-sra.c (candidates):
24652 Change type to hash_table. Update dependent calls and types.
24654 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
24656 (int_tree_map_hash): Likewise.
24658 * tree-ssa-dom.c (htab_t avail_exprs):
24659 Change type to hash_table. Update dependent calls and types.
24661 * tree-ssa-live.c (var_map_base_init::tree_to_index):
24662 Change type to hash_table. Update dependent calls and types.
24664 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
24665 Change type to hash_table. Update dependent calls and types.
24667 * tree-ssa-phiopt.c (seen_ssa_names):
24668 Change type to hash_table. Update dependent calls and types.
24670 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
24671 Change type to hash_table. Update dependent calls and types.
24673 * tree-ssa-uncprop.c (equiv):
24674 Change type to hash_table. Update dependent calls and types.
24676 2013-04-25 Jakub Jelinek <jakub@redhat.com>
24678 PR rtl-optimization/57003
24679 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
24680 call note_stores with kill_clobbered_value callback again after
24681 killing regs_invalidated_by_call.
24683 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24685 * config/aarch64/aarch64-simd.md
24686 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
24687 (aarch64_simd_bsl<mode>): Likewise.
24688 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
24690 2013-04-25 Marek Polacek <polacek@redhat.com>
24692 PR tree-optimization/57066
24693 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
24695 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24697 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
24699 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24701 * config/aarch64/aarch64-builtins.c
24702 (aarch64_fold_builtin): New.
24703 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
24704 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
24705 * config/aarch64/aarch64-simd-builtins.def (abs): New.
24706 * config/aarch64/arm_neon.h
24707 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
24709 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24710 Tejas Belagod <tejas.belagod@arm.com>
24712 * config/aarch64/aarch64-builtins.c
24713 (aarch64_gimple_fold_builtin): New.
24714 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
24715 * config/aarch64/aarch64-simd-builtins.def (addv): New.
24716 * config/aarch64/aarch64-simd.md (addpv4sf): New.
24717 (addvv4sf): Update.
24718 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
24720 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24722 * config/aarch64/aarch64.md
24723 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
24725 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24727 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
24728 (*ngcsi_uxtw): New pattern.
24730 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24731 Julian Brown <julian@codesourcery.com>
24733 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
24734 (TB_DREG): Add T_V4HF.
24735 (v4hf_UP): New macro.
24736 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24737 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24738 Handle initialisation of V4HF. Adjust initialisation of reinterpret
24740 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24741 (arm_vector_mode_supported_p): Handle V4HF.
24742 (arm_mangle_map): Handle V4HFmode.
24743 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
24744 * config/arm/arm_neon_builtins.def: Add entries for
24745 vcvtv4hfv4sf, vcvtv4sfv4hf.
24746 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
24747 (neon_vcvtv4hfv4sf): Likewise.
24748 * config/arm/neon-gen.ml: Handle half-precision floating point
24750 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
24751 * config/arm/arm_neon.h: Regenerate.
24752 * config/arm/neon.ml (type elts): Add F16.
24753 (type vectype): Add T_float16x4, T_floatHF.
24754 (type vecmode): Add V4HF.
24755 (type features): Add Requires_FP_bit feature.
24756 (elt_width): Handle F16.
24757 (elt_class): Likewise.
24758 (elt_of_class_width): Likewise.
24759 (mode_of_elt): Refactor.
24760 (type_for_elt): Handle F16, fix error messages.
24761 (vectype_size): Handle T_float16x4.
24762 (vcvt_sh): New function.
24763 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
24764 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
24765 (string_of_mode): Handle V4HF.
24766 * doc/arm-neon-intrinsics.texi: Regenerate.
24768 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24770 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
24771 format specifier in 'X' case.
24773 2013-04-25 Alan Modra <amodra@gmail.com>
24776 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
24777 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
24778 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
24779 Repeat for many other rotate/shift and mask patterns using subregs.
24780 Name lshiftrt insns.
24781 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
24782 on WORDS_BIG_ENDIAN.
24784 2013-04-25 Alan Modra <amodra@gmail.com>
24786 * config.gcc: Support little-endian powerpc-linux targets.
24787 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
24788 (LINK_OS_LINUX_SPEC): Define.
24789 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
24790 Preserve MASK_LITTLE_ENDIAN.
24791 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
24792 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
24793 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
24794 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
24795 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
24796 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
24797 Correct fp word order for little-endian. Don't shift toc entries
24798 smaller than a word for little-endian.
24799 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
24800 (bswapdi2 splits): Correct low-part subreg for little-endian.
24801 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
24802 low/high where such is correct only for be.
24803 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
24804 little-endian for -mcall-aixdesc.
24806 2013-04-25 Alan Modra <amodra@gmail.com>
24808 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
24809 replace_equiv_address_nv.
24811 2013-04-25 Alan Modra <amodra@gmail.com>
24813 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
24815 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24818 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24819 * rtl.h (struct rtx_def): ...
24821 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24823 PR rtl-optimizations/57046
24824 * lra-constraints (split_reg): Set up lra_risky_transformations_p
24825 for multi-reg splits.
24827 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
24829 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
24831 2013-04-24 Sterling Augustine <saugustine@google.com>
24833 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
24834 (comp_dir_string, debug_str_dwo_section): New.
24835 (DEBUG_STR_DWO_SECTION): Rename to ...
24836 (DEBUG_DWO_STR_SECTION): ... this.
24837 (DEBUG_NORM_STR_SECTION): Delete.
24838 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
24839 (DEBUG_STR_DWO_SECTION_FLAGS): New.
24840 (find_AT_string): Move most logic to ...
24841 (find_AT_string_in_table): ... here. New.
24842 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
24843 add_skeleton_AT_string. Delete logic.
24844 (output_skeleton_debug_sections): Remove call to
24845 add_top_level_skeleton_die_attrs.
24846 (add_comp_dir_attribute): Move logic to comp_dir_string.
24847 (dwarf2out_init): Initialize debug_str_dwo_section.
24848 (output_indirect_string): Call find_string_form.
24849 (output_indirect_strings): Rewrite.
24850 (prune_unused_types): Empty skeleton_debug_str_hash.
24851 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
24852 (dwarf2out_finish): Call output_indirect_strings.
24854 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24856 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
24858 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24860 * rtl.h (struct rtx_def): Add comment for field jump.
24861 (LRA_SUBREG_P): New macro.
24862 * recog.c (register_operand): Check LRA_SUBREG_P.
24863 * lra.c (lra): Add note at the end of RTL code. Align non-empty
24865 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
24866 (lra_final_code_change): Skip subreg change for operators.
24867 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
24868 if there are no operand changes.
24869 * lra-constraints.c (curr_insn_set): New.
24870 (match_reload): Set LRA_SUBREG_P.
24871 (emit_spill_move): Ditto.
24872 (check_and_process_move): Use curr_insn_set. Process only single
24873 set insns. Don't initialize sec_mem_p and change_p.
24874 (simplify_operand_subreg): Use LRA_SUBREG_P.
24875 (reg_in_class_p): New function.
24876 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
24877 of #ifdef. Add code to remove cycling.
24878 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
24879 non-null disp. Reload inner instead of disp when base and index
24880 are null. Try to put lo_sum into register.
24881 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
24882 (check_and_process_move): Move code for move cost check to
24883 simple_move_p. Remove equiv_substitution.
24884 (simple_move_p): New function.
24885 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
24886 curr_insn_set. Call check_and_process_move only for single set
24887 insns. Use the new function. Move call of check_and_process_move
24888 after operand equiv substitution and address process.
24890 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
24892 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
24893 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
24894 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
24896 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24898 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
24900 2013-04-24 Marek Polacek <polacek@redhat.com>
24902 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
24903 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
24904 (select_loops_exit_conditions): Likewise.
24905 (number_of_iterations_for_all_loops): Likewise.
24906 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
24907 (scev_analysis): Likewise.
24909 2013-04-02 Catherine Moore <clm@codesourcery.com>
24910 Chao-ying Fu <fu@mips.com>
24912 * config/mips/micromips.md (jraddiusp): New pattern.
24913 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
24914 instruction if possible.
24916 2013-04-24 Alan Modra <amodra@gmail.com>
24918 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
24920 2013-04-24 Julian Brown <julian@codesourcery.com>
24921 Chung-Lin Tang <cltang@codesourcery.com>
24923 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
24924 dependency behavior in enumeration type DIE generation. Add TODO note
24925 to comments about future DW_FORM_sdata/udata re-work of related code.
24927 2013-04-23 Lawrence Crowl <crowl@google.com>
24929 * Makefile.in: Update as needed below.
24931 * hash-table.h (class hash_table):
24932 Correct many methods with parameter types compare_type to the correct
24933 value_type. (Correct code was unlikely to notice the change.)
24934 (hash_table::elements_with_deleted) New.
24935 (class hashtable::iterator): New.
24936 (hashtable::begin()): New.
24937 (hashtable::end()): New.
24938 (FOR_EACH_HASH_TABLE_ELEMENT): New.
24940 * statistics.c (statistics_hashes):
24941 Change type to hash_table. Update dependent calls and types.
24943 * tree-into-ssa.c (var_infos):
24944 Change type to hash_table. Update dependent calls and types.
24946 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
24947 Change type to hash_table. Update dependent calls and types.
24949 * tree-ssa-loop-im.c (struct mem_ref.refs):
24950 Change type to hash_table. Update dependent calls and types.
24952 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
24953 Change type to hash_table. Update dependent calls and types.
24955 * tree-ssa-sccvn.c (vn_tables_s::nary):
24956 Change type to hash_table. Update dependent calls and types.
24957 (vn_tables_s::phis): Likewise.
24958 (vn_tables_s::references): Likewise.
24960 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
24961 (vn_reference_eq): Update parameter and return types.
24963 * tree-ssa-structalias.c (pointer_equiv_class_table):
24964 Change type to hash_table. Update dependent calls and types.
24965 (location_equiv_class_table): Likewise.
24967 * tree-vect-data-refs.c: Consequential changes for making
24968 peeling a hash_table.
24970 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
24971 (destroy_loop_vec_info): Dependent hash_table update.
24973 * tree-vectorizer.h (peeling_htab):
24974 Change type to hash_table. Update dependent calls and types.
24976 2013-04-23 Shiva Chen <shiva0217@gmail.com>
24978 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
24979 to check the register content is equal or not.
24980 * lra-constraints.c (match_reload): Use lra_assign_reg_val
24981 to assign register content record.
24982 * lra-eliminations.c (update_reg_eliminate): Use
24983 lra_update_reg_val_offset to update register content offset.
24984 * lra-int.h (struct lra_reg): Add offset member.
24985 (lra_reg_val_equal_p): New static inline function.
24986 (lra_update_reg_val_offset): New static inline function.
24987 (lra_assign_reg_val): New static inline function.
24988 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
24989 to assign register content record.
24990 (initialize_lra_reg_info_element): Initial offset to zero.
24992 2013-04-23 Catherine Moore <clm@codesourcery.com>
24994 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
24995 operands. Record compression.
24997 2013-04-23 Xinliang David Li <davidxl@google.com>
24999 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
25001 2013-04-23 Richard Biener <rguenther@suse.de>
25003 PR middle-end/57036
25004 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
25005 parameter, only add abnormal goto edges from the copied body
25006 if the call could perform abnormal gotos.
25007 (copy_cfg_body): Adjust.
25009 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
25011 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
25013 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
25015 * coretypes.h (gimple_stmt_iterator): Add struct to make
25018 2013-04-23 Richard Biener <rguenther@suse.de>
25020 PR tree-optimization/57026
25021 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
25022 from SSA names occuring in abnormal PHI nodes.
25024 2013-04-22 Andi Kleen <ak@linux.intel.com>
25026 * lto/lto.c (print_lto_report_1): Fix LTO report names.
25028 2013-04-22 Andi Kleen <ak@linux.intel.com>
25030 * lto/lto.c (print_lto_report_1): Declare early.
25031 (read_cgraph_and_symbols): Call print_lto_report_1 early.
25033 2013-04-22 Andi Kleen <ak@linux.intel.com>
25035 * common.opt (-flto-report-wpa): Add.
25036 * doc/invoke.texi (-flto-report-wpa): Add.
25037 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
25040 2013-04-22 Xinliang David Li <davidxl@google.com>
25042 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
25043 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
25044 * Makefile.in: New dependency
25046 David Daney <ddaney.cavm@gmail.com>
25048 * configure.ac (gcc_cv_as_micromips_support): Use the
25049 --fatal-warnings option.
25050 * configure: Regenerate.
25052 2013-04-22 Marek Polacek <polacek@redhat.com>
25055 * tsan.c (instrument_expr): Don't instrument expression
25056 in case its size is zero.
25058 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
25062 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25064 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25066 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
25068 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
25069 (gimple_stmt_iterator): New typedef.
25070 * gimple.h (gimple_stmt_iterator): Rename to...
25071 (gimple_stmt_iterator_d): ... This.
25072 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
25073 trees be valid for GIMPLE and GENERIC.
25074 (TARGET_GIMPLE_FOLD_BUILTIN): New.
25075 * gimple-fold.c (gimple_fold_call): Call target hook
25076 gimple_fold_builtin.
25077 * hooks.c (hook_bool_gsiptr_false): New.
25078 * hooks.h (hook_bool_gsiptr_false): New.
25079 * target.def (fold_stmt): New.
25080 * doc/tm.texi: Regenerate.
25082 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
25085 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
25086 a set sp if no stack realignment.
25088 2013-04-22 Nick Clifton <nickc@redhat.com>
25090 * config.gcc (tilegx-linux): Extend extra_objs rather than
25092 (tilepro-linux): Likewise.
25094 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
25096 * config/aarch64/aarch64-builtins.c
25098 (CF0, CF1, CF2, CF3, CF4, CF10): New.
25099 (VAR<1-12>): Add MAP parameter.
25100 (BUILTIN_*): Likewise.
25101 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
25102 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
25103 (aarch64_ushl_n<mode>): Likewise.
25104 (aarch64_sshr_n<mode>): Likewise.
25105 (aarch64_ushr_n<mode>): Likewise.
25106 (aarch64_<maxmin><mode>): Likewise.
25107 (aarch64_sqrt<mode>): Likewise.
25108 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
25109 (vshr<q>_n_*): Likewise.
25111 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
25113 * config/aarch64/aarch64-builtins.c
25114 (aarch64_simd_builtin_type_mode): Handle SF types.
25116 (BUILTIN_GPF): Define.
25117 (aarch64_init_simd_builtins): Handle SF types.
25118 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
25119 (frecps): Likewise.
25120 (frecpx): Likewise.
25121 * config/aarch64/aarch64-simd.md
25122 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
25123 (aarch64_frecpe<mode>): New.
25124 (aarch64_frecps<mode>): Likewise.
25125 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
25126 (v8type): Add frecp<esx>.
25127 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
25128 (aarch64_frecps<mode>): Likewise.
25129 * config/aarch64/iterators.md (FRECP): New.
25130 (frecp_suffix): Likewise.
25131 * config/aarch64/arm_neon.h
25132 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
25134 2013-04-22 Christian Bruel <christian.bruel@st.com>
25137 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
25138 (REG_CLASS_NAMES): Idem.
25139 (REG_CLASS_CONTENTS): Idem.
25140 (REGCLASS_HAS_FP_REG): Idem.
25141 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
25142 (sh_conditional_register_usage): Idem.
25144 2013-04-21 Jeff Law <law@redhat.com>
25146 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
25147 (ssa_forward_propagate_and_combine): Use it.
25149 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
25151 * lra.c: Update the flow chart diagram.
25153 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
25155 PR rtl-optimization/56847
25156 * lra-constraints.c (process_alt_operands): Discourage alternative
25157 with non-matche doffsettable memory constraint fro memory with
25160 2013-04-19 Richard Biener <rguenther@suse.de>
25162 PR tree-optimization/56982
25163 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
25165 * gimplify.c (gimplify_call_expr): Notice special calls.
25166 (gimplify_modify_expr): Likewise.
25167 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
25168 abnormal control flow receivers.
25169 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
25170 in the same way as cfun->has_nonlocal_labels.
25171 (gimple_purge_dead_abnormal_call_edges): Likewise.
25172 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
25173 receivers start a basic-block.
25175 2013-04-19 Richard Biener <rguenther@suse.de>
25177 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
25179 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
25180 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
25181 (SLP_TREE_LOAD_PERMUTATION): Add.
25182 (vect_transform_slp_perm_load): Adjust prototype.
25183 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
25184 (vect_free_slp_instance): Likewise.
25185 (vect_create_new_slp_node): Likewise.
25186 (vect_supported_slp_permutation_p): Remove.
25187 (vect_slp_rearrange_stmts): Adjust.
25188 (vect_supported_load_permutation_p): Likewise. Inline
25189 vect_supported_slp_permutation_p here.
25190 (vect_analyze_slp_instance): Compute load permutations per
25191 slp node instead of per instance.
25192 (vect_get_slp_defs): Adjust.
25193 (vect_transform_slp_perm_load): Likewise.
25194 (vect_schedule_slp_instance): Remove redundant code.
25195 (vect_schedule_slp): Remove hack for PR56270, add it ...
25196 * tree-vect-stmts.c (vectorizable_load): ... here, do not
25197 CSE loads for SLP. Adjust.
25199 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
25201 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
25202 spelling in two comments.
25204 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
25207 * config/arm/arm.c (load_multiple_sequence): Require SP
25208 as base register for loads if SP is in the register list.
25210 2013-04-19 Martin Jambor <mjambor@suse.cz>
25212 PR tree-optimization/56718
25213 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
25214 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
25215 and made public. Adjusted all callers.
25216 (ipa_intraprocedural_devirtualization): New function.
25217 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
25218 (ipa_intraprocedural_devirtualization): Likewise.
25219 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
25221 2013-04-19 Richard Biener <rguenther@suse.de>
25223 PR tree-optimization/57000
25224 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
25226 2013-04-19 Terry Guo <terry.guo@arm.com>
25228 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
25230 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
25231 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
25232 (cortex_m4_fmacs): Use new reservations.
25233 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
25235 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
25237 PR rtl-optimization/56999
25238 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
25240 (lra_coalesce): Remove split_origin_bitmap and related code.
25241 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
25242 ranges if necessary.
25244 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
25246 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
25248 (ix86_expand_call): Remove clobbered_registers array and use
25249 x86_64_ms_sysv_extra_clobbered_registers instead.
25250 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
25252 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
25254 * config/i386/i386.md (*call_rex64_ms_sysv): Use
25255 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
25256 (*call_value_rex64_ms_sysv): Ditto.
25258 2013-04-18 Cary Coutant <ccoutant@google.com>
25260 * dwarf2out.c (output_pubnames): Check die_perennial_p of
25261 parent instead of die_mark.
25263 2013-04-18 Diego Novillo <dnovillo@google.com>
25265 * gimple.c (create_gimple_tmp): New.
25266 (get_expr_type): New.
25267 (build_assign): New.
25268 (build_type_cast): New.
25269 * gimple.h (enum ssa_mode): Define.
25270 (gimple_seq_set_location): New.
25271 * asan.c (build_check_stmt): Change some gimple_build_* calls
25272 to use build_assign and build_type_cast.
25274 2013-04-18 Richard Biener <rguenther@suse.de>
25276 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
25277 handle negative step. Remove redundant checks.
25278 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
25279 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
25280 for negative step and grouped loads fail to vectorize.
25282 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
25284 * emit-rtl.c (reset_insn_used_flags): New function.
25285 (reset_all_used_flags): Use it.
25286 (verify_insn_sharing): New function.
25287 (verify_rtl_sharing): Fix verification for SEQUENCEs.
25289 2013-04-18 Jakub Jelinek <jakub@redhat.com>
25291 PR tree-optimization/56984
25292 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
25293 and (x >> M) >= N don't register any assertion if N << M is the
25296 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
25298 * lower-subreg.c (resolve_simple_move): If called self-recursive,
25299 do not delete_insn insns that have not yet been emitted, only
25300 unlink them with remove_insn.
25301 * df-scan.c (df_insn_delete): Revert r197492.
25303 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
25305 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
25306 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
25308 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
25310 * config/arm/arm.md (movsicc_insn): Convert define_insn into
25311 define_insn_and_split.
25312 (and_scc,ior_scc,negscc): Likewise.
25313 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
25315 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
25317 * config/arm/arm.c (use_return_insn): Return 0 for targets that
25318 can benefit from using a sequence of LDRD instructions in epilogue
25319 instead of a single LDM instruction.
25321 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
25324 * doc/extend.texi: Fix typo.
25326 2013-04-17 Richard Biener <rguenther@suse.de>
25328 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
25329 (vect_build_slp_tree): ... here.
25330 (vect_build_slp_tree_1): Compute which stmts of the SLP group
25331 match. Remove special-casing of mismatched complex loads.
25332 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
25333 re-try the match with swapped commutative operands.
25334 (vect_supported_load_permutation_p): Remove special-casing of
25335 mismatched complex loads.
25336 (vect_analyze_slp_instance): Adjust.
25338 2013-04-17 Richard Biener <rguenther@suse.de>
25340 PR rtl-optimization/56921
25341 * cfgloop.h (struct loop): Add simple_loop_desc member.
25342 (struct niter_desc): Mark with GTY(()).
25343 (simple_loop_desc): Do not use aux field but simple_loop_desc.
25344 * loop-iv.c (get_simple_loop_desc): Likewise.
25345 (free_simple_loop_desc): Likewise.
25348 2013-04-16 Richard Biener <rguenther@suse.de>
25350 PR rtl-optimization/56921
25351 * loop-init.c (pass_rtl_move_loop_invariants): Add
25352 TODO_do_not_ggc_collect to todo_flags_finish.
25353 (pass_rtl_unswitch): Same.
25354 (pass_rtl_unroll_and_peel_loops): Same.
25355 (pass_rtl_doloop): Same.
25357 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
25359 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
25360 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
25361 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
25363 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
25364 * tree-streamer.c (record_common_node): Adjust reference in comment.
25366 2013-04-17 Terry Guo <terry.guo@arm.com>
25368 * config/arm/cortex-m4.md: Add a new bypass.
25370 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25372 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
25374 (*subs_<optab><mode>_multp2): New pattern.
25375 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
25376 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
25378 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25380 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
25381 (*subs_mul_imm_<mode>): New pattern.
25383 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
25386 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
25387 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
25388 (vsx_movti_32bit): Change j->wa to O->wa.
25390 2013-04-16 Richard Biener <rguenther@suse.de>
25392 PR rtl-optimization/56921
25393 * loop-init.c (pass_rtl_move_loop_invariants): Add
25394 TODO_do_not_ggc_collect to todo_flags_finish.
25395 (pass_rtl_unswitch): Same.
25396 (pass_rtl_unroll_and_peel_loops): Same.
25397 (pass_rtl_doloop): Same.
25399 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
25401 * config/arm/arm.c (emit_multi_reg_push): New declaration
25402 for an existing function.
25403 (arm_emit_strd_push): New function.
25404 (arm_expand_prologue): Used here.
25405 (arm_emit_ldrd_pop): New function.
25406 (arm_expand_epilogue): Used here.
25407 (arm_get_frame_offsets): Update condition.
25408 (arm_emit_multi_reg_pop): Add a special case for load of a single
25409 register with writeback.
25411 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
25413 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
25416 2013-04-16 Richard Biener <rguenther@suse.de>
25418 PR tree-optimization/56756
25419 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
25420 (first_mem_ref_loc): New.
25421 (execute_sm): Place the load temporarily before a previous
25422 access instead of in the latch edge to ensure its SSA dependencies
25423 are defined at points dominating the load.
25425 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
25427 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
25428 correct fix by moving header and footer insn to the footer of
25429 the merged basic block. Clear BB_END of the merged-away block.
25431 PR middle-end/43631
25432 * emit-rtl.c (make_note_raw): New function.
25433 (link_insn_into_chain): New static inline function.
25434 (add_insn): Use it.
25435 (add_insn_before, add_insn_after): Factor insn chain linking code...
25436 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
25437 using link_insn_into_chain.
25438 (note_outside_basic_block_p): New helper function for emit_note_after
25439 and emit_note_before.
25440 (emit_note_after): Use nobb variant of add_insn_after if the note
25441 should not be contained in a basic block.
25442 (emit_note_before): Use nobb variant of add_insn_before if the note
25443 should not be contained in a basic block.
25444 (emit_note_copy): Use make_note_raw.
25445 (emit_note): Likewise.
25446 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
25447 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
25448 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
25449 the moved barrier the tail of the basic block it follows.
25450 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
25452 2013-04-15 Jakub Jelinek <jakub@redhat.com>
25454 PR tree-optimization/56962
25455 * gimple-ssa-strength-reduction.c (record_increment): Only set
25456 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
25457 either rhs1 or rhs2 is equal to c->base_expr.
25459 2013-04-15 Richard Biener <rguenther@suse.de>
25461 PR tree-optimization/56933
25462 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
25464 (GROUP_READ_WRITE_DEPENDENCE): Remove.
25465 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
25466 * tree-vect-data-refs.c (vect_analyze_group_access): Move
25467 dependence check ...
25468 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
25470 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
25471 GROUP_READ_WRITE_DEPENDENCE.
25473 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25475 * emit-rtl.c (reset_all_used_flags): New function.
25476 (verify_rtl_sharing): Call reset_all_used_flags before and after
25477 performing the checks.
25479 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25481 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
25482 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
25483 * config/arm/constraints.md (De): New constraint.
25484 * config/arm/neon.md (anddi3_neon): Delete.
25485 (neon_vand<mode>): Expand to standard anddi3 pattern.
25486 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
25487 Move earlier in the file.
25488 (neon_inv_logic_op2): Likewise.
25489 (arm_anddi_operand_neon): New predicate.
25491 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25493 * configure.ac (gcc_cv_ld_as_needed): Set
25494 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
25495 Use -z ignore, -z record on *-*-solaris2*.
25496 (HAVE_LD_AS_NEEDED): Update comment.
25497 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
25498 * configure: Regenerate.
25499 * config.in: Regenerate.
25500 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
25501 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
25502 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
25503 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
25504 equivalents. Fix markup.
25505 * doc/tm.texi: Regenerate.
25507 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
25509 * config/i386/i386.opt: New option mstack-protector-guard=.
25510 * config/i386/i386-opts.h: Add enum stack_protector_guard.
25511 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
25512 TARGET_SSP_TLS_GUARD.
25513 * config/i386/i386.c (ix86_option_override_internal): Set
25514 ix86_stack_protector_guard.
25515 * config/i386/i386.md (stack_protect_set): Enable for
25516 TARGET_SSP_TLS_GUARD only.
25517 (stack_protect_set_<mode>): Ditto.
25518 (stack_protect_test): Ditto.
25519 (stack_protect_test_<mode>): Ditto.
25520 * doc/invoke.texi (i386 Option): Document.
25522 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
25525 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
25526 (S_MODES): Set H_MODE bit.
25527 (SF_MODES): Set only S_MODE and SF_MODE bits.
25528 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
25529 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
25530 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
25531 <MODE_FLOAT>: Likewise.
25533 2013-04-15 Joey Ye <joey.ye@arm.com>
25535 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
25537 2013-04-15 Joey Ye <joey.ye@arm.com>
25539 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
25541 (thumb_far_jump_used_p): Count instruction size and set
25544 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
25546 * reorg.c (fill_simple_delay_slots): Reindent block of code.
25547 * resource.c (mark_target_live_regs): Reformat conditional block.
25549 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
25551 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
25552 notes, they are emitted only just before final.
25553 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
25555 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
25557 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
25558 * cfgrtl.c (delete_insn): Call it here instead.
25559 * lra-spills.c (lra_final_code_change): Use delete_insn.
25560 * haifa-sched.c (sched_remove_insn): Likewise.
25561 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
25562 returning to the nop pool.
25563 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
25564 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
25566 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
25568 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
25569 * doc/tm.texi: Regenerated.
25571 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
25573 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
25576 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
25578 * df-core.c (df_find_def): Compare register numbers.
25579 (df_find_use): Likewise.
25581 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
25584 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
25585 lra_in_progress for return.
25587 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
25589 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
25590 define_insn into define_insn_and_split and emit movsicc patterns.
25592 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
25594 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
25596 2013-04-12 Richard Biener <rguenther@suse.de>
25598 * tree-pass.h (TODO_do_not_ggc_collect): New.
25599 * passes.c (execute_one_ipa_transform_pass): Honor
25600 TODO_do_not_ggc_collect.
25601 (execute_one_pass): Likewise.
25604 2013-04-10 Richard Biener <rguenther@suse.de>
25606 * passes.c (init_optimization_passes): Remove reload pass.
25607 * ira.c (do_reload): Merge into ...
25609 (rest_of_handle_reload): Remove.
25610 (pass_reload): Likewise.
25611 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25612 of reload for vzeroupper pass placement.
25614 2013-04-12 Jakub Jelinek <jakub@redhat.com>
25616 PR tree-optimization/56918
25617 PR tree-optimization/56920
25618 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
25619 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
25620 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
25621 use wide_mul_with_sign method.
25623 2013-04-12 Richard Biener <rguenther@suse.de>
25625 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
25626 not be considered a gimple constant.
25628 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
25630 * fold-const.c (const_binop): Handle vector shifts by a scalar.
25631 (fold_binary_loc): Call const_binop also for mixed vector-scalar
25634 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
25635 Jakub Jelinek <jakub@redhat.com>
25637 * opts.c: Include diagnostic-color.h.
25638 (common_handle_option): Handle OPT_fdiagnostics_color_.
25639 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
25640 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
25641 (diagnostic-color.o): New.
25642 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
25643 (diagnostic_color_rule): New enum.
25644 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
25645 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
25646 the location string.
25647 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
25648 either NULL, or color kind.
25649 * diagnostic-color.c: New file.
25650 * diagnostic-color.h: New file.
25651 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
25653 * doc/invoke.texi (-fdiagnostics-color): Document.
25654 * pretty-print.h (pp_show_color): Define.
25655 (struct pretty_print_info): Add show_color field.
25656 * diagnostic.c: Include diagnostic-color.h.
25657 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
25658 macros. Colorize error:, warning: etc. strings and also the location
25660 (diagnostic_show_locus): Colorize the caret line.
25661 * pretty-print.c: Include diagnostic-color.h.
25662 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
25663 inside of %< %> quotes or quoted through q format modifier.
25665 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25667 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
25669 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25671 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
25672 code in CC_NZ mode.
25673 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
25676 2013-04-11 Marek Polacek <polacek@redhat.com>
25678 PR tree-optimization/48184
25679 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
25681 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
25683 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
25684 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
25685 (skip_simple_arithmetic): Tidy up.
25686 * tree.h (skip_simple_constant_arithmetic): Declare.
25688 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25690 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
25692 2013-04-11 Richard Biener <rguenther@suse.de>
25694 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25695 generate vector constants.
25697 2013-04-11 Richard Biener <rguenther@suse.de>
25699 PR tree-optimization/56878
25700 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
25701 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
25703 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
25704 Prefer to align the DR with the most invariant base address.
25706 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25708 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
25711 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
25713 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
25714 floating-point vector comparisons against 0.
25716 2013-04-11 Jakub Jelinek <jakub@redhat.com>
25718 PR tree-optimization/56899
25719 * fold-const.c (extract_muldiv_1): Apply distributive law
25720 only if TYPE_OVERFLOW_WRAPS (ctype).
25722 2013-04-11 Bin Cheng <bin.cheng@arm.com>
25725 * ira-costs.c (scan_one_insn): Check whether the source rtx of
25726 loading has side effect.
25728 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
25730 * config/sparc/sparc.c: Include tree-pass.h.
25731 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
25732 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
25733 head of file. Change return type. Split off gate function.
25734 (sparc_gate_work_around_errata): New function.
25735 (pass_work_around_errata): New pass definition.
25736 (insert_pass_work_around_errata) New pass insert definition to
25737 insert pass_work_around_errata just after delayed-branch scheduling.
25738 (sparc_option_override): Insert the pass.
25739 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
25741 2013-04-10 David S. Miller <davem@davemloft.net>
25743 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
25744 or -mcpu=hypersparc.
25746 * target.def (cstore_mode): New hook.
25747 * target.h: Include insn-codes.h
25748 * targhooks.c: Likewise.
25749 (default_cstore_mode): New function.
25750 * targhooks.h: Declare it.
25751 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
25752 * doc/tm.texi: Rebuild.
25753 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
25754 target hook, rather than inspecting the insn_data.
25755 * config/sparc/sparc.c (sparc_cstore_mode): New function.
25756 (TARGET_CSTORE_MODE): Redefine.
25757 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
25759 * config/sparc/predicates.md (cstore_result_operand): New special
25761 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
25762 Use it for operand 0.
25763 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
25764 (*snesi_special): Likewise.
25765 (*snesi_zero): Likewise.
25766 (*seqsi_zero): Likewise.
25767 (*sltu_insn): Likewise.
25768 (*sgeu_insn): Likewise.
25769 (*seqdi_special): Make operand 0 and comparison operation be of
25771 (*snedi_special): Likewise.
25772 (*snedi_special_vis3): Likewise.
25773 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
25774 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
25775 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
25776 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
25777 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
25778 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
25779 (*sltu_extend_sp64): Likewise.
25780 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
25781 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
25782 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
25783 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
25784 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
25786 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
25788 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
25789 (aarch64_start_file): Use the new function.
25791 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25792 Jason Merrill <jason@redhat.com>
25794 * common.opt: Add -gdwarf.
25795 * opts.c (common_handle_option): Handle it.
25796 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
25798 2013-04-10 Richard Biener <rguenther@suse.de>
25800 * passes.c (execute_todo): Do not call ggc_collect conditional here.
25801 (execute_one_ipa_transform_pass): But unconditionally here.
25802 (execute_one_pass): And here.
25803 (init_optimization_passes): Remove reload pass.
25804 * tree-pass.h (TODO_ggc_collect): Remove.
25805 (pass_reload): Likewise.
25806 * ira.c (do_reload): Merge into ...
25808 (rest_of_handle_reload): Remove.
25809 (pass_reload): Likewise.
25810 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25811 of reload for vzeroupper pass placement.
25812 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
25813 and todo_flags_finish of all passes.
25815 2013-04-10 Richard Biener <rguenther@suse.de>
25817 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
25818 first_const_oprnd field, rename first_def_type to first_op_type.
25819 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
25820 (vect_get_and_check_slp_defs): Always use the type of the
25821 operand. Allow mixed vect_external_def, vect_constant_def types.
25822 (vect_get_constant_vectors): Handle mixed vect_external_def,
25823 vect_constant_def types.
25825 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
25827 PR tree-optimization/55524
25828 * tree-ssa-math-opts.c
25829 (convert_mult_to_fma): Don't use an fms construct
25830 when we don't have an fms operation, but fnma, and it looks
25831 likely that we'll be able to use the latter.
25833 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
25835 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
25837 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
25838 inline fail caused by overwritable functions.
25840 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
25842 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
25843 unnecessary bits in the constant power of two case.
25845 2013-04-10 Richard Biener <rguenther@suse.de>
25847 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
25848 broken code swapping operands.
25849 (vect_build_slp_tree): Do not compute load permutations here.
25850 (vect_analyze_slp_instance): Compute load permutations here,
25851 after building the SLP tree.
25853 2013-04-09 Christian Bruel <christian.bruel@st.com>
25855 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
25856 of next/prev_real_insn.
25858 2013-04-09 Jan Hubicka <jh@suse.cz>
25860 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
25861 Drop aliased parameter.
25862 (function_and_variable_visibility): Do not handle alias pairs.
25863 * cgraph.c (varpool_externally_visible_p): Update prototype.
25864 * varpool.c (varpool_add_new_variable): Update.
25866 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25868 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
25870 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
25872 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
25874 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
25876 2013-04-09 Marek Polacek <polacek@redhat.com>
25878 PR tree-optimization/48762
25879 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
25881 2013-04-09 Richard Biener <rguenther@suse.de>
25883 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
25885 (vect_build_slp_tree): Likewise.
25886 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
25887 calculating the cost of a SLP instance.
25888 (vect_analyze_slp_instance): Use it from here, after building
25891 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25893 PR middle-end/56883
25894 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
25895 expand_omp_for_static_chunk): Use simple_p = true in
25896 force_gimple_operand_gsi calls when assigning to addressable decls.
25898 2013-04-09 Jeff Law <law@redhat.com>
25900 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
25901 when the boolean was created by converting a wider object which
25902 had a boolean range.
25904 2013-04-09 Richard Biener <rguenther@suse.de>
25906 * tree-vectorizer.h (slp_void_p): Remove.
25907 (slp_tree): Typedef before _slp_tree declaration.
25908 (struct _slp_tree): Use a vector of slp_tree as children.
25909 (vect_get_place_in_interleaving_chain): Remove.
25910 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
25912 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
25914 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
25915 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
25916 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
25917 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
25918 Use slp_node instead of slp_void_p and adjust.
25920 2013-04-09 Richard Biener <rguenther@suse.de>
25922 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
25923 work that is not necessary.
25925 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25927 PR tree-optimization/56854
25928 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
25929 forward into clobber stmts if it would change MEM_REF lhs into
25932 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
25934 * tree.c (type_hash_lookup, type_hash_add): Make static.
25935 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
25937 2013-04-09 Richard Biener <rguenther@suse.de>
25939 * tree.h (unsave_expr_now): Remove.
25940 * tree-inline.c (mark_local_for_remap_r): Remove.
25941 (unsave_expr_1): Likewise.
25942 (unsave_r): Likewise.
25943 (unsave_expr_now): Likewise.
25944 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
25945 (propagate_tree_value): Likewise.
25947 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
25949 * doc/rtl.texi (sequence): Rewrite documentation to match the
25950 current use of SEQUENCE rtl objects.
25951 * rtl.def (SEQUENCE): Likewise.
25953 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
25954 Update documentation.
25955 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
25956 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
25958 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
25960 2013-04-08 Teresa Johnson <tejohnson@google.com>
25962 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
25963 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
25965 (estimate_edge_size_and_time): Add comment to suggest using rounding
25967 (estimate_node_size_and_time): Ditto.
25968 (remap_edge_change_prob): Use helper rounding divide methods.
25969 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
25970 (gimple_mod_pow2_value_transform): Ditto.
25971 (gimple_mod_subtract_transform): Ditto.
25972 (gimple_ic_transform): Ditto.
25973 (gimple_stringops_transform): Ditto.
25974 * stmt.c (conditional_probability): Ditto.
25975 (emit_case_dispatch_table): Ditto.
25976 * lto-cgraph.c (merge_profile_summaries): Ditto.
25977 * tree-optimize.c (execute_fixup_cfg): Ditto.
25978 * cfgcleanup.c (try_forward_edges): Ditto.
25979 * cfgloopmanip.c (scale_loop_profile): Ditto.
25981 (duplicate_loop_to_header_edge): Ditto.
25982 (lv_adjust_loop_entry_edge): Ditto.
25983 * tree-vect-loop.c (vect_transform_loop): Ditto.
25984 * profile.c (compute_branch_probabilities): Ditto.
25985 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
25986 * lto-streamer-in.c (input_cfg): Ditto.
25987 * gimple-streamer-in.c (input_bb): Ditto.
25988 * ipa-cp.c (update_profiling_info): Ditto.
25989 (update_specialized_profile): Ditto.
25990 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
25991 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
25993 * sched-rgn.c (compute_dom_prob_ps): Ditto.
25994 (compute_trg_info): Ditto.
25995 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
25996 (purge_dead_edges): Ditto.
25997 * loop-unswitch.c (unswitch_loop): Ditto.
25998 * cgraphclones.c (cgraph_clone_edge): Ditto.
25999 (cgraph_clone_node): Ditto.
26000 * tree-inline.c (copy_bb): Ditto.
26001 (copy_edges_for_bb): Ditto.
26002 (initialize_cfun): Ditto.
26003 (copy_cfg_body): Ditto.
26004 (expand_call_inline): Ditto.
26006 2013-04-08 Kai Tietz <ktietz@redhat.com>
26008 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
26009 TARGET_CYGWIN64 by TARGET_64BIT.
26011 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
26013 * config/epiphany/epiphany.md (GPR_1): New constant.
26014 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
26015 * config/epiphany/epiphany.c (gen_compare_reg):
26016 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
26017 is already in place.
26018 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
26019 Don't require being called during rtl expansion; If y operlaps r0,
26021 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
26022 (epiphany_expand_epilogue): Likewise.
26024 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
26025 Don't use CC_FPmode for ORDERED / UNORDERED.
26026 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
26028 * config/epiphany/constraints.md (CnL): New constraint.
26029 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
26030 * config/epiphany/predicates.md (add_operand): Allow 1024.
26032 * config/epiphany/epiphany.md (logical_op): New code iterator.
26033 (op_mnc): New code attribute.
26034 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
26035 (mov_f+1, mov_f+2): New peephole2 patterns.
26037 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
26038 (cstoresi4): Also allow re-use of zero result when doing a NE
26039 comparison to a non-zero operand.
26040 Use (clobber (scratch)) for first insn if the gpr output is not needed.
26042 * config/epiphany/epiphany.md (<insn_opname>v2si3):
26043 Use gen_addsi3_i / gen_subsi3_i.
26045 2013-04-08 Jakub Jelinek <jakub@redhat.com>
26049 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
26050 contain anything but clobbers, at most one __builtin_stack_restore,
26051 optionally debug stmts and final resx, and if it has at least one
26052 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
26053 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
26054 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
26055 which isn't defaut definition, remove them.
26056 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
26057 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
26058 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
26059 with MEM_REF LHS with SSA_NAME address.
26061 2013-04-08 Jeff Law <law@redhat.com>
26063 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
26065 2013-04-08 Richard Biener <rguenther@suse.de>
26067 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
26069 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
26070 determined vector type.
26071 (vect_analyze_data_refs): Likewise.
26072 (vect_get_new_vect_var): Adjust.
26073 (vect_create_destination_var): Preserve SSA name versions.
26074 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
26075 not dump anything here.
26077 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
26079 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
26080 Add member lr_slot_known.
26081 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
26083 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
26084 Remove code that sets lr_slot_offset according to what a previous
26085 version of epiphany_emit_save_restore used to do.
26086 (epiphany_emit_save_restore): When doing an lr save or restore,
26087 set/verify lr_slot_known and lr_slot_offset.
26089 2013-04-08 Xinyu Qi <xyqi@marvell.com>
26092 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
26095 2013-04-08 Richard Biener <rguenther@suse.de>
26097 * alias.c (find_base_term): Fix thinko in previous change.
26099 2013-04-08 Jakub Jelinek <jakub@redhat.com>
26101 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
26102 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
26103 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
26104 if possible to compute val.
26105 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
26106 For QImode integers don't require anything about precision. Use
26107 const_with_all_bytes_same to find out if the constant doesn't have
26108 repeated bytes in it.
26110 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26112 * config/s390/s390.c (s390_expand_insv): Only accept insertions
26115 2013-04-08 Marek Polacek <polacek@redhat.com>
26117 PR rtl-optimization/48182
26118 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
26121 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26124 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
26125 nuses, make sure we have a label.
26127 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
26130 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
26131 (rs6000_emit_swdiv_low_precision): Remove.
26132 (rs6000_emit_swdiv): Rewrite to handle between one and four
26133 iterations of Newton-Raphson generally; modify required number of
26134 iterations for some cases.
26135 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
26137 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
26139 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
26140 set-but-unused variable.
26142 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
26143 basic blocks of released function bodies garbage-collectable.
26145 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
26146 (struct rtl_opt_pass): Add TODO_df_finish.
26148 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
26150 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26152 * config/arm/constraints.md (q): New constraint.
26153 * config/arm/ldrdstrd.md: New file.
26154 * config/arm/arm.md (ldrdstrd.md) New include.
26155 (arm_movdi): Use "q" instead of "r" constraint
26156 for double-word memory access.
26157 (movdf_soft_insn): Likewise.
26158 * config/arm/vfp.md (movdi_vfp): Likewise.
26159 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
26160 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
26161 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
26162 (mem_ok_for_ldrd_strd): Likewise.
26163 (output_move_double): Update assertion.
26165 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26167 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
26169 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26171 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
26172 define_insn_and_split.
26173 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
26175 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26177 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
26178 define_insn_and_split.
26179 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
26180 (shiftsi3_compare): New pattern.
26181 (rrx): New pattern.
26182 * config/arm/unspecs.md (UNSPEC_RRX): New.
26184 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26186 * config/arm/arm.md (negdi_extendsidi): New pattern.
26187 (negdi_zero_extendsidi): Likewise.
26189 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26191 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
26192 define_insn_and_split.
26193 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
26194 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
26196 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26198 * config/arm/arm.md (arm_subdi3): Convert define_insn into
26199 define_insn_and_split.
26200 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
26201 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
26203 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26205 * config/arm/arm.md (subsi3_carryin): New pattern.
26206 (subsi3_carryin_const): Likewise.
26207 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
26208 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
26210 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26212 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
26214 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
26216 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
26217 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
26219 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26221 * config/arm/arm.c (arm_expand_builtin): Change fcode
26222 type to unsigned int.
26224 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26226 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
26228 2013-04-04 Ian Lance Taylor <iant@google.com>
26230 * doc/standards.texi (Standards): The Go frontend supports the Go 1
26233 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
26235 PR middle-end/56729
26236 * df-scan.c (df_insn_delete): Disable failing assert.
26238 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26240 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
26241 New function prototype.
26242 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
26243 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
26244 (arm_builtin_vectorized_function): New function.
26246 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26248 * config/arm/arm_neon_builtins.def: New file.
26249 * config/arm/arm.c (neon_builtin_data): Move contents to
26250 arm_neon_builtins.def.
26251 (enum arm_builtins): Include neon builtin definitions.
26252 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
26253 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
26255 2013-04-04 Marek Polacek <polacek@redhat.com>
26257 PR tree-optimization/48186
26258 * predict.c (maybe_hot_frequency_p): Return false if
26259 HOT_BB_FREQUENCY_FRACTION is 0.
26260 (cgraph_maybe_hot_edge_p): Likewise.
26262 2013-04-04 Richard Biener <rguenther@suse.de>
26264 PR tree-optimization/56826
26265 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
26268 2013-04-04 Richard Biener <rguenther@suse.de>
26270 PR tree-optimization/56213
26271 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
26272 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
26274 2013-04-04 Richard Biener <rguenther@suse.de>
26276 PR tree-optimization/56837
26277 * tree-loop-distribution.c (classify_partition): For non-zero
26278 values require that the value has the same precision as its
26279 mode to be useful as memset value.
26281 2013-04-03 Nick Clifton <nickc@redhat.com>
26283 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
26284 (fmssf4): Use fmsf.s on E3V5 architectures.
26285 (fnmasf4): Use fnmaf.s on E3V5 architectures.
26286 (fnmssf4): Use fnmsf.s on E3V5 architectures.
26288 2013-04-03 Jeff Law <law@redhat.com>
26290 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
26291 (lra-eliminations.o): Likewise.
26293 2013-04-03 Teresa Johnson <tejohnson@google.com>
26295 * gcov-io.c (compute_working_sets): Moved most of body of old
26296 compute_working_sets here from profile.c.
26297 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
26298 (gcov_working_set_t): Moved typedef here from basic-block.h
26299 (compute_working_set): Declare.
26300 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
26301 (get_working_sets): Renamed from compute_working_set,
26302 replace most of body with call to new compute_working_sets.
26303 (get_exec_counts): Replace call to compute_working_sets
26304 to get_working_sets.
26305 * profile.h (get_working_sets): Renamed from compute_working_set.
26306 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
26307 to get_working_sets.
26308 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
26309 * gcov-dump.c (dump_working_sets): New function.
26311 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
26313 * hwint.c (sext_hwi, zext_hwi): New functions.
26314 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
26315 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
26316 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
26317 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
26318 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
26319 (sext_hwi, zext_hwi): New functions.
26321 2013-04-03 Jeff Law <law@redhat.com>
26323 PR tree-optimization/56799
26324 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
26325 back test for widening conversion erroneously dropped in prior change.
26327 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26330 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
26331 instead of next_real_insn.
26333 2013-04-03 Marek Polacek <polacek@redhat.com>
26336 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
26338 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26341 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
26343 (thumb1_output_casesi): Likewise.
26344 (thumb2_output_casesi): Likewise.
26346 2013-04-03 Richard Biener <rguenther@suse.de>
26348 PR tree-optimization/56817
26349 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
26351 (tree_unroll_loops_completely_1): ... new function to manually
26352 walk the loop tree, properly defering outer loops of unrolled
26353 loops to later iterations.
26355 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26357 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
26358 (vectorizable_load): Likewise.
26359 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
26360 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
26362 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26364 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
26367 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
26369 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
26371 2013-04-03 Bin Cheng <bin.cheng@arm.com>
26373 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
26375 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26377 PR tree-optimization/56790
26378 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
26381 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26383 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
26385 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
26386 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
26389 2013-04-03 Jakub Jelinek <jakub@redhat.com>
26392 * tree.h (force_folding_builtin_constant_p): New decl.
26393 * builtins.c (force_folding_builtin_constant_p): New variable.
26394 (fold_builtin_constant_p): Fold immediately also if
26395 force_folding_builtin_constant_p.
26397 2013-04-03 Richard Biener <rguenther@suse.de>
26399 PR tree-optimization/56812
26400 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
26401 DRs of the same interleaving chain are independent.
26403 2013-04-02 Jason Merrill <jason@redhat.com>
26405 * gdbinit.in (pbb): Use debug fn.
26407 2013-04-02 Lawrence Crowl <crowl@google.com>
26409 * sese.h (struct ivtype_map_elt_s): Remove unused.
26410 (extern debug_ivtype_map): Remove unused.
26411 (extern eq_ivtype_map_elts): Remove unused.
26412 * sese.c (debug_ivtype_map): Removed unused.
26413 (debug_ivtype_map_1): Removed unused.
26414 (debug_ivtype_elt): Remove unused.
26415 (eq_ivtype_map_elts): Remove unused.
26417 2013-04-02 Kai Tietz <ktietz@redhat.com>
26420 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
26421 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
26422 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
26424 (legitimize_pe_coff_symbol): Likewise.
26425 (is_imported_p): New helper-function.
26426 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
26427 for Windows x64 targets.
26428 (ix86_expand_prologue): Optimize for pe-coff targets.
26429 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
26430 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
26431 medium/large code-model.
26432 (legitimize_pic_address): Likewise.
26433 (legitimize_tls_address): Likewise.
26434 (ix86_expand_call): Likewise.
26435 (x86_output_mi_thunk): Likewise.
26436 (get_dllimport_decl): Add new beimport argument.
26437 (construct_plt_address): Don't assert for x64 pe-coff targets.
26438 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
26440 (SYMBOL_FLAG_STUBVAR): New macro.
26441 (SYMBOL_REF_STUBVAR_P): Likewise.
26442 * config/i386/winnt.c (stub_list): New structure.
26443 (stub_head): New local variable.
26444 (i386_pe_record_stub): New function.
26445 (i386_pe_file_end): Emit refptr-stubs.
26447 2013-04-02 Jakub Jelinek <jakub@redhat.com>
26449 PR rtl-optimization/56745
26450 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
26451 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
26454 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
26455 and both of them are MEM_REFs, just compare first argument for
26456 equality and attempt to deal even with differing offsets.
26459 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
26460 of gimple_clobber_p to be MEM_REF.
26461 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
26462 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
26463 after gimplification.
26464 * asan.c (get_mem_ref_of_assignment): Don't instrument
26465 gimple_clobber_p stmts.
26466 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
26467 gimple_clobber_p stmt if they have MEM_REF lhs and
26468 are dead because of another gimple_clobber_p stmt.
26469 * tree-ssa-live.c (clear_unused_block_pointer): Treat
26470 gimple_clobber_p stmts like debug stmts.
26471 (remove_unused_locals): Remove clobbers with MEM_REF lhs
26472 that refer to unused VAR_DECLs or uninitialized values.
26473 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
26474 gimple_clobber_p stmts if they refer to removed parameters.
26475 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
26478 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
26480 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
26481 using SWI48 mode attribute.
26483 2013-04-02 Wei Mi <wmi@google.com>
26485 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
26486 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
26487 *<rotate_insn><mode>3_mask in i386.md.
26489 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
26491 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
26493 2013-04-02 Richard Biener <rguenther@suse.de>
26495 PR tree-optimization/56778
26496 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
26497 Runtime alias tests are not supported for gather loads.
26498 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
26499 stmts referenced from SSA operands before updating SSA form.
26501 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
26502 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26504 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
26505 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
26506 * config/arm/cortex-a53.md: New file.
26507 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
26508 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
26509 * config/arm/arm.c (arm_issue_rate): Likewise.
26510 * config/arm/arm-tune.md: Regenerate
26511 * config/arm/arm-tables.opt: Regenerate.
26512 * config/arm/arm-cores.def: Add cortex-a53.
26514 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
26516 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
26519 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
26521 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
26522 scalar load/store operations using B/H registers.
26523 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
26525 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
26527 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
26529 * config/aarch64/aarch64.c
26530 (aarch64_simd_scalar_immediate_valid_for_move): New.
26531 * config/aarch64/aarch64-protos.h
26532 (aarch64_simd_scalar_immediate_valid_for_move): New.
26533 * config/aarch64/constraints.md (Dh, Dq): New.
26534 * config/aarch64/iterators.md (hq): New.
26536 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
26538 * reorg.c (get_branch_condition): Deal with conditional returns.
26539 (fill_simple_delay_slots): Remove dead code dealing with jumps.
26541 2013-04-01 Wei Mi <wmi@google.com>
26543 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
26544 Truncate operand 2 using %b asm operand modifier.
26545 (*<shift_insn><mode>3_mask): Ditto.
26546 (*<rotate_insn><mode>3_mask): Ditto.
26548 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
26550 PR middle-end/56798
26551 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
26553 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
26555 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
26557 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
26559 2013-03-30 Lawrence Crowl <crowl@google.com>
26561 * dse.c (clear_alias_sets): Remove never set.
26562 (disqualified_clear_alias_sets): Remove never set.
26563 (clear_alias_mode_pool): Remove never set.
26564 (dse_step0): Remove condition that is never true.
26565 (canon_address): Remove condition that is never true.
26566 (dse_step7): Remove condition that is never true.
26567 (rest_of_handle_dse): Remove condition that is never true.
26568 (rest_of_handle_dse::did_global): Remove never read from above.
26569 (dse_step2_spill): Remove never called from above.
26570 (dse_step5_spill): Remove never called from above.
26572 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
26574 * doc/md.texi (Standard Names) <casesi>: Update documentation for
26575 JUMP_TABLE_DATA changes.
26576 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
26577 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
26578 (Insns) <jump_table_data>: New entry.
26579 * doc/tm.texi: Regenerate.
26581 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
26583 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
26584 for table jump at the end of a basic block using tablejump_p.
26585 * targhooks.c (default_invalid_within_doloop): Likewise.
26586 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
26587 target hook implementation that is identical to the default hook.
26588 (rs6000_invalid_within_doloop): Remove.
26590 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
26591 unused variable from tablejump_p call.
26593 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
26594 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
26595 (INSN_DELETED_P): Likewise.
26596 (emit_jump_table_data): New prototype.
26597 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
26598 after 4th as unused.
26599 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
26600 * sched-vis.c (print_insn): Likewise.
26601 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
26602 insn for compatibility with back ends that use next_active_insn to
26603 identify jump table data.
26604 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
26605 (remove_insn): Likewise.
26606 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
26608 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
26609 (emit_jump_table_data): New function.
26611 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
26612 basic block, a JUMP_TABLE_DATA never is.
26613 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
26614 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
26615 off from code handling real insns.
26616 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
26617 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
26618 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
26619 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
26620 is not a NONDEBUG_INSN_P.
26621 * ira-costs.c (scan_one_insn): Likewise.
26622 * jump.c (mark_all_labels): Likewise.
26623 (mark_jump_label_1): Likewise.
26624 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
26625 * lra.c (get_insn_freq): Expect all insns reaching here to be in
26627 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
26628 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
26629 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
26630 JUMP_TABLE_DATA_P insns.
26631 (calculate_elim_costs_all_insns): Likewise.
26632 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
26633 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
26634 (delete_output_reload): Code style fixups.
26635 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
26636 insn flags on this non-insn.
26637 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
26638 as scheduling barriers, for pre-change compatibility.
26639 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
26640 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
26642 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
26643 redundant JUMP_TABLE_DATA_P test.
26644 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
26645 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
26646 (frv_for_each_packet): Likewise.
26647 * config/i386/i386.c (min_insn_size): Likewise.
26648 (ix86_avoid_jump_mispredicts): Likewise.
26649 * config/m32r/m32r.c (m32r_is_insn): Likewise.
26650 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
26651 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
26652 (mips16_insn_length): Robustify.
26653 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
26654 (mips16_split_long_branches): Likewise.
26655 * config/pa/pa.c (pa_combine_instructions): Likewise.
26656 * config/rs6000/rs6000.c (get_next_active_insn): Treat
26657 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
26658 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
26659 as contributing to pool range lengths.
26660 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
26661 Remove redundant JUMP_TABLE_DATA_P test.
26662 (sh_loop_align): Likewise.
26663 (split_branches): Likewise.
26664 (sh_insn_length_adjustment): Likewise.
26665 * config/spu/spu.c (get_branch_target): Likewise.
26667 2013-03-29 Jan Hubicka <jh@suse.cz>
26669 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
26670 gcov streaming; stream hot bb threshold to ltrans.
26671 * predict.c (get_hot_bb_threshold): Break out from ....
26672 (maybe_hot_count_p): ... here.
26673 (set_hot_bb_threshold): New function.
26674 * lto-section-in.c (lto_section_name): Add profile.
26675 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
26676 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
26677 and data-streamer.h
26678 (histogram_entry): New structure.
26679 (histogram, histogram_pool): New global vars.
26680 (histogram_hash): New structure.
26681 (histogram_hash::hash): New method.
26682 (histogram_hash::equal): Likewise.
26683 (account_time_size): New function.
26684 (cmp_counts): New function.
26685 (dump_histogram): New function.
26686 (ipa_profile_generate_summary): New function.
26687 (ipa_profile_write_summary): New function.
26688 (ipa_profile_read_summary): New function.
26689 (ipa_profile): Decide on threshold.
26690 (pass_ipa_profile): Add ipa_profile_write_summary and
26691 ipa_profile_read_summary.
26692 * Makefile.in (ipa.o): Update dependencies.
26693 * lto-streamer.h (LTO_section_ipa_profile): New section.
26695 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
26697 * tree.h (VAR_P): New.
26699 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
26702 * doc/invoke.texi ([-fwhole-program]): Fix typo.
26704 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
26706 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
26707 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
26708 (control_flow_insn_p): Likewise.
26709 * cfgrtl.c (duplicate_insn_chain): Likewise.
26710 * final.c (get_attr_length_1): Likewise.
26711 (shorten_branches): Likewise.
26712 (final_scan_insn): Likewise.
26713 * function.c (instantiate_virtual_regs): Likewise.
26714 * gcse.c (insert_insn_end_basic_block): Likewise.
26715 * ira-costs.c (scan_one_insn): Likewise.
26716 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
26717 * lra.c (check_rtl): Likewise.
26718 * reload1.c (elimination_costs_in_insn): Likewise.
26719 * reorg.c (follow_jumps): Likewise.
26721 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
26722 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
26723 (thumb_far_jump_used_p): Likewise.
26724 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
26725 (workaround_speculation): Likewise.
26726 (add_sched_insns_for_speculation): Likewise.
26727 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
26728 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
26729 (frv_for_each_packet): Likewise.
26730 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
26731 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
26732 (final_emit_insn_group_barriers): Likewise.
26733 * config/m32r/m32r.c (m32r_is_insn): Likewise.
26734 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
26735 (mips16_insn_length): Likewise.
26736 * config/pa/pa.c (pa_reorg): Likewise.
26737 (pa_combine_instructions): Likewise.
26738 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
26739 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
26740 (sh_reorg): Likewise.
26741 (split_branches): Likewise.
26742 * config/spu/spu.c (get_branch_target): Likewise.
26744 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
26747 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
26749 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
26750 Fix declaration name.
26752 2013-03-28 Lawrence Crowl <crowl@google.com>
26754 * graphds.h (struct graph.indicies): Remove unused.
26755 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
26756 (SCOP_ORIGINAL_PDDRS): Remove unused.
26757 * sese.h (extern insert_loop_close_phis): Removed unused.
26758 (extern insert_guard_phis): Removed unused.
26759 (extern ivtype_map_elt_info): Removed unused.
26760 (new_ivtype_map_elt): Removed unused.
26761 * sese.c (ivtype_map_elt_info): Removed unused.
26763 2013-03-28 Lawrence Crowl <crowl@google.com>
26765 * Makefile.in: Add several missing include dependences.
26767 (test-dump.o): New. This object is not added to any executable,
26768 but is present for ad-hoc testing.
26770 (debug (const bitmap_head_def &)): New.
26771 (debug (const bitmap_head_def *)): New.
26773 (extern debug (const bitmap_head_def &)): New.
26774 (extern debug (const bitmap_head_def *)): New.
26776 (debug (edge_def &)): New.
26777 (debug (edge_def *)): New.
26779 (debug (basic_block_def &)): New.
26780 (debug (basic_block_def *)): New.
26782 (dump_node (const_tree, int, FILE *)): Correct source file.
26784 (debug (die_struct &)): New.
26785 (debug (die_struct *)): New.
26787 (extern debug (die_struct &)): New.
26788 (extern debug (die_struct *)): New.
26789 * gimple-pretty-print.c
26790 (debug (gimple_statement_d &)): New.
26791 (debug (gimple_statement_d *)): New.
26792 * gimple-pretty-print.h
26793 (extern debug (gimple_statement_d &)): New.
26794 (extern debug (gimple_statement_d *)): New.
26796 (debug (ira_allocno_copy &)): New.
26797 (debug (ira_allocno_copy *)): New.
26798 (debug (ira_allocno &)): New.
26799 (debug (ira_allocno *)): New.
26801 (extern debug (ira_allocno_copy &)): New.
26802 (extern debug (ira_allocno_copy *)): New.
26803 (extern debug (ira_allocno &)): New.
26804 (extern debug (ira_allocno *)): New.
26806 (debug (live_range &)): New.
26807 (debug (live_range *)): New.
26809 (debug (lra_live_range &)): New.
26810 (debug (lra_live_range *)): New.
26812 (debug (lra_live_range &)): New.
26813 (debug (lra_live_range *)): New.
26815 (debug (omega_pb_d &)): New.
26816 (debug (omega_pb_d *)): New.
26818 (extern debug (omega_pb_d &)): New.
26819 (extern debug (omega_pb_d *)): New.
26821 (debug (const rtx_def &)): New.
26822 (debug (const rtx_def *)): New.
26824 (debug_tree (tree): Move within file.
26825 (debug_raw (const tree_node &)): New.
26826 (debug_raw (const tree_node *)): New.
26827 (dump_tree_via_hooks (const tree_node *, int)): New.
26828 (debug (const tree_node &)): New.
26829 (debug (const tree_node *)): New.
26830 (debug_verbose (const tree_node &)): New.
26831 (debug_verbose (const tree_node *)): New.
26832 (debug_head (const tree_node &)): New.
26833 (debug_head (const tree_node *)): New.
26834 (debug_body (const tree_node &)): New.
26835 (debug_body (const tree_node *)): New.
26836 (debug_vec_tree (tree): Move and reimplement in terms of dump.
26837 (debug (vec<tree, va_gc> &)): New.
26838 (debug (vec<tree, va_gc> *)): New.
26840 (extern debug (const rtx_def &)): New.
26841 (extern debug (const rtx_def *)): New.
26843 (debug_raw (simple_bitmap_def &)): New.
26844 (debug_raw (simple_bitmap_def *)): New.
26845 (debug (simple_bitmap_def &)): New.
26846 (debug (simple_bitmap_def *)): New.
26848 (extern debug (simple_bitmap_def &)): New.
26849 (extern debug (simple_bitmap_def *)): New.
26850 (extern debug_raw (simple_bitmap_def &)): New.
26851 (extern debug_raw (simple_bitmap_def *)): New.
26853 (debug (vinsn_def &)): New.
26854 (debug (vinsn_def *)): New.
26855 (debug_verbose (vinsn_def &)): New.
26856 (debug_verbose (vinsn_def *)): New.
26857 (debug (expr_def &)): New.
26858 (debug (expr_def *)): New.
26859 (debug_verbose (expr_def &)): New.
26860 (debug_verbose (expr_def *)): New.
26861 (debug (vec<rtx> &)): New.
26862 (debug (vec<rtx> *)): New.
26864 (extern debug (vinsn_def &)): New.
26865 (extern debug (vinsn_def *)): New.
26866 (extern debug_verbose (vinsn_def &)): New.
26867 (extern debug_verbose (vinsn_def *)): New.
26868 (extern debug (expr_def &)): New.
26869 (extern debug (expr_def *)): New.
26870 (extern debug_verbose (expr_def &)): New.
26871 (extern debug_verbose (expr_def *)): New.
26872 (extern debug (vec<rtx> &)): New.
26873 (extern debug (vec<rtx> *)): New.
26875 (_list_iter_cond_expr): Make inline instead of static.
26877 (debug (sreal &)): New.
26878 (debug (sreal *)): New.
26880 (extern debug (sreal &)): New.
26881 (extern debug (sreal *)): New.
26883 (extern debug_raw (const tree_node &)): New.
26884 (extern debug_raw (const tree_node *)): New.
26885 (extern debug (const tree_node &)): New.
26886 (extern debug (const tree_node *)): New.
26887 (extern debug_verbose (const tree_node &)): New.
26888 (extern debug_verbose (const tree_node *)): New.
26889 (extern debug_head (const tree_node &)): New.
26890 (extern debug_head (const tree_node *)): New.
26891 (extern debug_body (const tree_node &)): New.
26892 (extern debug_body (const tree_node *)): New.
26893 (extern debug (vec<tree, va_gc> &)): New.
26894 (extern debug (vec<tree, va_gc> *)): New.
26896 (debug (struct loop &)): New.
26897 (debug (struct loop *)): New.
26898 (debug_verbose (struct loop &)): New.
26899 (debug_verbose (struct loop *)): New.
26900 * tree-dump.c: Add header dependence.
26902 (extern debug (struct loop &)): New.
26903 (extern debug (struct loop *)): New.
26904 (extern debug_verbose (struct loop &)): New.
26905 (extern debug_verbose (struct loop *)): New.
26907 (debug (data_reference &)): New.
26908 (debug (data_reference *)): New.
26909 (debug (vec<data_reference_p> &)): New.
26910 (debug (vec<data_reference_p> *)): New.
26911 (debug (vec<ddr_p> &)): New.
26912 (debug (vec<ddr_p> *)): New.
26914 (extern debug (data_reference &)): New.
26915 (extern debug (data_reference *)): New.
26916 (extern debug (vec<data_reference_p> &)): New.
26917 (extern debug (vec<data_reference_p> *)): New.
26918 (extern debug (vec<ddr_p> &)): New.
26919 (extern debug (vec<ddr_p> *)): New.
26921 (debug (pt_solution &)): New.
26922 (debug (pt_solution *)): New.
26924 (extern debug (pt_solution &)): New.
26925 (extern debug (pt_solution *)): New.
26927 (debug (_var_map &)): New.
26928 (debug (_var_map *)): New.
26929 (debug (tree_live_info_d &)): New.
26930 (debug (tree_live_info_d *)): New.
26932 (extern debug (_var_map &)): New.
26933 (extern debug (_var_map *)): New.
26934 (extern debug (tree_live_info_d &)): New.
26935 (extern debug (tree_live_info_d *)): New.
26937 2013-03-28 Jan Hubicka <jh@suse.cz>
26939 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
26941 2013-03-28 Ian Bolton <ian.bolton@arm.com>
26943 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
26944 record only when desired or required.
26946 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
26948 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
26949 *vec_extractv2di_1_rex64. Use x64 isa attribute.
26951 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26953 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
26954 (*andsi3_compare0_uxtw): New pattern.
26955 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
26956 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
26958 2013-03-28 Jan Hubicka <jh@suse.cz>
26960 * data-streamer-in.c (streamer_read_gcov_count): New function.
26961 * gimple-streamer-out.c: Include value-prof.h.
26962 (output_gimple_stmt): Output histogram.
26963 (output_bb): Use streamer_write_gcov_count.
26964 * value-prof.c: Include data-streamer.h
26965 (dump_histogram_value): Add HIST_TYPE_MAX.
26966 (stream_out_histogram_value): New function.
26967 (stream_in_histogram_value): New function.
26968 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
26969 (stream_out_histogram_value, stream_in_histogram_value): Declare.
26970 * data-streamer-out.c (streamer_write_gcov_count): New function.
26971 (streamer_write_gcov_count_stream): New function.
26972 * lto-cgraph.c (lto_output_edge): Update counter streaming.
26973 (lto_output_node): Likewise.
26974 (input_node, input_edge): Likewise.
26975 * lto-streamer-out.c (output_cfg): Update streaming.
26976 * lto-streamer-in.c (input_cfg): Likewise.
26977 * data-streamer.h (streamer_write_gcov_count,
26978 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
26979 * gimple-streamer-in.c: Include value-prof.h
26980 (input_gimple_stmt): Input histograms.
26981 (input_bb): Update profile streaming.
26983 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26985 * genmodes.c (emit_max_int): New function.
26986 (emit_insn_modes_h): Added call to emit_max_function.
26987 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
26989 * machmode.def: Fixed comment.
26991 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26993 * combine.c (try_combine): Removed useless assert.
26994 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
26996 2013-03-28 Marek Polacek <polacek@redhat.com>
26997 Richard Biener <rguenther@suse.de>
26999 PR tree-optimization/56695
27000 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
27001 build signed result of a vector comparison.
27002 * tree-cfg.c (verify_gimple_comparison): Check that a result
27003 of a vector comparison has signed type.
27005 2013-03-28 Richard Biener <rguenther@suse.de>
27007 PR tree-optimization/37021
27008 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
27009 do not restrict gaps between groups.
27010 * tree-vect-stmts.c (vectorizable_load): Properly account for
27011 a gap between groups.
27013 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
27015 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
27016 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
27019 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
27021 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
27022 * doc/extend.texi (Named Address Spaces): Ditto.
27023 (Variable Attributes): Ditto.
27025 2013-03-27 Kai Tietz <ktietz@redhat.com>
27027 * config.build: Add support for cygwin x64 target.
27028 * config.gcc: Likewise.
27029 * config.host: Likewise.
27030 * configure.ac: Likewise
27031 * configure: Regenerated.
27033 2013-03-27 Kai Tietz <ktietz@redhat.com>
27035 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
27036 * config/i386/t-cygwin-w64: New file.
27037 * config/i386/cygwin-w64.h: New file.
27038 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
27039 and add support for x64-cygwin target.
27040 (CPP_SPEC): Likewise.
27041 (CXX_WRAP_SPEC_LIST): Undefine before define.
27042 (LIBGCJ_SONAME): Use 15 as version.
27044 2013-03-27 Richard Biener <rguenther@suse.de>
27046 PR tree-optimization/56716
27047 * tree-ssa-structalias.c (perform_var_substitution): Adjust
27048 dumping for ref nodes.
27050 2013-03-27 Martin Jambor <mjambor@suse.cz>
27052 PR tree-optimization/55334
27053 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
27054 restricted pointers to arrays.
27056 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
27058 * Makefile.in (.SUFFIXES): Add .cc.
27059 (.c.o): Apply same recipe for implicit rule .cc.o.
27061 2013-03-27 Richard Biener <rguenther@suse.de>
27063 PR tree-optimization/37021
27064 * tree-vect-data-refs.c (vect_check_strided_load): Allow
27065 REALPART/IMAGPART_EXPRs around the supported refs.
27066 * tree-ssa-structalias.c (find_func_aliases): Assume that
27067 floating-point values are not used to transfer pointers.
27069 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
27071 * target.def (TARGET_HAS_IFUNC_P): New target hook.
27072 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
27073 * doc/tm.texi: Regenerate.
27074 * targhooks.h (default_has_ifunc_p): New.
27075 * targhooks.c (default_has_ifunc_p): Ditto.
27076 * config/linux-protos.h: New file.
27077 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
27078 hook for linux which disables support of indirect functions in android.
27079 * config/linux-android.c: New file.
27080 * config/t-linux-android.c: Ditto.
27081 * config.gcc: Added new object file linux-android.o.
27082 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
27083 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
27084 * varasm.c (do_assemble_alias): Likewise.
27085 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
27086 doesn't support indirect functions.
27087 * configure: Regenerate.
27089 2013-03-27 Bin Cheng <bin.cheng@arm.com>
27092 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
27093 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
27096 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27098 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
27100 2013-03-27 Terry Guo <terry.guo@arm.com>
27102 * config/arm/arm-cores.def: Added core cortex-r7.
27103 * config/arm/arm-tune.md: Regenerated.
27104 * config/arm/arm-tables.opt: Regenerated.
27105 * doc/invoke.texi: Added entry for core cortex-r7.
27107 2013-03-27 Walter Lee <walt@tilera.com>
27109 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
27110 double-decrement of next_scratch_regno.
27112 2013-03-27 Walter Lee <walt@tilera.com>
27114 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
27116 (insn_v1mulus): Ditto.
27117 (insn_v2muls): Ditto.
27119 2013-03-27 Walter Lee <walt@tilera.com>
27121 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
27122 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
27124 2013-03-27 Walter Lee <walt@tilera.com>
27126 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
27127 (*sibcall_value): Ditto.
27129 2013-03-27 Walter Lee <walt@tilera.com>
27131 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
27132 (insn_mnz_v8qi): ... this ...
27133 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
27135 (insn_v<n>mnz): Replaced by ...
27136 (insn_v1mnz): ... this ...
27137 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
27139 (insn_mz_<mode>): Replaced by ...
27140 (insn_mz_v8qi): ... this ...
27141 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
27143 (insn_v<n>mz): Replaced by ...
27144 (insn_v1mz): ... this ...
27145 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
27148 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
27150 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
27152 2013-03-26 Roland McGrath <mcgrathr@google.com>
27154 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
27155 than fprintf with a non-constant, non-format string.
27157 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
27159 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
27160 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
27161 operand 0 predicate.
27162 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
27163 attribute. Use general_x64nomem_operand as operand 1 predicate.
27164 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
27165 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
27166 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
27167 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
27168 (mov<mode>_insv_1): Remove expander. Merge insn with
27169 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
27170 Use general_x64nomem_operand as operand 1 predicate.
27171 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
27172 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
27173 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
27174 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
27175 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
27176 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
27177 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
27178 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
27179 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
27180 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
27181 (general_x64nomem_operand): Ditto.
27183 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
27185 * config/rtems.opt: Add -pthread option.
27187 2013-03-26 Richard Biener <rguenther@suse.de>
27189 * alias.c (find_base_term): Avoid redundant and not used recursion.
27190 (base_alias_check): Get the initial base term from the caller.
27191 (true_dependence_1): Compute and pass base terms to base_alias_check.
27192 (write_dependence_p): Likewise.
27193 (may_alias_p): Likewise.
27195 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
27197 * config/aarch64/aarch64.c (aarch64_classify_address): Support
27198 PC-relative load in SI modes and above only.
27200 2013-03-26 Xinyu Qi <xyqi@marvell.com>
27202 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
27203 * config/arm/iwmmxt.md (WCGR0): Update.
27204 (WCGR1, WCGR2, WCGR3): Likewise.
27206 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
27208 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
27209 Use x64 and nox64 isa attributes.
27211 2013-03-26 Richard Biener <rguenther@suse.de>
27213 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
27214 alignment computations and rely on get_object_alignment_1
27215 for the !TYPE_P case.
27216 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
27218 2013-03-26 Walter Lee <walt@tilera.com>
27220 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
27221 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
27223 2013-03-25 Jeff Law <law@redhat.com>
27225 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
27226 check for INTEGRAL_TYPE_P that was missing due to checking in
27227 wrong version of prior patch.
27229 2013-03-25 Walter Lee <walt@tilera.com>
27231 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
27232 TILEGX_INSN_SHUFFLEBYTES1.
27233 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
27235 (tilegx_builtins): Ditto.
27236 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
27238 2013-03-25 Walter Lee <walt@tilera.com>
27240 * config/tilegx/tilegx.md (floatsisf2): New pattern.
27241 (floatunssisf2): New pattern.
27242 (floatsidf2): New pattern.
27243 (floatunssidf2): New pattern.
27245 2013-03-25 Walter Lee <walt@tilera.com>
27247 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
27248 tests for constraint J, K, N, P.
27250 2013-03-25 Walter Lee <walt@tilera.com>
27252 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
27253 Use indirect/pcrel encoding.
27254 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
27257 2013-03-25 Steve Ellcey <sellcey@mips.com>
27259 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
27260 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
27261 * config/mips/mips.c (mips_option_override): Set IMADD default.
27262 * config/mips/mips.h (PTF_AVOID_IMADD): New.
27263 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
27264 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
27265 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
27266 * doc/invoke.texi (-mimadd/-mno-imadd): New.
27268 2013-03-25 Jeff Law <law@redhat.com>
27270 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
27271 slightly to avoid creating and folding useless trees. Simplify
27272 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
27274 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
27276 * config/i386/i386.md (*zero_extendsidi2): Merge with
27277 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
27278 * config/i386/predicates.md (x86_64_zext_operand): Rename from
27279 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
27280 targets. Clarify comment.
27282 2013-03-25 Martin Jambor <mjambor@suse.cz>
27284 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
27285 pass-through jump functions differently.
27286 (ipa_read_jump_function): Likewise. Also use setter functions to set
27289 2013-03-25 Martin Jambor <mjambor@suse.cz>
27291 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
27292 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
27294 (ipa_get_indirect_edge_target): New function.
27295 (devirtualization_time_bonus): New parameter known_aggs, pass it to
27296 ipa_get_indirect_edge_target. Update all callers.
27297 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
27298 ipa_get_indirect_edge_target_1 instead of calling
27299 ipa_get_indirect_edge_target.
27300 (create_specialized_node): Pass aggvlas to
27301 ipcp_discover_new_direct_edges.
27303 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27305 * config/arm/arm.md (f_sels, f_seld): New types.
27306 (*cmov<mode>): New pattern.
27307 * config/arm/predicates.md (arm_vsel_comparison_operator): New
27310 2013-03-25 Kai Tietz <ktietz@redhat.com>
27312 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
27313 POSIX-printf for mingw-hosted builds.
27315 2013-03-25 Richard Biener <rguenther@suse.de>
27317 PR middle-end/56694
27318 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
27319 must-not-throw stmt location.
27321 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27323 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
27324 Emit load-acquire versions when acq is true.
27325 (arm_emit_store_exclusive): Add rel parameter.
27326 Emit store-release versions when rel is true.
27327 (arm_split_compare_and_swap): Use acquire-release instructions
27329 of barriers when appropriate.
27330 (arm_split_atomic_op): Likewise.
27331 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
27332 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
27333 (VUNSPEC_SLX): Likewise.
27334 (VUNSPEC_LDA): Likewise.
27335 (VUNSPEC_STL): Likewise.
27336 * config/arm/sync.md (atomic_load<mode>): New pattern.
27337 (atomic_store<mode>): Likewise.
27338 (arm_load_acquire_exclusive<mode>): Likewise.
27339 (arm_load_acquire_exclusivesi): Likewise.
27340 (arm_load_acquire_exclusivedi): Likewise.
27341 (arm_store_release_exclusive<mode>): Likewise.
27343 2013-03-25 Catherine Moore <clm@codesourcery.com>
27345 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
27346 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
27347 * config/mip/predicates.md (lwsp_swsp_operand,
27348 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
27349 sb16_operand, db4_operand, db7_operand, ib3_operand,
27350 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
27351 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
27352 andi16_operand): New predicates.
27353 * config/mips/mips.md (compression): New attribute.
27354 (enabled): New attribute.
27355 (length): Consider compression in computing length.
27356 (shift_compression): New code attribute.
27357 (*add<mode>3): New operands. Record compression.
27358 (sub<mode>3): Likewise.
27359 (one_cmpl<mode>2): Likewise.
27360 (*and<mode>3): Likewise.
27361 (*ior<mode>3): Likewise.
27362 (unnamed pattern for xor): Likewise.
27363 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
27364 (*<optab><mode>3): Likewise.
27365 (*mov<mode>_internal: Likewise.
27366 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
27367 (mips_unsigned_immediate_p): New.
27368 (umips_lwsp_swsp_address_p): New.
27369 (m16_based_address_p): New.
27370 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
27371 (mips_unsigned_immediate_p): New prototype.
27372 (lwsp_swsp_address_p): New prototype.
27373 (m16_based_address_p): New prototype.
27374 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
27375 (mips_signed_immediate_p): New function.
27376 (m16_based_address_p): New function.
27377 (lwsp_swsp_address_p): New function.
27378 (mips_print_operand_punctuation): Recognize short delay slot insns
27379 for microMIPS.add<mode>3"
27381 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27384 * config/arm/iterators.md (v_cmp_result): New mode attribute.
27385 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
27387 2013-03-25 Richard Biener <rguenther@suse.de>
27389 PR tree-optimization/56689
27390 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
27393 2013-03-25 Richard Biener <rguenther@suse.de>
27395 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
27397 (memory_references): Likewise.
27398 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
27399 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
27400 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
27401 (gather_mem_refs_in_loops): Fold into ...
27402 (analyze_memory_references): ... this. Move initialization
27403 to tree_ssa_lim_initialize.
27404 (fill_always_executed_in): Rename to ...
27405 (fill_always_executed_in_1): ... this.
27406 (fill_always_executed_in): Move contains_call computation to
27407 this new function from ...
27408 (tree_ssa_lim_initialize): ... here.
27409 (tree_ssa_lim): Call fill_always_executed_in.
27411 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
27413 * postreload.c (reload_combine): Fix code detecting returns.
27415 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
27417 * function.c (emit_use_return_register_into_block): On cc0 targets,
27418 do not emit the sequence between cc0 setter and user.
27420 2013-03-25 Kai Tietz <ktietz@redhat.com>
27422 * config/i386/predicates.md (local_symbolic_operand): Interpret
27423 dll-imported symbols as none-local.
27425 2013-03-25 Richard Biener <rguenther@suse.de>
27427 * tree-ssa-loop-im.c (struct depend): Remove.
27428 (struct lim_aux_data): Make depends a vec of gimples.
27429 (free_lim_aux_data): Adjust.
27430 (add_dependency): Likewise.
27431 (set_level): Likewise.
27433 2013-03-25 Richard Biener <rguenther@suse.de>
27435 PR middle-end/56434
27436 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
27437 the pointer returned by calls with ECF_MALLOC set.
27439 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
27441 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
27443 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
27445 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
27446 using MMXMODE mode iterator.
27447 (*move<mode>_internal): Merge with *movv2sf_internal and
27448 *movv2sf_internal_rex64 using MMXMODE mode iterator.
27450 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
27452 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
27453 (record_last_mem_set_info): Likewise.
27455 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
27456 of XNEWVEC followed by memset.
27457 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
27459 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
27461 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
27462 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
27463 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
27464 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
27465 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
27466 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
27467 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
27468 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
27469 BARRIER_P instead of GET_CODE.
27471 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27473 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
27474 inaccuracy in the probing code.
27476 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
27477 (ctrapdi4): Likewise.
27479 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27481 * calls.c (expand_call): Add missing guard to code handling return
27482 of non-BLKmode structures in MSB.
27483 * function.c (expand_function_end): Likewise.
27485 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27487 * combine.c (try_combine): Adjust comment. Do not add the set of
27488 insn #0 if the destination indirectly is set or dies in insn #2.
27489 Tidy up code to distribute a new note.
27491 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
27493 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
27494 also for alternatives 16 and 17.
27496 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
27498 * config/i386/sse.md (*mov<mode>_internal): Merge with
27499 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
27500 Emit insn template depending on type attribute. Use
27501 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
27502 movd instead of movq mnemonic for interunit moves. Rewrite mode
27503 attribute calculation. Remove unit attribute calculation.
27504 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
27505 Set prefix_data16 attribute for DImode ssemov types.
27506 Use Ym instead of y for SSE-MMX conversion alternatives.
27507 Reorder operand constraints.
27509 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
27511 * df.h (df_insn_delete): Adjust prototype.
27512 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
27513 and let it decide whether mark the basic block dirty.
27514 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
27515 * df-scan.c (df_insn_info_delete): New helper function, split
27516 off from df_insn_delete.
27517 (df_scan_free_bb_info): Use it.
27518 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
27520 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
27521 that the insn is actually an insn and it has a non-NULL basic block.
27522 Do not mark basic block dirty if only deleting a DEBUG_INSN.
27524 2013-03-22 Richard Biener <rguenther@suse.de>
27526 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
27528 (mem_ref_alloc): Do not allocate them.
27529 (refs_independent_p): Do not query or maintain a cache.
27531 2013-03-22 Richard Biener <rguenther@suse.de>
27533 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
27534 (gather_mem_refs_in_loops): Do not compute it.
27535 (analyze_memory_references): Do not allocate it.
27536 (tree_ssa_lim_finalize): Do not free it.
27537 (for_all_locs_in_loop): Do not query all_refs_in_loop.
27539 2013-03-22 Richard Biener <rguenther@suse.de>
27541 * is-a.h (as_a): Use gcc_checking_assert.
27543 2013-03-22 Ian Bolton <ian.bolton@arm.com>
27545 * config/aarch64/aarch64.c (aarch64_print_operand): New
27546 format specifier for printing a constant in hex.
27547 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
27548 format specifier for printing second operand.
27550 2013-03-22 Richard Biener <rguenther@suse.de>
27552 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
27554 (gather_mem_refs_in_loops): Perform store accumulation here.
27555 (create_vop_ref_mapping_loop): Remove.
27556 (create_vop_ref_mapping): Likewise.
27557 (analyze_memory_references): Initialize refs_stored_in_loop.
27558 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
27559 (record_indep_loop): Remove.
27560 (record_dep_loop): New function.
27561 (ref_indep_loop_p_1): Adjust to only walk over references
27562 in the loop, not its subloops.
27563 (ref_indep_loop_p): Rename to ...
27564 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
27565 maintaining a more fine-grained cache.
27566 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
27567 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
27569 2013-03-22 Richard Biener <rguenther@suse.de>
27571 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
27572 (struct mem_ref): Make accesses_in_loop a vec of a vec of
27573 aggregate mem_ref_loc.
27574 (free_mem_ref_locs): Inline into ...
27575 (memref_free): ... this and adjust.
27576 (mem_ref_alloc): Adjust.
27577 (mem_ref_locs_alloc): Remove.
27578 (record_mem_ref_loc): Adjust.
27579 (get_all_locs_in_loop): Rewrite into ...
27580 (for_all_locs_in_loop): ... this iterator.
27581 (rewrite_mem_ref_loc): New functor.
27582 (rewrite_mem_refs): Use for_all_locs_in_loop.
27583 (sm_set_flag_if_changed): New functor.
27584 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
27585 (ref_always_accessed): New functor.
27586 (ref_always_accessed_p): Use for_all_locs_in_loop.
27588 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27590 * tree-pass.h (PROP_gimple_lvec): New.
27591 * passes.c (dump_properties): Handle PROP_gimple_lvec.
27592 (init_optimization_passes): Move pass_lower_vector.
27593 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
27595 (pass_lower_vector): Provide PROP_gimple_lvec.
27596 (pass_lower_vector_ssa): Likewise.
27597 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
27599 2013-03-21 Mark Wielaard <mjw@redhat.com>
27601 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
27603 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
27605 * config/i386/i386.md (*movdi_internal): Disparage slightly
27606 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
27607 conversion alternatives.
27609 2013-03-21 Jakub Jelinek <jakub@redhat.com>
27611 PR middle-end/48087
27612 * diagnostic.def (DK_WERROR): New kind.
27613 * diagnostic.h (werrorcount): Define.
27614 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
27615 promoted to DK_ERROR, increment DK_WERROR counter instead of
27617 * toplev.c (toplev_main): Call print_ignored_options even if
27618 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
27619 even if just werrorcount is non-zero.
27622 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
27624 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
27625 (string_cst_pool_decl): New function.
27626 (optimize_one_addr_into_implicit_ptr): New function.
27627 (resolve_addr_in_expr): Optimize DWARF location expression
27628 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
27629 which doesn't live in memory, but has DW_AT_location or
27630 DW_AT_const_value, or refers to a string literal, into
27631 DW_OP_GNU_implicit_pointer.
27632 (optimize_location_into_implicit_ptr): New function.
27633 (resolve_addr): If removing DW_AT_location of a variable because
27634 it was DW_OP_addr of address of the variable, but the variable doesn't
27635 live in memory, try to emit const value attribute for the initializer.
27637 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27639 * tree.h (VECTOR_TYPE_P): New macro.
27640 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
27641 TYPE_MODE): Use it.
27642 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
27643 VEC_COND_EXPR cannot be lvalues.
27644 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
27646 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27648 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
27649 Restrict the transformation to equal modes.
27651 2013-03-21 Richard Biener <rguenther@suse.de>
27653 PR tree-optimization/39326
27654 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
27655 (MEM_ANALYZABLE): Adjust.
27656 (record_mem_ref_loc): Move bitmap ops ...
27657 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
27658 unanalyzable refs, do not record locations for it.
27659 (analyze_memory_references): Allocate ref zero as shared
27661 (refs_independent_p): Do not test for unanalyzed mems here.
27662 (ref_indep_loop_p_1): Special-case disambiguation against
27663 the unanalyzed ref.
27664 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
27666 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
27668 * config/arm/arm-protos.h (tune_params): Add
27669 prefer_neon_for_64bits field.
27670 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
27671 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
27672 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
27673 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
27674 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
27675 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
27676 (arm_option_override): Handle -mneon-for-64bits new option.
27677 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
27678 (prefer_neon_for_64bits): Declare new variable.
27679 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
27680 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
27681 (arch_enabled): Handle new arch types. Remove support for onlya8
27683 (one_cmpldi2): Use new arch names.
27684 (zero_extend<mode>di2, extend<mode>di2): Ditto.
27685 * config/arm/arm.opt (mneon-for-64bits): Add option.
27686 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
27687 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
27688 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
27690 * doc/invoke.texi (-mneon-for-64bits): Document.
27692 2013-03-21 Richard Biener <rguenther@suse.de>
27694 PR tree-optimization/39326
27695 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
27696 (sort_bbs_in_loop_postorder_cmp): New function.
27697 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
27699 2013-03-21 Richard Biener <rguenther@suse.de>
27701 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
27702 (vect_insert_into_interleaving_chain): Likewise.
27703 (vect_drs_dependent_in_basic_block): Inline ...
27704 (vect_slp_analyze_data_ref_dependence): ... here. New function,
27706 (vect_analyze_data_ref_dependence): ... here. Simplify.
27707 (vect_check_interleaving): Simplify.
27708 (vect_analyze_data_ref_dependences): Likewise. Split out ...
27709 (vect_slp_analyze_data_ref_dependences): ... this new function.
27710 (dr_group_sort_cmp): New function.
27711 (vect_analyze_data_ref_accesses): Compute data-reference groups
27712 here instead of in vect_analyze_data_ref_dependence. Use
27713 a more efficient algorithm.
27714 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
27715 vect_slp_analyze_data_ref_dependences. Call
27716 vect_analyze_data_ref_accesses earlier.
27717 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
27718 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
27719 (vect_slp_analyze_data_ref_dependences): New prototype.
27721 2013-03-21 Richard Biener <rguenther@suse.de>
27723 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
27724 ref is stored in the loop.
27725 (find_refs_for_sm): Walk only over all stores.
27726 (store_motion_loop): Allocate from lim_bitmap_obstack.
27727 (store_motion): Likewise.
27729 2013-03-21 Richard Biener <rguenther@suse.de>
27731 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
27732 Update virtual SSA form.
27734 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
27736 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
27737 * configure: Regenerate.
27738 * config.in: Regenerate.
27739 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
27740 if !HAVE_LD_EH_FRAME_CIEV3.
27742 2013-03-21 Richard Biener <rguenther@suse.de>
27744 * tree-cfg.c (verify_expr_no_block): New function.
27745 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
27746 nor DECL_VALUE_EXPR have locations with associated blocks.
27747 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
27748 (clear_unused_block_pointer): Remove code dealing with
27749 blocks in DECL_DEBUG_EXPR locations.
27751 2013-03-21 Richard Biener <rguenther@suse.de>
27753 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
27754 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
27755 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
27756 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
27757 instead of DECL_DEBUG_EXPR_IS_FROM.
27758 * gimplify.c (gimplify_modify_expr): Likewise.
27759 * tree-cfg.c (verify_expr_location_1): Likewise.
27760 * tree-complex.c (create_one_component_var): Likewise.
27761 * tree-sra.c (create_access_replacement): Likewise.
27762 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
27763 (clear_unused_block_pointer): Likewise.
27764 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
27765 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
27766 * var-tracking.c (var_debug_decl): Likewise.
27767 (track_expr_p): Likewise.
27768 * tree-inline.c (add_local_variables): Likewise. Set
27769 DECL_HAS_DEBUG_EXPR_P after copying it.
27770 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
27771 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
27773 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
27776 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
27777 * configure: Regenerate.
27778 * config.in: Regenerate.
27779 * config/i386/i386.md (*movdf_internal): Use
27780 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
27781 movd instead of movq mnemonic for interunit moves.
27782 (*movdi_internal): Ditto.
27784 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27786 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
27787 (abd<mode>_3): New pattern.
27788 (aba<mode>_3): New pattern.
27789 (fabd<mode>_3): New pattern.
27791 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27793 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
27794 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
27795 occurrence of REGISTER_PREFIX as its empty string.
27797 2013-03-20 Jeff Law <law@redhat.com>
27799 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
27800 addititional equivalences for equality comparisons between an SSA_NAME
27801 and a constant where the SSA_NAME was set from a widening conversion.
27803 2013-03-20 Walter Lee <walt@tilera.com>
27805 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
27807 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27809 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
27810 depending on type attribute.
27811 (*movti_internal): Ditto.
27812 (*movtf_internal): Ditto.
27813 (*movxf_internal): Ditto.
27814 (*movdf_internal): Ditto.
27815 (*movsf_internal): Ditto.
27817 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27819 * config/i386/i386.md (*movti_internal): Set prefix attribute to
27820 maybe_vex for sselog1 and ssemov types.
27821 (*movdi_internal): Reorder operand constraints.
27822 (*movsi_internal): Ditto. Set prefix attribute to
27823 maybe_vex for sselog1 and ssemov types.
27824 (*movtf_internal): Set prefix attribute to maybe_vex
27825 for sselog1 and ssemov types.
27826 (*movdf_internal): Ditto. Set prefix_data16 attribute for
27827 DImode ssemov types. Reorder operand constraints.
27828 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
27829 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
27830 attribute for SImode ssemov types. Reorder operand constraints.
27832 2013-03-20 Martin Jambor <mjambor@suse.cz>
27834 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
27835 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
27837 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
27839 * config/rs6000/predicates.md (indexed_address, update_address_mem
27840 update_indexed_address_mem): New predicates.
27841 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
27842 attribute for load/store instructions.
27843 * config/rs6000/dfp.md (movsd_store): Likewise.
27844 (movsd_load): Likewise.
27845 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
27846 (unnamed HI->DI extend define_insn): Likewise.
27847 (unnamed SI->DI extend define_insn): Likewise.
27848 (unnamed QI->SI extend define_insn): Likewise.
27849 (unnamed QI->HI extend define_insn): Likewise.
27850 (unnamed HI->SI extend define_insn): Likewise.
27851 (unnamed HI->SI extend define_insn): Likewise.
27852 (extendsfdf2_fpr): Likewise.
27853 (movsi_internal1): Likewise.
27854 (movsi_internal1_single): Likewise.
27855 (movhi_internal): Likewise.
27856 (movqi_internal): Likewise.
27857 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
27858 attribute for load/store instructions.
27859 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
27861 (mov<mode>_softfloat): Likewise.
27862 (mov<mode>_hardfloat32): Likewise.
27863 (mov<mode>_hardfloat64): Likewise.
27864 (mov<mode>_softfloat64): Likewise.
27865 (movdi_internal32): Likewise.
27866 (movdi_internal64): Likewise.
27867 (probe_stack_<mode>): Likewise.
27869 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
27871 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
27872 floating point, and decimal floating point to reload iterator.
27874 * config/rs6000/constraints.md (wl constraint): New constraints to
27875 return FLOAT_REGS if certain options are used to reduce the number
27876 of separate patterns that exist in the file.
27877 (wx constraint): Likewise.
27878 (wz constraint): Likewise.
27880 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27881 -mdebug=reg, print wg, wl, wx, and wz constraints.
27882 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27883 Initialize the reload functions for 64-bit binary/decimal floating
27885 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27886 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27887 create the buffer on the stack to overcome not having a 32-bit
27889 (rs6000_emit_move): Likewise.
27890 (rs6000_secondary_memory_needed_rtx): Likewise.
27891 (rs6000_alloc_sdmode_stack_slot): Likewise.
27892 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27893 via xxlxor, just like DFmode 0.0.
27895 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
27896 define as 1 if we are running on a power7 or newer.
27897 (enum r6000_reg_class_enum): Add new constraints.
27899 * config/rs6000/dfp.md (movsd): Delete, combine with binary
27900 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27901 with other moves by using conditional constraits (wg). Use LFIWZX
27902 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
27903 (movsd splitter): Likewise.
27904 (movsd_hardfloat): Likewise.
27905 (movsd_softfloat): Likewise.
27907 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
27908 binary and decimal floating point moves.
27909 (fmove_ok): New attributes to combine binary and decimal floating
27910 point moves, and to combine power6x (mfpgpr) moves along normal
27912 (real_value_to_target): Likewise.
27913 (f32_lr): Likewise.
27914 (f32_lm): Likewise.
27915 (f32_li): Likewise.
27916 (f32_sr): Likewise.
27917 (f32_sm): Likewise.
27918 (f32_si): Likewise.
27919 (movsf): Combine binary and decimal floating point moves. Combine
27920 power6x (mfpgpr) moves with other moves by using conditional
27921 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
27922 (mov<mode> for SFmode/SDmode); Likewise.
27923 (SFmode/SDmode splitters): Likewise.
27924 (movsf_hardfloat): Likewise.
27925 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
27926 (movsf_softfloat): Likewise.
27927 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
27929 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
27930 wx and wz constraints.
27932 * config/rs6000/constraints.md (wg constraint): New constraint to
27933 return FLOAT_REGS if -mmfpgpr (power6x) was used.
27935 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
27938 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27939 -mdebug=reg, print wg, wl, wx, and wz constraints.
27940 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27941 Initialize the reload functions for 64-bit binary/decimal floating
27943 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27944 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27945 create the buffer on the stack to overcome not having a 32-bit
27947 (rs6000_emit_move): Likewise.
27948 (rs6000_secondary_memory_needed_rtx): Likewise.
27949 (rs6000_alloc_sdmode_stack_slot): Likewise.
27950 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27951 via xxlxor, just like DFmode 0.0.
27953 * config/rs6000/dfp.md (movdd): Delete, combine with binary
27954 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27955 with other moves by using conditional constraits (wg). Use LFIWZX
27956 and STFIWX for loading SDmode on power7.
27957 (movdd splitters): Likewise.
27958 (movdd_hardfloat32): Likewise.
27959 (movdd_softfloat32): Likewise.
27960 (movdd_hardfloat64_mfpgpr): Likewise.
27961 (movdd_hardfloat64): Likewise.
27962 (movdd_softfloat64): Likewise.
27964 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
27965 64-bit binary and decimal floating point moves.
27966 (FMOVE64X): Likewise.
27967 (movdf): Combine 64-bit binary and decimal floating point moves.
27968 Combine power6x (mfpgpr) moves with other moves by using
27969 conditional constraits (wg).
27970 (mov<mode> for DFmode/DDmode): Likewise.
27971 (DFmode/DDmode splitters): Likewise.
27972 (movdf_hardfloat32): Likewise.
27973 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
27974 (movdf_softfloat32): Likewise.
27975 (movdf_hardfloat64_mfpgpr): Likewise.
27976 (movdf_hardfloat64): Likewise.
27977 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
27978 (movdf_softfloat64): Likewise.
27979 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
27980 (reload_<mode>_load): Move to later in the file so they aren't in
27981 the middle of the floating point move insns.
27982 (reload_<mode>_store): Likewise.
27984 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
27987 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
27988 constraint if -mdebug=reg.
27989 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
27990 Enable using dd reload support if needed.
27992 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
27993 binary and decimal floating point moves in rs6000.md.
27994 (movtd_internal): Likewise.
27996 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
27997 decimal floating point moves.
27999 (movtf_internal): Likewise.
28000 (mov<mode>_internal, TDmode/TFmode): Likewise.
28001 (movtf_softfloat): Likewise.
28002 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
28004 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
28005 movdi_internal64, using wg constraint for move direct operations.
28006 (movdi_internal64): Likewise.
28008 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
28009 MODES_TIEABLE_P for selected modes. Print the numerical value of
28010 the various virtual registers. Use GPR/FPR first/last values,
28011 instead of hard coding the register numbers. Print which modes
28012 have reload functions registered.
28013 (rs6000_option_override_internal): If -mdebug=reg, trace the options
28014 settings before/after setting cpu, target and subtarget settings.
28015 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
28016 and for secondary reload failures in rs6000_secondary_reload_inner.
28017 (rs6000_secondary_reload_fail): Likewise.
28018 (rs6000_secondary_reload_inner): Likewise.
28020 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
28021 macros for first/last GPR and FPR registers.
28022 (LAST_GPR_REGNO): Likewise.
28023 (FIRST_FPR_REGNO): Likewise.
28024 (LAST_FPR_REGNO): Likewise.
28026 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
28027 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
28028 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
28029 (vcond<mode><mode>): Likewise.
28030 (vcondu<mode><mode>): Likewise.
28031 (vector_gtu<mode>): Likewise.
28032 (vector_gte<mode>): Likewise.
28033 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
28034 to prevent the compiler from converting DImode operations to TImode.
28035 (ior<mode>3): Likewise.
28036 (and<mode>3): Likewise.
28037 (one_cmpl<mode>2): Likewise.
28038 (nor<mode>3): Likewise.
28039 (andc<mode>3): Likewise.
28041 * config/rs6000/constraints.md (wt constraint): New constraint
28042 that returns VSX_REGS if TImode is allowed in VSX registers.
28044 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
28045 constant under VSX.
28047 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
28048 similar to TImode, but it is restricted to being in the GPRs.
28050 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
28051 TImode to occupy a single VSX register.
28053 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
28054 -mvsx-timode for power7/power8.
28055 (power7 cpu): Likewise.
28056 (power8 cpu): Likewise.
28058 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
28059 sure that TFmode/TDmode take up two registers if they are ever
28060 allowed in the upper VSX registers.
28061 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
28063 (rs6000_init_hard_regno_mode_ok): Likewise.
28064 (rs6000_debug_reg_global): Add debugging for PTImode and wt
28065 constraint. Print if LRA is turned on.
28066 (rs6000_option_override_internal): Give an error if -mvsx-timode
28067 and VSX is not enabled.
28068 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
28069 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
28070 to reg+offset addressing. Use PTImode when checking offset
28071 addresses for validity.
28072 (reg_offset_addressing_ok_p): Likewise.
28073 (rs6000_legitimate_offset_address_p): Likewise.
28074 (rs6000_legitimize_address): Likewise.
28075 (rs6000_legitimize_reload_address): Likewise.
28076 (rs6000_legitimate_address_p): Likewise.
28077 (rs6000_eliminate_indexed_memrefs): Likewise.
28078 (rs6000_emit_move): Likewise.
28079 (rs6000_secondary_reload): Likewise.
28080 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
28081 reloads to fpr registers to continue to use reg+offset addressing,
28082 but 64-bit reloads to altivec registers need reg+reg addressing.
28083 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
28084 it. Treat LO_SUM like a PLUS operation.
28085 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
28086 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
28087 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
28088 registers to share a register with a smaller sized type, since VSX
28089 puts scalars in the upper 64-bits.
28090 (print_operand): Add support for PTImode.
28091 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
28092 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
28093 registers, but don't have arithmetic support.
28094 (rs6000_memory_move_cost): Add test for VSX.
28095 (rs6000_opt_masks): Add -mvsx-timode.
28097 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
28100 (VSr): Use wt constraint for TImode.
28101 (VSv): Drop TImode support.
28102 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
28103 (vsx_movti_64bit): Likewise.
28104 (vsx_movti_32bit): Likewise.
28105 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
28106 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
28107 one '?' on the appropriate output constraint. Do not allow TImode
28108 logical operations on 32-bit systems.
28109 (vsx_ior<mode>3): Likewise.
28110 (vsx_xor<mode>3): Likewise.
28111 (vsx_one_cmpl<mode>2): Likewise.
28112 (vsx_nor<mode>3): Likewise.
28113 (vsx_andc<mode>3): Likewise.
28114 (vsx_concat_<mode>): Likewise.
28115 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
28117 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
28118 OPTION_MASK_VSX_TIMODE.
28119 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
28120 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
28122 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
28123 (TI2 iterator): New iterator for TImode, PTImode.
28124 (wd mode attribute): Add values for vector types.
28125 (movti_string): Replace TI move operations with operations for TImode
28126 and PTImode. Add support for TImode being allowed in VSX registers.
28127 (mov<mode>_string, TImode/PTImode): Likewise.
28128 (movti_ppc64): Likewise.
28129 (mov<mode>_ppc64, TImode/PTImode): Likewise.
28130 (TI mode splitters): Likewise.
28132 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
28135 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
28137 PR tree-optimization/56355
28138 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
28139 Also handle integers with undefined overflow.
28141 2013-03-20 Catherine Moore <clm@codesourcery.com>
28142 Maciej W. Rozycki <macro@codesourcery.com>
28143 Tom de Vries <tom@codesourcery.com>
28144 Nathan Sidwell <nathan@codesourcery.com>
28145 Iain Sandoe <iain@codesourcery.com>
28146 Nathan Froyd <froydnj@codesourcery.com>
28147 Chao-ying Fu <fu@mips.com>
28149 * doc/extend.texi: (micromips, nomicromips, nocompression):
28150 Document new function attributes.
28151 * doc/invoke.texi (minterlink-compressed, mmicromips,
28152 m14k, m14ke, m14kec): Document new options.
28153 (minterlink-mips16): Update documentation.
28154 * doc/md.texi (ZC, ZD): Document new constraints.
28155 * configure.ac (gcc_cv_as_micromips): Check if linker
28156 supports the .set micromips directive.
28157 * configure: Regenerate.
28158 * config.in: Regenerate.
28159 * config/mips/mips-tables.opt: Regenerate.
28160 * config/mips/micromips.md: New file.
28161 * constraints.md (ZC, ZD): New constraints.
28162 * config/mips/predicates.md (movep_src_register): New predicate.
28163 (movep_src_operand): New predicate.
28164 (non_volatile_mem_operand): New predicate.
28165 * config/mips/mips.md (multimem): New type.
28166 (length): Differentiate between 17-bit and 18-bit branch offsets.
28167 (MOVEP1, MOVEP2): New mode iterator.
28168 (mov_<load>l): Use ZC constraint.
28169 (mov_<load>r): Likewise.
28170 (mov_<store>l): Likewise.
28171 (mov_<store>r): Likewise.
28172 (*branch_equality<mode>_inverted): Add microMIPS support.
28173 (*branch_equality<mode>): Likewise.
28174 (*jump_absolute): Likewise.
28175 (indirect_jump_<mode>): Likewise.
28176 (tablejump_<mode>): Likewise.
28177 (<optab>_internal): Likewise.
28178 (sibcall_internal): Likewise.
28179 (sibcall_value_internal): Likewise.
28180 (prefetch): Use constraint ZD.
28181 * config/mips/mips.opt (minterlink-compressed): New option.
28182 (minterlink-mips16): Now an alias for minterlink-compressed.
28183 (mmicromips): New option.
28184 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
28185 (compare_and_swap_12): Likewise.
28186 (sync_add<mode>): Likewise.
28187 (sync_<optab>_12): Likewise.
28188 (sync_old_<optab>_12): Likewise.
28189 (sync_new_<optab>_12): Likewise.
28190 (sync_nand_12): Likewise.
28191 (sync_old_nand_12): Likewise.
28192 (sync_new_nand_12): Likewise.
28193 (sync_sub<mode>): Likewise.
28194 (sync_old_add<mode>): Likewise.
28195 (sync_old_sub<mode>): Likewise.
28196 (sync_new_add<mode>): Likewise.
28197 (sync_new_sub<mode>): Likewise.
28198 (sync_<optab><mode>): Likewise.
28199 (sync_old_<optab><mode>): Likewise.
28200 (sync_new_<optab><mode>): Likewise.
28201 (sync_nand<mode>): Likewise.
28202 (sync_old_nand<mode>): Likewise.
28203 (sync_new_nand<mode>): Likewise.
28204 (sync_lock_test_and_set<mode>): Likewise.
28205 (test_and_set_12): Likewise.
28206 (atomic_compare_and_swap<mode>): Likewise.
28207 (atomic_exchange<mode>_llsc): Likewise.
28208 (atomic_fetch_add<mode>_llsc): Likewise.
28209 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
28210 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
28211 (umips_save_restore_pattern_p): Likewise.
28212 (umips_load_store_pair_p): Likewise.
28213 (umips_output_load_store_pair): Likewise.
28214 (umips_movep_target_p): Likewise.
28215 (umips_12bit_offset_address_p): Likewise.
28216 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
28217 (mips_base_mips16): Rename this...
28218 (mips_base_compression_flags): ...to this. Update all uses.
28219 (mips_attribute_table): Add micromips, nomicromips and nocompression.
28220 (mips_mips16_decl_p): Delete.
28221 (mips_nomips16_decl_p): Delete.
28222 (mips_get_compress_on_flags): New function.
28223 (mips_get_compress_off_flags): New function.
28224 (mips_get_compress_mode): New function.
28225 (mips_get_compress_on_name): New function.
28226 (mips_get_compress_off_name): New function.
28227 (mips_insert_attributes): Support multiple compression types.
28228 (mips_merge_decl_attributes): Likewise.
28229 (umips_12bit_offset_address_p): New function.
28230 (mips_start_function_definition): Emit .set micromips directive.
28231 (mips_call_may_need_jalx_p): New function.
28232 (mips_function_ok_for_sibcall): Add microMIPS support.
28233 (mips_print_operand_punctuation): Support short delay slots and
28235 (umips_swm_mask, umips_swm_encoding): New.
28236 (umips_build_save_restore): New function.
28237 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
28238 (was_mips16_p): Remove.
28239 (old_compression_mode): New.
28240 (mips_set_compression_mode): New function.
28241 (mips_set_current_function): Add microMIPS support.
28242 (mips_option_override): Likewise.
28243 (umips_save_restore_pattern_p): New function.
28244 (umips_output_save_restore): New function.
28245 (umips_load_store_pair_p_1): New function.
28246 (umips_load_store_pair_p): New function.
28247 (umips_output_load_store_pair_1): New function.
28248 (umips_output_load_store_pair): New function.
28249 (umips_movep_target_p) New function.
28250 (mips_prepare_pch_save): Add microMIPS support.
28251 * config/mips/mips.h (TARGET_COMPRESSION): New.
28252 (TARGET_CPU_CPP_BUILTINS): Update macro
28253 to use new compression flags and to support microMIPS.
28254 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
28255 (MIPS_ARCH_FLOAT_SPEC): Likewise.
28256 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
28257 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
28258 (ASM_SPEC): Support mmicromips and mno-micromips.
28259 (M16STORE_REG_P): New macro.
28260 (MIPS_CALL): Support TARGET_MICROMIPS.
28261 (MICROMIPS_J): New macro.
28262 (mips_base_mips16): Rename this...
28263 (mips_base_compression_flags): ...to this.
28264 (UMIPS_12BIT_OFFSET_P): New macro.
28265 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
28266 (MULTILIB_DIRNAMES): Likewise.
28267 2013-03-20 Richard Biener <rguenther@suse.de>
28269 PR tree-optimization/56661
28270 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
28271 the result does not have to be distinct.
28273 2013-03-20 Richard Biener <rguenther@suse.de>
28275 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
28278 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
28279 Steven Bosscher <steven@gcc.gnu.org>
28281 PR rtl-optimization/56605
28282 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
28284 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
28287 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
28288 that require movd instead of movq.
28290 2013-03-20 Richard Biener <rguenther@suse.de>
28292 * tree-ssa-structalias.c (struct variable_info): Add pointer
28293 to the first field of an aggregate with sub-vars. Make
28294 this and the pointer to the next subfield its ID.
28295 (vi_next): New function.
28296 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
28297 storedanything_id, integer_id): Increment by one.
28298 (new_var_info, get_call_vi, lookup_call_clobber_vi,
28299 get_call_clobber_vi): Adjust.
28300 (solution_set_expand): Simplify and speedup.
28301 (solution_set_add): Inline into ...
28302 (set_union_with_increment): ... this. Adjust accordingly.
28303 (do_sd_constraint): Likewise.
28304 (do_ds_constraint): Likewise.
28305 (do_complex_constraint): Simplify.
28306 (build_pred_graph): Adjust.
28307 (solve_graph): Likewise. Simplify and speedup.
28308 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
28309 get_constraint_for_component_ref, get_constraint_for_1,
28310 first_vi_for_offset, first_or_preceding_vi_for_offset,
28311 create_function_info_for, create_variable_info_for_1,
28312 create_variable_info_for, intra_create_variable_infos): Adjust.
28313 (init_base_vars): Push NULL for ID zero.
28314 (compute_points_to_sets): Adjust.
28316 2013-03-20 Richard Biener <rguenther@suse.de>
28318 * cfgloop.c (verify_loop_structure): Streamline and avoid
28319 ICEing on corrupt loop tree.
28320 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
28323 2013-03-20 Richard Biener <rguenther@suse.de>
28325 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
28326 check whether an SSA update is needed.
28328 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
28330 * config/mips/constraints.md (T): Rename to...
28334 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
28335 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
28337 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28339 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
28340 (*subsi3_carryin_uxtw): Likewise.
28342 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28344 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
28345 (*rorsi3_insn_uxtw): Likewise.
28347 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28349 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
28350 (*extrsi5_insn_uxtw): Likewise.
28352 2013-03-19 Richard Biener <rguenther@suse.de>
28354 PR tree-optimization/56273
28355 * passes.c (init_optimization_passes): Move second VRP after DOM.
28357 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
28359 * config/i386/i386.md (*movti_internal): Merge from
28360 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
28361 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
28362 nox64 isa attributes.
28364 2013-03-18 Richard Biener <rguenther@suse.de>
28366 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
28368 (merge_node_constraints): Likewise.
28369 (build_succ_graph): Likewise.
28370 (valid_graph_edge): Inline into single caller.
28371 (unify_nodes): Likewise. Use bitmap_set_bit return value
28373 (scc_visit): Fix formatting and variable use.
28374 (do_sd_constraint): Use gcc_checking_assert.
28375 (do_ds_constraint): Likewise.
28376 (do_complex_constraint): Likewise.
28377 (condense_visit): Likewise. Cleanup.
28378 (dump_pred_graph): New function.
28379 (perform_var_substitution): Dump the pred-graph before
28380 variable substitution.
28381 (find_equivalent_node): Use gcc_checking_assert.
28382 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
28384 2013-03-18 Richard Biener <rguenther@suse.de>
28386 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
28387 Remove cond_expr_stmt_list argument and do not gimplify the
28389 (vect_loop_versioning): Adjust.
28390 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
28391 Cleanup to use less temporaries.
28392 (vect_create_data_ref_ptr): Cleanup.
28394 2013-03-18 Jakub Jelinek <jakub@redhat.com>
28396 PR tree-optimization/56635
28397 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
28398 require types_compatible_p types.
28400 2013-03-18 Nick Clifton <nickc@redhat.com>
28402 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
28403 spurious backslash.
28405 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
28406 Add missing line to comment describing function.
28408 2013-03-18 Richard Biener <rguenther@suse.de>
28410 PR tree-optimization/56210
28411 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
28412 Handle string / character search functions.
28413 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
28415 2013-03-18 Richard Biener <rguenther@suse.de>
28417 PR middle-end/56483
28418 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
28419 and implement properly.
28420 * gimple.h (gimple_cond_single_var_p): Remove.
28422 2013-03-18 Richard Biener <rguenther@suse.de>
28424 * tree-data-ref.h (find_data_references_in_loop): Declare.
28425 * tree-data-ref.c (get_references_in_stmt): Use a stack
28426 vector pre-allocated in the callers.
28427 (find_data_references_in_stmt): Adjust.
28428 (graphite_find_data_references_in_stmt): Likewise.
28429 (create_rdg_vertices): Likewise.
28430 (find_data_references_in_loop): Export.
28431 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
28432 Compute dependences here...
28433 (vect_analyze_data_refs): ...not here. When we encounter
28434 a non-vectorizable data reference in basic-block vectorization
28435 truncate the data reference vector. Do not bother to
28436 fixup data-dependence information for gather loads.
28437 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
28438 of data references, as reported.
28440 2013-03-18 Richard Biener <rguenther@suse.de>
28442 PR tree-optimization/3713
28443 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
28444 has_constants and expr.
28445 (stmt_has_constants): Properly valueize SSA names when deciding
28446 whether the stmt has constants.
28448 2013-03-18 Richard Biener <rguenther@suse.de>
28450 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
28451 whole function when there is nothing to do.
28452 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
28453 * tree-vectorizer.c (vectorize_loops): Update virtual and
28454 loop-closed SSA once.
28455 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
28457 2013-03-18 Richard Biener <rguenther@suse.de>
28459 PR middle-end/56113
28460 * domwalk.c (bb_postorder): New global static.
28461 (cmp_bb_postorder): New function.
28462 (walk_dominator_tree): Replace scheme imposing an order for
28463 visiting dominator sons by one sorting them at the time they
28464 are pushed on the stack.
28466 2013-03-18 Richard Biener <rguenther@suse.de>
28468 PR tree-optimization/39326
28469 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
28470 (struct mem_ref): Replace mem member with ao_ref typed member.
28471 (MEM_ANALYZABLE): Adjust.
28472 (memref_eq): Likewise.
28473 (mem_ref_alloc): Likewise.
28474 (gather_mem_refs_stmt): Likewise.
28475 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
28476 (execute_sm_if_changed_flag_set): Adjust.
28477 (execute_sm): Likewise.
28478 (ref_always_accessed_p): Likewise.
28479 (refs_independent_p): Likewise.
28480 (can_sm_ref_p): Likewise.
28482 2013-03-18 Jakub Jelinek <jakub@redhat.com>
28485 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
28486 return 1 even for !unsignedp.
28488 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
28490 * config/i386/i386.md (isa): Add x64 and nox64.
28491 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
28492 (*pushtf): Enable *roF alternative for x64 isa only.
28493 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
28494 mode attribute of integer alternatives to DImode for TARGET_64BIT.
28495 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
28496 (*movtf_internal): Merge from *movtf_internal_rex64 and
28497 *movtf_internal_sse. Use x64 and nox64 isa attributes.
28498 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
28499 nox64 isa attributes.
28500 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
28501 nox64 isa attributes.
28502 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
28504 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
28506 * config/alpha/alpha.c (TARGET_LRA_P): New define.
28508 2013-03-17 Jakub Jelinek <jakub@redhat.com>
28511 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
28512 class names. Remove trailing comma after "ALL_REGS".
28514 2013-03-16 Jan Hubicka <jh@suse.cz>
28516 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
28517 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
28518 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
28519 of cgraph_get_create_node.
28520 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
28522 2013-03-16 Jason Merrill <jason@redhat.com>
28525 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
28526 with DW_AT_default_value.
28528 2013-03-16 Jakub Jelinek <jakub@redhat.com>
28530 * BASE-VER: Set to 4.9.0.
28532 2013-03-14 Andi Kleen <ak@linux.intel.com>
28535 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
28536 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
28537 Document _x* TSX intrinsics.
28539 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
28540 David Holsgrove <david.holsgrove@xilinx.com>
28542 * configure.ac: Add MicroBlaze TLS support detection.
28543 * configure: Regenerate.
28544 * config/microblaze/microblaze-protos.h
28545 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
28546 symbol_mentioned_p, label_mentioned_p): Add prototypes.
28547 * config/microblaze/microblaze.c (microblaze_address_type): Add
28548 ADDRESS_TLS and tls_reloc address types.
28549 (microblaze_address_info): Add tls_reloc.
28550 (TARGET_HAVE_TLS): Define.
28551 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
28552 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
28553 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
28554 load_tls_operand, microblaze_call_tls_get_addr,
28555 microblaze_legitimize_tls_address): New functions.
28556 (microblaze_classify_unspec): Handle UNSPEC_TLS.
28557 (get_base_reg): Use microblaze_tls_symbol_p.
28558 (microblaze_classify_address): Handle TLS.
28559 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
28560 label_mentioned_p and microblaze_tls_referenced_p.
28561 (microblaze_legitimize_address): Handle TLS.
28562 (microblaze_address_insns): Handle ADDRESS_TLS.
28563 (pic_address_needs_scratch): Handle TLS.
28564 (print_operand_address): Handle TLS.
28565 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
28566 (microblaze_expand_move): Handle TLS.
28567 (microblaze_legitimate_constant_p): Check
28568 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
28569 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
28570 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
28571 (PIC_OFFSET_TABLE_REGNUM): Set.
28572 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
28573 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
28574 (addsi3, movsi_internal2, movdf_internal): Update constraints
28575 * config/microblaze/predicates.md (arith_plus_operand): Define
28576 (move_operand): Redefine as move_src_operand,
28577 check microblaze_tls_referenced_p.
28579 2013-03-14 Ian Bolton <ian.bolton@arm.com>
28581 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
28582 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
28584 2013-03-14 Ian Bolton <ian.bolton@arm.com>
28586 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
28589 2013-03-14 Jakub Jelinek <jakub@redhat.com>
28591 PR tree-optimization/53265
28592 * common.opt (Waggressive-loop-optimizations): New option.
28593 * tree-ssa-loop-niter.c: Include tree-pass.h.
28594 (do_warn_aggressive_loop_optimizations): New function.
28595 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
28596 if number_of_latch_executions returned constant.
28597 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
28598 early. If number_of_latch_executions returned constant, set
28599 nb_iterations_upper_bound back to it.
28600 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
28602 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
28603 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
28605 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
28606 (MULTILIB_OSDIRNAMES): Set.
28607 * genmultilib: If defaultosdirname doesn't start with :: , set
28608 defaultosdirname2 instead, clear it and emit two . multilib_raw
28609 entries instead of just one.
28611 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
28613 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
28614 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
28615 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
28616 (SUBTARGET_OVERRIDE_OPTIONS): New.
28618 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
28621 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
28622 (musermode): Convert to Var(TARGET_USERMODE).
28623 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
28624 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
28625 * config/sh/sh.c (sh_option_override): Use
28626 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
28627 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
28629 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
28631 (udivsi3_i4_single, divsi3_i4_single): Use
28632 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
28634 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
28636 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
28639 2013-03-13 Richard Biener <rguenther@suse.de>
28641 PR tree-optimization/56608
28642 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
28643 calls when vectorizing basic-blocks.
28645 2013-03-13 Jakub Jelinek <jakub@redhat.com>
28648 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
28651 2013-03-12 Jakub Jelinek <jakub@redhat.com>
28653 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
28655 2013-03-11 Jan Hubicka <jh@suse.cz>
28658 * lto-streamer-out.c (output_symbol_p): Skip references from
28659 constructors of external variables.
28661 2013-03-11 Jan Hubicka <jh@suse.cz>
28663 PR middle-end/56571
28664 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
28666 * emit-rtl.c (verify_rtx_sharing): Likewise.
28667 (copy_insn_1): Likewise.
28668 * rtl.c (copy_rtx): Likewise.
28670 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
28673 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
28674 output_operand_lossage message.
28676 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
28679 * arm.c (shift_op): Validate RTL pattern on the fly.
28680 (arm_print_operand, case 'S'): Don't use shift_operator to validate
28683 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28686 * config/pa/pa.md (call_value): Check for calls to powf and direct to
28687 new call patterns that clobber %fr12.
28688 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
28689 split and postreload patterns.
28690 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
28691 registers %fr12 and %fr12R as call used.
28693 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
28695 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
28696 (canon_address, record_store, replace_read, check_mem_read_rtx,
28697 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
28698 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
28699 rest_of_handle_dse): Likewise.
28701 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
28703 PR middle-end/56524
28704 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
28706 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
28707 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
28708 (save_optabs_if_changed): Replace with...
28709 (init_tree_optimization_optabs): ...this.
28710 * optabs.c (save_optabs_if_changed): Rename to...
28711 (init_tree_optimization_optabs): ...this. Take the optimization node
28712 as argument. Do nothing if the base optabs are already correct.
28713 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
28714 to recompute optabs.
28715 * function.h (function): Remove optabs field.
28716 * function.c (invoke_set_current_function_hook): Call
28717 init_tree_optimization_optabs. Use the result to initialize
28720 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
28722 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
28723 if GTMA_HAS_NO_INSTRUMENTATION.
28724 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
28725 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
28726 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
28727 * gimple-pretty-print.c (dump_gimple_transaction): Handle
28728 GTMA_HAS_NO_INSTRUMENTATION.
28730 2013-03-08 Jakub Jelinek <jakub@redhat.com>
28732 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
28735 2013-03-08 Marek Polacek <polacek@redhat.com>
28736 Jakub Jelinek <jakub@redhat.com>
28738 PR tree-optimization/56478
28739 * predict.c (is_comparison_with_loop_invariant_p): Change the
28740 type of loop_step to tree.
28741 (predict_loops): Adjust.
28742 (predict_iv_comparison): Perform the computations on double_ints.
28744 2013-03-08 Richard Biener <rguenther@suse.de>
28746 PR tree-optimization/56570
28747 * tree-cfg.c (verify_expr_location_1): Verify locations for
28749 * tree-sra.c (create_access_replacement): Strip locations
28750 from DECL_DEBUG_EXPRs.
28752 2013-03-08 Richard Biener <rguenther@suse.de>
28754 * tree-inline.c (expand_call_inline): Do not associate
28755 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
28756 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
28758 2013-03-08 Richard Biener <rguenther@suse.de>
28760 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
28761 or block changes with -Og. Fix for location / block encoding
28762 changes and PHI arguments with locations.
28764 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
28766 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
28768 (struct output_info): Likewise.
28769 (register_overhead): Remove bad gcc_assert.
28770 (bitmap_find_bit): If there is only a single bitmap element, do not
28771 count a miss as a search.
28772 (print_statistics): Update for counter type changes.
28773 (dump_bitmap_statistics): Likewise. Print headers such that they
28774 are properly lined up with the printed counters.
28776 2013-03-07 Jakub Jelinek <jakub@redhat.com>
28778 PR tree-optimization/56559
28779 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
28780 check that it has only a single use.
28782 2013-03-07 Richard Biener <rguenther@suse.de>
28784 * doc/invoke.texi (fwhole-program): Discourage use in combination
28787 2013-03-06 Jakub Jelinek <jakub@redhat.com>
28789 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
28791 PR tree-optimization/56539
28792 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
28793 instead of GSI_CONTINUE_LINKING as last argument to
28794 force_gimple_operand_gsi. Adjust function comment.
28796 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
28799 PR middle-end/56548
28800 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
28801 promoted mode, convert the result back to the original mode.
28803 2013-03-06 Richard Biener <rguenther@suse.de>
28805 PR middle-end/56294
28806 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
28807 (insert_updated_phi_nodes_compare_uids): New function.
28808 (update_ssa): Sort symbols_to_rename after UID before
28809 traversing it to insert PHI nodes.
28811 2013-03-06 Richard Biener <rguenther@suse.de>
28813 PR middle-end/50494
28814 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
28815 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
28818 2013-02-13 Richard Biener <rguenther@suse.de>
28821 * varasm.c (output_constant_def_1): Get the decl representing
28822 the constant as argument.
28823 (output_constant_def): Wrap output_constant_def_1.
28824 (make_decl_rtl): Use output_constant_def_1 with the decl
28825 representing the constant.
28826 (build_constant_desc): Optionally re-use a decl already
28827 representing the constant.
28828 (tree_output_constant_def): Adjust.
28830 2013-03-06 Joey Ye <joey.ye@arm.com>
28833 * gcc.c (convert_white_space): New function.
28834 (main): Handles white space in function name.
28836 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
28839 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
28840 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
28841 to SH_DIV_CALL_TABLE for TARGET_SH2.
28842 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
28844 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
28845 call-table options.
28847 2013-03-05 Sterling Augustine <saugustine@google.com>
28848 Cary Coutant <ccoutant@google.com>
28851 * dwarf2out.c (resolve_addr): Don't call
28852 remove_loc_list_addr_table_entries a second time for the same
28855 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28858 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
28859 (avoid_complex_debug_insns): New function.
28860 (expand_debug_locations): Call it.
28862 PR rtl-optimization/56484
28863 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
28864 lifetimes of hard registers on small register class machines.
28866 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28868 * config/microblaze/microblaze-protos.h: Rename
28869 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
28870 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
28872 (microblaze_fast_interrupt_function_p): New function.
28873 (microblaze_is_interrupt_handler): Rename to
28874 microblaze_is_interrupt_variant and add fast_interrupt check.
28875 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
28876 (save_restore_insns): Likewise.
28877 (compute_frame_size): Likewise.
28878 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
28879 (microblaze_globalize_label): Likewise.
28880 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
28881 * config/microblaze/microblaze.md: Use wrapper
28882 microblaze_is_interrupt_variant.
28884 2013-03-05 Kai Tietz <ktietz@redhat.com>
28886 * sdbout.c (sdbout_one_type): Switch to current function's section
28887 supporting cold/hot.
28889 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28891 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
28894 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28896 PR middle-end/56461
28897 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
28898 if VALGRIND_GET_VBITS is defined, temporarily make object
28899 memory all defined, and restore previous valgrind addressability
28900 and definability afterwards. Free this_object at the end.
28902 PR middle-end/56461
28903 * lra.c (lra): Call lra_clear_live_ranges if live_p,
28904 right before calling lra_create_live_ranges, also call it
28905 when clearing live_p. Only call lra_clear_live_ranges
28906 at the end if live_p.
28908 PR middle-end/56461
28909 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
28911 2013-03-05 Richard Biener <rguenther@suse.de>
28913 PR tree-optimization/56521
28914 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
28917 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
28920 * except.h (remove_unreachable_eh_regions): New prototype.
28921 * except.c (remove_eh_handler_splicer): New function, split out
28922 of remove_eh_handler.
28923 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
28924 warning about running it on many EH regions one at a time.
28925 (remove_unreachable_eh_regions_worker): New function, walk the
28926 EH tree in depth-first order and remove non-marked regions.
28927 (remove_unreachable_eh_regions): New function.
28928 * tree-eh.c (mark_reachable_handlers): New function, split out
28929 from remove_unreachable_handlers.
28930 (remove_unreachable_handlers): Use mark_reachable_handlers and
28931 remove_unreachable_eh_regions.
28932 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
28933 and remove_unreachable_eh_regions.
28935 2013-03-05 Richard Biener <rguenther@suse.de>
28937 PR middle-end/56525
28938 * loop-init.c (fix_loop_structure): Remove loops in two stages,
28939 not freeing them until the end.
28941 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28943 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
28945 2013-03-05 Richard Biener <rguenther@suse.de>
28947 PR tree-optimization/56270
28948 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
28949 of loads after scheduling an SLP instance.
28951 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28953 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
28955 (check_gcc_parallelize): Run guality.exp as a separate job from
28956 vect.exp with unsorted.exp and $(dg_target_exps) separately from
28957 struct-layout-1.exp with stackalign.exp.
28959 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
28961 PR middle-end/56461
28962 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
28963 load_index sbitmap even if some bit in it isn't set.
28965 PR middle-end/56461
28966 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
28967 (discover_iteration_bound_by_body_walk): Change queues to
28968 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
28969 spelling in comment. Call safe_push on queues[bound_index] directly.
28970 Release queues[queue_index] in every iteration unconditionally.
28971 Release bounds vector.
28973 PR middle-end/56461
28974 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
28975 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
28976 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
28979 2013-03-05 Richard Biener <rguenther@suse.de>
28982 * tree-inline.c (remap_blocks_to_null): New function.
28983 (expand_call_inline): When expanding a call stmt without
28984 an associated block inline remap all callee blocks to NULL.
28986 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28988 PR rtl-optimization/56494
28989 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
28990 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
28991 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
28993 PR middle-end/56461
28994 * sel-sched-ir.c (free_sched_pools): Release
28995 succs_info_pool.stack[succs_info_pool.max_top] vectors too
28996 if succs_info_pool.max_top isn't -1.
28999 * opts.c (opts_obstack, opts_concat): Moved to...
29000 * opts-common.c (opts_obstack, opts_concat): ... here.
29002 2013-03-04 Jakub Jelinek <jakub@redhat.com>
29004 PR middle-end/56461
29005 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
29007 2013-03-04 Martin Jambor <mjambor@suse.cz>
29009 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
29010 all appropriate places.
29012 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
29014 PR tree-optimization/56424
29015 * ipa-split.c (split_function): Do not set the RSO flag if result is
29016 not by reference and its type is a register type.
29018 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
29020 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
29021 (microblaze_legitimate_pic_operand): Likewise
29022 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
29023 new function microblaze_legitimate_pic_operand
29024 * config/microblaze/microblaze-protos.h
29025 (microblaze_legitimate_pic_operand): Declare.
29027 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29029 * config/microblaze/predicates.md (call_insn_simple_operand):
29030 New predicate for supported rtx code types.
29031 * config/microblaze/microblaze.md (call_internal1): Use
29032 call_insn_simple_operand predicate.
29034 2013-03-04 Jakub Jelinek <jakub@redhat.com>
29036 PR middle-end/56461
29037 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
29038 partitions.ordered_remove.
29040 PR middle-end/56461
29041 * tree-vect-stmts.c (vectorizable_conversion): Don't call
29042 vec_oprnds0.create (1) for modifier == NONE.
29044 PR middle-end/56461
29045 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
29046 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
29047 vec_oprnds1 right before pushing anything to it for
29050 PR middle-end/56461
29051 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
29052 set nbbs to 0 instead of having separate code path.
29053 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
29054 instead of false as last argument if returning NULL.
29056 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
29058 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
29059 the attribute is now called "target" instead of "option".
29060 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
29061 * doc/tm.texi.in (Target Attributes): Likewise document the correct
29062 attribute/pragma name for TARGET_OPTION_VALID_P and
29063 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
29064 * doc/tm.texi: Regenerated.
29066 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
29068 * config/microblaze/microblaze.c:
29069 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
29070 * config/microblaze/microblaze.h: Add -mxl-reorder to
29072 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
29073 instructions emitted if TARGET_REORDER.
29074 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
29075 or 0 for -m/-mno case, but initialises as 2 to detect default use case
29078 2013-03-01 Xinliang David Li <davidxl@google.com>
29080 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
29083 2013-03-01 Jakub Jelinek <jakub@redhat.com>
29085 PR middle-end/56461
29086 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
29087 vector even when returning true. Fix up function comment formatting.
29089 PR middle-end/56461
29090 * ira-build.c (ira_loop_nodes_count): New variable.
29091 (create_loop_tree_nodes): Initialize it.
29092 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
29094 PR middle-end/56461
29095 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
29096 method on dr_chain and result_chain.
29097 * tree-vect-stmts.c (vectorizable_store): Only call
29098 result_chain.create if j == 0.
29100 PR middle-end/56461
29101 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
29102 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
29103 before overwriting it.
29105 2013-03-01 Tobias Burnus <burnus@net-b.de>
29107 * doc/extended.texi (C Extensions): Change order in @menu
29109 (Other MIPS Built-in Functions): Move last MIPS entry before
29110 "picoChip Built-in Functions".
29111 (SH Built-in Functions): Move after RX Built-in Functions.
29112 * doc/gcc.texi (Introduction): Change order in @menu
29114 * doc/md.texi (Constraints): Ditto.
29115 * gty.texi (Type Information): Ditto.
29116 (User-provided marking routines for template types): Make
29118 * doc/invoke.texi (AArch64 Options): Move before
29119 "Adapteva Epiphany Options".
29121 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
29122 Jakub Jelinek <jakub@redhat.com>
29125 * asan.c (gate_asan): Lookup no_sanitize_address instead of
29126 no_address_safety_analysis attribute.
29127 * doc/extend.texi (no_address_safety_attribute): Rename to
29128 no_sanitize_address attribute, mention no_address_safety_analysis
29129 attribute as deprecated alias.
29131 2013-02-28 Jakub Jelinek <jakub@redhat.com>
29133 PR middle-end/56461
29134 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
29135 type to vec<vec<tree> > *.
29136 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
29137 to be vec<tree> instead of vec<tree> *, set vec_defs
29138 to vNULL and call vec_defs.create (number_of_vects), adjust other
29140 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
29141 vectorizable_condition): Adjust vect_get_slp_defs callers.
29143 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
29145 * config/aarch64/aarch64.c
29146 (aarch64_float_const_representable): Remove unused variable.
29148 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
29150 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
29152 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
29154 * config/aarch64/aarch64-builtins.c
29155 (aarch64_init_simd_builtins): Make static.
29157 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
29159 * config/aarch64/aarch64.c
29160 (aarch64_simd_make_constant): Make static.
29162 2013-02-28 Martin Jambor <mjambor@suse.cz>
29164 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
29165 with no initialization to the RHS of debug statements.
29167 2013-02-28 Martin Jambor <mjambor@suse.cz>
29169 PR tree-optimization/56294
29170 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
29172 (get_access_replacement): Do not call create_access_replacement.
29173 Assert a replacement exists.
29174 (get_repl_default_def_ssa_name): Create the replacement declaration
29177 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29179 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
29180 final_end_function.
29182 2013-02-28 Marek Polacek <polacek@redhat.com>
29184 PR rtl-optimization/56466
29185 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
29186 if we're changing a loop.
29187 (peel_loops_completely): Likewise.
29189 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
29192 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
29194 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
29197 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
29198 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
29199 INTX_FTYPE_FX, FX_FTYPE_INTX.
29200 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
29202 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
29204 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
29205 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
29206 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
29207 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
29208 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
29209 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
29210 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
29211 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
29212 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
29213 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
29214 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
29215 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
29216 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
29217 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
29218 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
29219 (avrxmega6): Increase max flash segments from 5 to 6.
29220 * config/avr/t-multilib: Regenerate.
29221 * config/avr/avr-tables.opt: Regenerate.
29222 * doc/avr-mmcu.texi: Regenerate.
29224 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
29226 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
29227 (avr_device_to_arch): Rename to avr_device_to_ld.
29228 (avr_device_to_as): New prototype.
29229 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
29230 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
29231 * config/avr/driver-avr.c (avr_device_to_as): New.
29232 (avr_device_to_arch): Rename to avr_device_to_ld.
29234 2013-02-27 Jakub Jelinek <jakub@redhat.com>
29236 PR middle-end/56461
29237 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
29238 method on dr_chain and result_chain.
29240 PR middle-end/56461
29241 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
29242 pointer_set_destroy on not_executed_last_iteration.
29244 PR middle-end/56461
29245 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
29247 PR middle-end/56461
29248 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
29249 FOR_EACH_DEFINED_FUNCTION when freeing state.
29251 PR middle-end/56461
29252 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
29254 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
29257 PR middle-end/56461
29258 * ipa-cp.c (decide_whether_version_node): Call vec_free on
29259 known_aggs[i].items and release known_aggs vector.
29261 PR middle-end/56461
29262 * ipa-reference.c (propagate): Free node_info even for alias nodes.
29264 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29266 * config/microblaze/microblaze.c (microblaze_emit_compare):
29267 Use xor for EQ/NE comparisions.
29268 * config/microblaze/microblaze.md (cstoresf4): Add constraints
29269 (cbranchsf4): Adjust operator to comparison_operator.
29271 2013-02-27 Jakub Jelinek <jakub@redhat.com>
29273 PR middle-end/56461
29274 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
29276 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
29277 vec_safe_push, always update *slot.
29278 (redirect_edge_var_map_clear): Use vec_free.
29279 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
29280 (free_var_map_entry): Use vec_free.
29281 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
29282 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
29284 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
29286 PR middle-end/45472
29287 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
29288 when the may_trap_p bit of the exprs being merged differs.
29289 Reorder tests for speculativeness in the logical and operator.
29291 2013-02-27 Jakub Jelinek <jakub@redhat.com>
29293 * incpath.c (add_standard_paths): Use reconcat instead of concat
29294 where appropriate and avoid leaking memory.
29296 * opts.h: Include obstack.h.
29297 (opts_concat): New prototype.
29298 (opts_obstack): New declaration.
29299 * opts.c (opts_concat): New function.
29300 (opts_obstack): New variable.
29301 (init_options_struct): Call gcc_init_obstack on opts_obstack.
29302 (finish_options): Use opts_concat instead of concat
29303 and XOBNEWVEC instead of XNEWVEC.
29304 * opts-common.c (generate_canonical_option, decode_cmdline_option,
29305 generate_option): Likewise.
29306 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
29307 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
29310 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
29311 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
29313 2013-02-26 Jakub Jelinek <jakub@redhat.com>
29315 PR middle-end/56461
29316 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
29318 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
29320 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
29321 (arm_block_move_unaligned_straight): Likewise.
29322 (arm_adjust_block_mem): Likewise.
29324 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
29327 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
29328 temp, cond and label.
29329 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
29332 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
29333 * config/c6x/c6x.h (dbx_register_map): Update declaration.
29336 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
29337 of prologue/epilogue functions.
29340 * config/tilegx/tilegx.c (tilegx_expand_prologue):
29341 Remove unused variable cfa_offset.
29342 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
29345 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
29346 type promotion to unsigned.
29349 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
29350 for HOST_WIDE_INT of 32 bit / same size as int.
29351 (arm_block_move_unaligned_straight): Likewise.
29352 (arm_adjust_block_mem): Likewise.
29355 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
29358 2013-02-26 Marek Polacek <polacek@redhat.com>
29360 PR tree-optimization/56426
29361 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
29363 2013-02-26 Richard Biener <rguenther@suse.de>
29366 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
29367 unused variable loops.
29369 2013-02-26 Jakub Jelinek <jakub@redhat.com>
29371 PR tree-optimization/56448
29372 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
29373 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
29374 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
29375 later operands of the references, or even first operand for
29376 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
29378 PR tree-optimization/56443
29379 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
29380 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
29381 to type_for_mode langhook.
29383 2013-02-25 Matt Turner <mattst88@gmail.com>
29385 * doc/invoke.texi: Document r4700.
29387 2013-02-25 Richard Biener <rguenther@suse.de>
29389 PR tree-optimization/56175
29390 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
29392 (simplify_bitwise_binary): ... here. Also guard the conversion
29393 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
29395 2013-02-25 Catherine Moore <clm@codesourcery.com>
29398 2013-02-24 Catherine Moore <clm@codesourcery.com>
29399 Maciej W. Rozycki <macro@codesourcery.com>
29400 Tom de Vries <tom@codesourcery.com>
29401 Nathan Sidwell <nathan@codesourcery.com>
29402 Iain Sandoe <iain@codesourcery.com>
29403 Nathan Froyd <froydnj@codesourcery.com>
29404 Chao-ying Fu <fu@mips.com>
29406 * doc/extend.texi: (micromips, nomicromips, nocompression):
29407 Document new function attributes.
29408 * doc/invoke.texi (minterlink-compressed, mmicromips,
29409 m14k, m14ke, m14kec): Document new options.
29410 (minterlink-mips16): Update documentation.
29411 * doc/md.texi (ZC, ZD): Document new constraints.
29412 * configure.ac (gcc_cv_as_micromips): Check if linker
29413 supports the .set micromips directive.
29414 * configure: Regenerate.
29415 * config.in: Regenerate.
29416 * config/mips/mips-tables.opt: Regenerate.
29417 * config/mips/micromips.md: New file.
29418 * constraints.md (ZC, AD): New constraints.
29419 * config/mips/predicates.md (movep_src_register): New predicate.
29420 (movep_src_operand): New predicate.
29421 (non_volatile_mem_operand): New predicate.
29422 * config/mips/mips.md (multimem): New type.
29423 (length): Differentiate between 17-bit and 18-bit branch offsets.
29424 (MOVEP1, MOVEP2): New mode iterator.
29425 (mov_<load>l): Use ZC constraint.
29426 (mov_<load>r): Likewise.
29427 (mov_<store>l): Likewise.
29428 (mov_<store>r): Likewise.
29429 (*branch_equality<mode>_inverted): Add microMIPS support.
29430 (*branch_equality<mode>): Likewise.
29431 (*jump_absolute): Likewise.
29432 (indirect_jump_<mode>): Likewise.
29433 (tablejump_<mode>): Likewise.
29434 (<optab>_internal): Likewise.
29435 (sibcall_internal): Likewise.
29436 (sibcall_value_internal): Likewise.
29437 (prefetch): Use constraint ZD.
29438 * config/mips/mips.opt (minterlink-compressed): New option.
29439 (minterlink-mips16): Now an alias for minterlink-compressed.
29440 (mmicromips): New option.
29441 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
29442 (compare_and_swap_12): Likewise.
29443 (sync_add<mode>): Likewise.
29444 (sync_<optab>_12): Likewise.
29445 (sync_old_<optab>_12): Likewise.
29446 (sync_new_<optab>_12): Likewise.
29447 (sync_nand_12): Likewise.
29448 (sync_old_nand_12): Likewise.
29449 (sync_new_nand_12): Likewise.
29450 (sync_sub<mode>): Likewise.
29451 (sync_old_add<mode>): Likewise.
29452 (sync_old_sub<mode>): Likewise.
29453 (sync_new_add<mode>): Likewise.
29454 (sync_new_sub<mode>): Likewise.
29455 (sync_<optab><mode>): Likewise.
29456 (sync_old_<optab><mode>): Likewise.
29457 (sync_new_<optab><mode>): Likewise.
29458 (sync_nand<mode>): Likewise.
29459 (sync_old_nand<mode>): Likewise.
29460 (sync_new_nand<mode>): Likewise.
29461 (sync_lock_test_and_set<mode>): Likewise.
29462 (test_and_set_12): Likewise.
29463 (atomic_compare_and_swap<mode>): Likewise.
29464 (atomic_exchange<mode>_llsc): Likewise.
29465 (atomic_fetch_add<mode>_llsc): Likewise.
29466 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
29467 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
29468 (umips_save_restore_pattern_p): Likewise.
29469 (umips_load_store_pair_p): Likewise.
29470 (umips_output_load_store_pair): Likewise.
29471 (umips_movep_target_p): Likewise.
29472 (umips_12bit_offset_address_p): Likewise.
29473 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
29474 (mips_base_mips16): Rename this...
29475 (mips_base_compression_flags): ...to this. Update all uses.
29476 (mips_attribute_table): Add micromips, nomicromips and nocompression.
29477 (mips_mips16_decl_p): Delete.
29478 (mips_nomips16_decl_p): Delete.
29479 (mips_get_compress_on_flags): New function.
29480 (mips_get_compress_off_flags): New function.
29481 (mips_get_compress_mode): New function.
29482 (mips_get_compress_on_name): New function.
29483 (mips_get_compress_off_name): New function.
29484 (mips_insert_attributes): Support multiple compression types.
29485 (mips_merge_decl_attributes): Likewise.
29486 (umips_12bit_offset_address_p): New function.
29487 (mips_start_function_definition): Emit .set micromips directive.
29488 (mips_call_may_need_jalx_p): New function.
29489 (mips_function_ok_for_sibcall): Add microMIPS support.
29490 (mips_print_operand_punctuation): Support short delay slots and
29492 (umips_swm_mask, umips_swm_encoding): New.
29493 (umips_build_save_restore): New function.
29494 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
29495 (was_mips16_p): Remove.
29496 (old_compression_mode): New.
29497 (mips_set_compression_mode): New function.
29498 (mips_set_current_function): Add microMIPS support.
29499 (mips_option_override): Likewise.
29500 (umips_save_restore_pattern_p): New function.
29501 (umips_output_save_restore): New function.
29502 (umips_load_store_pair_p_1): New function.
29503 (umips_load_store_pair_p): New function.
29504 (umips_output_load_store_pair_1): New function.
29505 (umips_output_load_store_pair): New function.
29506 (umips_movep_target_p) New function.
29507 (mips_prepare_pch_save): Add microMIPS support.
29508 * config/mips/mips.h (TARGET_COMPRESSION): New.
29509 (TARGET_CPU_CPP_BUILTINS): Update macro
29510 to use new compression flags and to support microMIPS.
29511 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
29512 (MIPS_ARCH_FLOAT_SPEC): Likewise.
29513 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
29514 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
29515 (ASM_SPEC): Support mmicromips and mno-micromips.
29516 (M16STORE_REG_P): New macro.
29517 (MIPS_CALL): Support TARGET_MICROMIPS.
29518 (MICROMIPS_J): New macro.
29519 (mips_base_mips16): Rename this...
29520 (mips_base_compression_flags): ...to this.
29521 (UMIPS_12BIT_OFFSET_P): New macro.
29522 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
29523 (MULTILIB_DIRNAMES): Likewise.
29525 2013-02-25 Tom de Vries <tom@codesourcery.com>
29527 PR rtl-optimization/56131
29528 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
29529 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
29530 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
29532 2013-02-25 Tobias Burnus <burnus@net-b.de>
29534 * doc/invoke.texi (-fsanitize=): Move from optimization
29535 to debugging options.
29537 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
29539 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
29541 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
29542 Alexander Monakov <amonakov@ispras.ru>
29544 PR middle-end/56077
29545 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
29546 flush pending lists also on non-jumps. Adjust comment.
29548 2013-02-24 Catherine Moore <clm@codesourcery.com>
29549 Maciej W. Rozycki <macro@codesourcery.com>
29550 Tom de Vries <tom@codesourcery.com>
29551 Nathan Sidwell <nathan@codesourcery.com>
29552 Iain Sandoe <iain@codesourcery.com>
29553 Nathan Froyd <froydnj@codesourcery.com>
29554 Chao-ying Fu <fu@mips.com>
29556 * doc/extend.texi: (micromips, nomicromips, nocompression):
29557 Document new function attributes.
29558 * doc/invoke.texi (minterlink-compressed, mmicromips,
29559 m14k, m14ke, m14kec): Document new options.
29560 (minterlink-mips16): Update documentation.
29561 * doc/md.texi (ZC, ZD): Document new constraints.
29562 * configure.ac (gcc_cv_as_micromips): Check if linker
29563 supports the .set micromips directive.
29564 * configure: Regenerate.
29565 * config.in: Regenerate.
29566 * config/mips/mips-tables.opt: Regenerate.
29567 * config/mips/micromips.md: New file.
29568 * constraints.md (ZC, AD): New constraints.
29569 * config/mips/predicates.md (movep_src_register): New predicate.
29570 (movep_src_operand): New predicate.
29571 (non_volatile_mem_operand): New predicate.
29572 * config/mips/mips.md (multimem): New type.
29573 (length): Differentiate between 17-bit and 18-bit branch offsets.
29574 (MOVEP1, MOVEP2): New mode iterator.
29575 (mov_<load>l): Use ZC constraint.
29576 (mov_<load>r): Likewise.
29577 (mov_<store>l): Likewise.
29578 (mov_<store>r): Likewise.
29579 (*branch_equality<mode>_inverted): Add microMIPS support.
29580 (*branch_equality<mode>): Likewise.
29581 (*jump_absolute): Likewise.
29582 (indirect_jump_<mode>): Likewise.
29583 (tablejump_<mode>): Likewise.
29584 (<optab>_internal): Likewise.
29585 (sibcall_internal): Likewise.
29586 (sibcall_value_internal): Likewise.
29587 (prefetch): Use constraint ZD.
29588 * config/mips/mips.opt (minterlink-compressed): New option.
29589 (minterlink-mips16): Now an alias for minterlink-compressed.
29590 (mmicromips): New option.
29591 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
29592 (compare_and_swap_12): Likewise.
29593 (sync_add<mode>): Likewise.
29594 (sync_<optab>_12): Likewise.
29595 (sync_old_<optab>_12): Likewise.
29596 (sync_new_<optab>_12): Likewise.
29597 (sync_nand_12): Likewise.
29598 (sync_old_nand_12): Likewise.
29599 (sync_new_nand_12): Likewise.
29600 (sync_sub<mode>): Likewise.
29601 (sync_old_add<mode>): Likewise.
29602 (sync_old_sub<mode>): Likewise.
29603 (sync_new_add<mode>): Likewise.
29604 (sync_new_sub<mode>): Likewise.
29605 (sync_<optab><mode>): Likewise.
29606 (sync_old_<optab><mode>): Likewise.
29607 (sync_new_<optab><mode>): Likewise.
29608 (sync_nand<mode>): Likewise.
29609 (sync_old_nand<mode>): Likewise.
29610 (sync_new_nand<mode>): Likewise.
29611 (sync_lock_test_and_set<mode>): Likewise.
29612 (test_and_set_12): Likewise.
29613 (atomic_compare_and_swap<mode>): Likewise.
29614 (atomic_exchange<mode>_llsc): Likewise.
29615 (atomic_fetch_add<mode>_llsc): Likewise.
29616 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
29617 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
29618 (umips_save_restore_pattern_p): Likewise.
29619 (umips_load_store_pair_p): Likewise.
29620 (umips_output_load_store_pair): Likewise.
29621 (umips_movep_target_p): Likewise.
29622 (umips_12bit_offset_address_p): Likewise.
29623 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
29624 (mips_base_mips16): Rename this...
29625 (mips_base_compression_flags): ...to this. Update all uses.
29626 (mips_attribute_table): Add micromips, nomicromips and nocompression.
29627 (mips_mips16_decl_p): Delete.
29628 (mips_nomips16_decl_p): Delete.
29629 (mips_get_compress_on_flags): New function.
29630 (mips_get_compress_off_flags): New function.
29631 (mips_get_compress_mode): New function.
29632 (mips_get_compress_on_name): New function.
29633 (mips_get_compress_off_name): New function.
29634 (mips_insert_attributes): Support multiple compression types.
29635 (mips_merge_decl_attributes): Likewise.
29636 (umips_12bit_offset_address_p): New function.
29637 (mips_start_function_definition): Emit .set micromips directive.
29638 (mips_call_may_need_jalx_p): New function.
29639 (mips_function_ok_for_sibcall): Add microMIPS support.
29640 (mips_print_operand_punctuation): Support short delay slots and
29642 (umips_swm_mask, umips_swm_encoding): New.
29643 (umips_build_save_restore): New function.
29644 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
29645 (was_mips16_p): Remove.
29646 (old_compression_mode): New.
29647 (mips_set_compression_mode): New function.
29648 (mips_set_current_function): Add microMIPS support.
29649 (mips_option_override): Likewise.
29650 (umips_save_restore_pattern_p): New function.
29651 (umips_output_save_restore): New function.
29652 (umips_load_store_pair_p_1): New function.
29653 (umips_load_store_pair_p): New function.
29654 (umips_output_load_store_pair_1): New function.
29655 (umips_output_load_store_pair): New function.
29656 (umips_movep_target_p) New function.
29657 (mips_prepare_pch_save): Add microMIPS support.
29658 * config/mips/mips.h (TARGET_COMPRESSION): New.
29659 (TARGET_CPU_CPP_BUILTINS): Update macro
29660 to use new compression flags and to support microMIPS.
29661 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
29662 (MIPS_ARCH_FLOAT_SPEC): Likewise.
29663 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
29664 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
29665 (ASM_SPEC): Support mmicromips and mno-micromips.
29666 (M16STORE_REG_P): New macro.
29667 (MIPS_CALL): Support TARGET_MICROMIPS.
29668 (MICROMIPS_J): New macro.
29669 (mips_base_mips16): Rename this...
29670 (mips_base_compression_flags): ...to this.
29671 (UMIPS_12BIT_OFFSET_P): New macro.
29672 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
29673 (MULTILIB_DIRNAMES): Likewise.
29675 2013-02-24 Jakub Jelinek <jakub@redhat.com>
29678 * target-globals.c (save_target_globals): For init_reg_sets and
29679 target_reinit remporarily set this_fn_optabs to this_target_optabs.
29681 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
29683 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
29684 * config/aarch64/t-aarch64
29685 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
29687 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
29689 PR inline-asm/56148
29690 * lra-constraints.c (process_alt_operands): Reload operand
29691 conflicting with earlier clobber only if no more other conflicting
29694 2013-02-22 Jakub Jelinek <jakub@redhat.com>
29697 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
29698 if not linking a shared library.
29700 2013-02-22 Seth LaForge <sethml@google.com>
29702 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
29704 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
29706 * config/arm/arm.md (split for extendsidi): Update condition.
29707 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
29708 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
29709 (qhs_zextenddi_cstr): Likewise.
29711 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29713 PR middle-end/56420
29714 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
29715 avoid signed wrapping.
29716 (expand_mult): Handle properly multiplication by
29717 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
29718 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
29719 in the compiler if coeff is HOST_WIDE_INT_MIN.
29720 (expand_divmod): Don't make ext_op1 static, change it's type to
29721 uhwi. Avoid undefined behavior in -INTVAL (op1).
29723 PR rtl-optimization/50339
29724 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
29726 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
29727 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
29728 into splitting_ashiftrt field.
29729 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
29731 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
29734 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
29736 PR middle-end/56108
29737 * trans-mem.c (execute_tm_mark): Do not expand transactions that
29738 are sure to go irrevocable.
29740 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
29742 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
29743 scalars are valid operands.
29745 2013-02-21 Martin Jambor <mjambor@suse.cz>
29747 PR tree-optimization/56310
29748 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
29749 only matching indices and non-negative final offsets.
29750 (intersect_aggregates_with_edge): Pass src_idx to
29751 agg_replacements_to_vector. Pass src_idx insstead of index to
29752 intersect_with_agg_replacements.
29754 2013-02-21 Martin Jambor <mjambor@suse.cz>
29756 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
29757 instead of hard-wired defaults.
29759 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
29761 * doc/invoke.texi (MIPS Options): Update documentation of the
29762 floating-point multiply-accumulate instruction restrictions.
29764 2013-02-21 Kostya Serebryany <kcc@google.com>
29766 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
29767 asan_shadow_offset on x86_64 linux.
29769 2013-02-21 Richard Biener <rguenther@suse.de>
29771 PR tree-optimization/56415
29773 2013-02-11 Richard Biener <rguenther@suse.de>
29775 PR tree-optimization/56273
29776 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
29779 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29782 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
29785 PR inline-asm/56405
29786 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
29787 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
29789 2013-02-20 Jan Hubicka <jh@suse.cz>
29791 PR tree-optimization/56265
29792 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
29793 when target is referenced for first time.
29795 2013-02-20 Richard Biener <rguenther@suse.de>
29797 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
29798 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
29799 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
29800 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
29801 not return anything.
29802 (rename_ssa_copies): Do not remove unused locals.
29803 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
29804 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
29805 * passes.c (execute_function_todo): Do not schedule unused locals
29806 removal if cleanup_tree_cfg did something.
29807 * tree-ssa-live.c (remove_unused_locals): Dump statistics
29808 about the number of removed locals.
29810 2013-02-20 Richard Biener <rguenther@suse.de>
29812 PR tree-optimization/56398
29813 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
29815 2013-02-20 Martin Jambor <mjambor@suse.cz>
29817 PR tree-optimization/55334
29818 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
29819 restricted pointers to arrays.
29821 2013-02-20 Richard Biener <rguenther@suse.de>
29822 Jakub Jelinek <jakub@redhat.com>
29824 PR tree-optimization/56396
29825 * tree-ssa-ccp.c (n_const_val): New static variable.
29826 (get_value): Return NULL for SSA names we don't have a lattice
29828 (ccp_initialize): Initialize n_const_val.
29829 * tree-ssa-copy.c (n_copy_of): New static variable.
29830 (init_copy_prop): Initialize n_copy_of.
29831 (get_value): Return NULL_TREE for SSA names we don't have a
29834 2013-02-20 Martin Jambor <mjambor@suse.cz>
29836 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
29838 2013-02-20 Richard Biener <rguenther@suse.de>
29840 * genpreds.c (write_lookup_constraint): Do not compare first
29841 letter of the constraint again.
29843 2013-02-20 Richard Biener <rguenther@suse.de>
29845 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
29847 (get_use_iv_cost): Terminate hashtable walk when coming across
29850 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
29852 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
29853 reassociation for avx2 targets.
29855 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29857 * config/microblaze/microblaze.c: microblaze_has_clz = 0
29858 Add version check for v8.10.a to enable microblaze_has_clz
29859 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
29860 version and TARGET_PATTERN_COMPARE check
29861 * config/microblaze/microblaze.md: New clzsi2 instruction
29863 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29865 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
29866 function before branching.
29868 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29870 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
29872 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
29874 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29876 PR middle-end/55889
29877 * sel-sched.c: Include ira.h.
29878 (implicit_clobber_conflict_p): New function.
29879 (moveup_expr): Use it.
29880 * Makefile.in (sel-sched.o): Depend on ira.h.
29882 2013-02-19 Richard Biener <rguenther@suse.de>
29884 PR tree-optimization/56384
29885 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
29886 (vn_hash_type): Split out from ...
29887 (vn_hash_constant_with_type): ... here.
29888 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
29889 (vn_phi_eq): Compare types from vn_phi_s structure.
29890 (vn_phi_lookup): Populate vn_phi_s type.
29891 (vn_phi_insert): Likewise.
29893 2013-02-19 Jakub Jelinek <jakub@redhat.com>
29895 PR tree-optimization/56350
29896 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
29897 if haven't found reduction or nested cycle operand, rather than
29898 asserting we must find it.
29900 PR tree-optimization/56381
29901 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
29904 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
29905 Jakub Jelinek <jakub@redhat.com>
29908 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
29909 (swap_optab_enable): Same.
29910 (init_all_optabs): Use argument instead of global.
29911 * tree.h (struct tree_optimization_option): New field target_optabs.
29912 * expr.h (init_all_optabs): Add argument to prototype.
29913 (TREE_OPTIMIZATION_OPTABS): New.
29914 (save_optabs_if_changed): Protoize.
29915 * optabs.h: Declare this_fn_optabs.
29916 * optabs.c (save_optabs_if_changed): New.
29917 Declare this_fn_optabs.
29918 (init_optabs): Add argument to init_all_optabs() call.
29919 * function.c (invoke_set_current_function_hook): Handle per
29921 * function.h (struct function): New field optabs.
29922 * config/mips/mips.c (mips_set_mips16_mode): Handle when
29923 optimization_current_node has changed.
29924 * target-globals.h (save_target_globals_default_opts): Protoize.
29925 * target-globals.c (save_target_globals_default_opts): New.
29927 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29930 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
29931 registers %fr12 and %fr12R as call used.
29934 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
29935 and HImode, require all displacements to be an integer multiple of
29937 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
29938 only allow QImode and HImode when reload is in progress and strict is
29939 true. Likewise for symbolic addresses. Use base14_operand to check
29940 displacements in REG+BASE addresses.
29942 2013-02-18 Richard Biener <rguenther@suse.de>
29944 PR tree-optimization/56366
29945 * tree-vect-loop.c (get_initial_def_for_induction): Properly
29946 handle sign-conversion of outer-loop initial induction value.
29948 2013-02-18 Richard Biener <rguenther@suse.de>
29950 PR middle-end/56349
29951 * cfghooks.c (merge_blocks): If we merge a latch into another
29952 block adjust references to it.
29953 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
29954 (verify_loop_structure): Verify that a recorded latch is in fact
29957 2013-02-18 Richard Biener <rguenther@suse.de>
29959 PR tree-optimization/56321
29960 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
29961 order SSA name release and virtual operand unlinking.
29963 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29965 * config/microblaze/microblaze.md (save_stack_block): Define.
29966 (restore_stack_block): Likewise.
29968 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29970 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
29971 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
29972 * config/microblaze/microblaze.c (microblaze_option_override):
29973 Bail out early for PIC modes when target does not support PIC.
29975 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29977 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
29978 Replace with a microblaze version.
29979 (microblaze_trampoline_init): Adapt for microblaze.
29980 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
29983 2013-02-16 Jakub Jelinek <jakub@redhat.com>
29984 Dodji Seketeli <dodji@redhat.com>
29987 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
29988 (instrument_mem_region_access): Do not forget to always put
29989 instrumentation of the of 'base' and 'base + len' in a "if (len !=
29990 0) statement, even for cases where either 'base' or 'base + len'
29991 are not instrumented -- because they have been previously
29992 instrumented. Simplify the logic by putting all the statements
29993 instrument 'base + len' inside a sequence, and then insert that
29994 sequence right before the current insertion point. Then, to
29995 instrument 'base + len', just get an iterator on that statement.
29996 And do not forget to update the pointer to iterator the function
29997 received as argument.
29999 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
30001 PR rtl-optimization/56348
30002 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
30004 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
30006 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
30007 (clean_graph_dump_file): Pass base to start_graph_dump.
30009 2013-02-14 Richard Henderson <rth@redhat.com>
30012 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
30014 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
30016 * collect2-aix.h: Define F_LOADONLY.
30018 2013-02-14 Richard Biener <rguenther@suse.de>
30021 * varasm.c (output_constant_def_1): Get the decl representing
30022 the constant as argument.
30023 (output_constant_def): Wrap output_constant_def_1.
30024 (make_decl_rtl): Use output_constant_def_1 with the decl
30025 representing the constant.
30026 (build_constant_desc): Optionally re-use a decl already
30027 representing the constant.
30028 (tree_output_constant_def): Adjust.
30030 2013-02-14 Dodji Seketeli <dodji@redhat.com>
30033 * asan.c (instrument_builtin_call): Really put the length of the
30034 second source argument into src1_len.
30036 2013-02-13 Jakub Jelinek <jakub@redhat.com>
30038 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
30039 argument. If it is false, don't create edge from then_bb to
30041 (insert_if_then_before_iter): Pass true to it.
30042 (build_check_stmt): Pass false to it.
30043 (transform_statements): Flush hash table only on extended basic
30044 block boundaries, rather than at the beginning of every bb.
30045 Don't flush hash table on nonfreeing_call_p calls.
30046 * tree-flow.h (nonfreeing_call_p): New prototype.
30047 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
30049 2013-02-13 David S. Miller <davem@davemloft.net>
30051 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
30053 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
30056 * ira.c (max_regno_before_ira): Move from ...
30058 (fix_reg_equiv_init): Use max_regno_before_ira instead of
30061 2013-02-13 Jakub Jelinek <jakub@redhat.com>
30063 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
30065 2013-02-13 Richard Biener <rguenther@suse.de>
30068 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
30069 globals in MEM_REFs.
30071 2013-02-13 Richard Biener <rguenther@suse.de>
30073 * loop-init.c (loop_optimizer_init): Clear loop state when
30074 re-initializing preserved loops.
30075 * loop-unswitch.c (unswitch_single_loop): Return whether
30076 we unswitched the loop. Do not verify loop state here.
30077 (unswitch_loops): When we unswitched a loop discover new loops.
30079 2013-02-13 Kostya Serebryany <kcc@google.com>
30081 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
30083 * sanitizer.def: Rename __asan_init to __asan_init_v1.
30085 2013-02-12 Dodji Seketeli <dodji@redhat.com>
30087 Avoid instrumenting duplicated memory access in the same basic block
30088 * Makefile.in (asan.o): Add new dependency on hash-table.h
30089 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
30090 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
30091 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
30092 (free_mem_ref_resources, has_mem_ref_been_instrumented)
30093 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
30094 (get_mem_ref_of_assignment): New functions.
30095 (get_mem_refs_of_builtin_call): Extract from
30096 instrument_builtin_call and tweak a little bit to make it fit with
30098 (instrument_builtin_call): Use the new
30099 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
30100 of is_gimple_builtin_call.
30101 (instrument_derefs, instrument_mem_region_access): Insert the
30102 instrumented memory reference into the hash table.
30103 (maybe_instrument_assignment): Renamed instrument_assignment into
30104 this, and change it to advance the iterator when instrumentation
30105 actually happened and return true in that case. This makes it
30106 homogeneous with maybe_instrument_assignment, and thus give a
30107 chance to callers to be more 'regular'.
30108 (transform_statements): Clear the memory reference hash table
30109 whenever we enter a new BB, when we cross a function call, or when
30110 we are done transforming statements. Use
30111 maybe_instrument_assignment instead of instrumentation. No more
30112 need to special case maybe_instrument_assignment and advance the
30113 iterator after calling it; it's now handled just like
30114 maybe_instrument_call. Update comment.
30116 2013-02-13 Richard Biener <rguenther@suse.de>
30118 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
30119 Fix loop discovery code.
30121 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
30123 PR inline-asm/56148
30124 * lra-constraints.c (process_alt_operands): Match early clobber
30125 operand with itself. Check conflicts with earlyclobber only if
30126 the operand is not reloaded. Prefer to reload conflicting operand
30127 if earlyclobber and matching operands are the same.
30129 2013-02-12 Richard Biener <rguenther@suse.de>
30132 * lto-streamer-out.c (write_symbol): Do not output symbols
30133 for hard register variables.
30135 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
30138 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
30139 (umulsidi3_insn, mulsidi3_insn): New insns.
30141 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
30143 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
30144 (struct tune_params): Add vec_costs field.
30145 * config/arm/arm.c (arm_builtin_vectorization_cost)
30146 (arm_add_stmt_cost): New functions.
30147 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
30148 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
30149 (arm_default_vec_cost): New struct of type cpu_vec_costs.
30150 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
30151 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
30152 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
30153 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
30155 2013-02-12 Richard Biener <rguenther@suse.de>
30158 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
30159 decls again if possible.
30161 2013-02-12 Richard Biener <rguenther@suse.de>
30163 PR middle-end/56288
30164 * tree-ssa.c (verify_ssa_name): Fix check, move
30165 SSA_NAME_IN_FREE_LIST check up.
30167 2013-02-12 Jakub Jelinek <jakub@redhat.com>
30168 Steven Bosscher <steven@gcc.gnu.org>
30170 PR rtl-optimization/56151
30171 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
30172 equal to op0 or op1, and last_insn pattern is CODE operation
30173 with MEM dest and one of the operands matches that MEM.
30175 2013-02-11 Sriraman Tallam <tmsriram@google.com>
30177 * doc/extend.texi: Document Function Multiversioning and "default"
30178 parameter string to target attribute.
30179 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
30180 target attribute parameter is "default".
30181 (ix86_compare_version_priority): Remove checks for target attribute.
30182 (ix86_mangle_function_version_assembler_name): Change error to sorry.
30183 Remove check for target attribute equal to NULL. Add assert.
30184 (ix86_generate_version_dispatcher_body): Change error to sorry.
30186 2013-02-11 Iain Sandoe <iain@codesourcery.com>
30187 Jack Howarth <howarth@bromo.med.uc.edu>
30188 Patrick Marlier <patrick.marlier@gmail.com>
30191 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
30192 define ENDFILE_SPEC as TM_DESTRUCTOR.
30193 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
30195 2013-02-11 Alexander Potapenko <glider@google.com>
30196 Jack Howarth <howarth@bromo.med.uc.edu>
30197 Jakub Jelinek <jakub@redhat.com>
30200 * config/darwin.c (cdtor_record): Rename ctor_record.
30201 (sort_cdtor_records): Rename sort_ctor_records.
30202 (finalize_dtors): New routine to sort destructors by
30203 priority before use in assemble_integer.
30204 (machopic_asm_out_destructor): Use finalize_dtors if needed.
30206 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
30208 PR rtl-optimization/56275
30209 * simplify-rtx.c (avoid_constant_pool_reference): Check that
30210 offset is non-negative and less than cmode size before
30211 calling simplify_subreg.
30213 2013-02-11 Richard Biener <rguenther@suse.de>
30215 PR tree-optimization/56264
30216 * cfgloop.h (fix_loop_structure): Adjust prototype.
30217 * loop-init.c (fix_loop_structure): Return the number of
30218 newly discovered loops.
30219 * tree-cfgcleanup.c (repair_loop_structures): When new loops
30220 are discovered, do a full loop-closed SSA rewrite.
30222 2013-02-11 Richard Biener <rguenther@suse.de>
30224 PR tree-optimization/56273
30225 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
30227 (check_array_ref): Fix missing newline in dumps.
30228 (search_for_addr_array): Likewise.
30230 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
30232 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
30234 2013-02-09 Jakub Jelinek <jakub@redhat.com>
30237 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
30239 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
30241 PR rtl-optimization/56246
30242 * lra-constraints.c (simplify_operand_subreg): Try to reuse
30244 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
30245 constraints are satisfied.
30247 2013-02-08 Jeff Law <law@redhat.com>
30250 * emit-rtl.c (reg_is_parm_p): New function.
30251 * regs.h (reg_is_parm_p): New prototype.
30252 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
30253 callee-clobbered registers.
30255 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
30258 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
30259 If there is no implicit builtin declaration, just return NULL.
30261 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
30263 * config/i386/sse.md (FMAMODEM): New mode iterator.
30264 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
30265 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
30267 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
30269 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
30270 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
30271 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
30273 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30275 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
30276 (microblaze*-*-elf): Likewise.
30277 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
30279 * config/microblaze/microblaze-c.c: Add builtin defines for
30280 _LITTLE_ENDIAN and _BIG_ENDIAN.
30281 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
30282 add to TARGET_DEFAULT flags.
30283 Expand ASM_SPEC and LINK_SPEC.
30284 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
30285 * config/microblaze/microblaze.md: Update extendsidi2 and
30286 movdi_internal instructions to use low-order / high-order reg
30288 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
30289 options and inversemask / mask of LITTLE_ENDIAN.
30290 * config/microblaze/t-microblaze: Expand multilib options to
30291 include mlittle-endian (le) and update exceptions patterns.
30293 2013-02-08 Jakub Jelinek <jakub@redhat.com>
30295 PR rtl-optimization/56195
30296 * lra-constraints.c (get_reload_reg): Don't reuse regs
30297 if they have smaller mode than requested, if they have
30298 wider mode than requested, try to return a SUBREG.
30300 PR tree-optimization/56250
30301 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
30302 if type is unsigned and code isn't MULT_EXPR.
30304 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
30306 PR tree-optimization/56064
30307 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
30308 bits according to mode.
30309 * fixed-value.h (fixed_from_double_int)
30310 (const_fixed_from_double_int): Adjust comments.
30312 2013-02-08 Richard Biener <rguenther@suse.de>
30315 * lto-streamer.h (struct data_in): Remove current_file, current_line
30316 and current_col members.
30317 * lto-streamer-out.c (lto_output_location): Stream changed bits
30318 en-block for efficiency.
30319 * lto-streamer-in.c (clear_line_info): Remove.
30320 (lto_input_location): Cache current file, line and column
30321 globally via local statics. Read changed bits en-block.
30322 (input_function): Do not call clear_line_info.
30323 (lto_read_body): Likewise.
30324 (lto_input_toplevel_asms): Likewise.
30326 2013-02-08 Michael Matz <matz@suse.de>
30328 PR tree-optimization/52448
30329 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
30330 (nt_call_phase): New static.
30331 (add_or_mark_expr): Only mark accesses with newer phase than any
30333 (nonfreeing_call_p): New.
30334 (nt_init_block): Update nt_call_phase, mark blocks as visited.
30335 (nt_fini_block): Keep blocks marked as visited.
30336 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
30338 2013-02-08 Richard Biener <rguenther@suse.de>
30340 * ira.c (ira): Free broken dominator information.
30342 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
30344 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
30346 2013-02-08 Marek Polacek <polacek@redhat.com>
30348 * cfgloop.c (verify_loop_structure): Add more checking of headers.
30350 2013-02-08 Richard Biener <rguenther@suse.de>
30352 PR middle-end/56181
30353 * cfgloop.h (flow_loops_find): Adjust.
30354 (bb_loop_header_p): Declare.
30355 * cfgloop.c (bb_loop_header_p): New function split out from ...
30356 (flow_loops_find): ... here. Adjust function signature,
30357 support incremental loop structure update.
30358 (verify_loop_structure): Cleanup. Verify a loop is a loop.
30359 * cfgloopmanip.c (fix_loop_structure): Move ...
30360 * loop-init.c (fix_loop_structure): ... here.
30361 (apply_loop_flags): Split out from ...
30362 (loop_optimizer_init): ... here.
30363 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
30364 in incremental mode, only remove dead loops here.
30366 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
30369 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
30370 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
30371 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
30372 (*round<mode>3.libgcc): New insns for fixed-modes.
30373 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
30374 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
30375 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
30376 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
30377 implementations. Define to __builtin_avr_absFX,
30378 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
30379 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
30380 __builtin_avr_countlsFX, respectively.
30381 * config/avr/avr-c.c (target.h): Include it.
30382 (enum avr_builtin_id): New enum.
30383 (avr_resolve_overloaded_builtin): New static function.
30384 (avr_register_target_pragmas): Use it to set
30385 targetm.resolve_overloaded_builtin.
30386 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
30387 tree nodes used by DEF_BUILTIN.
30388 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
30389 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
30390 <AVR_BUILTIN_xxBITS>: Same.
30392 2013-02-08 Richard Biener <rguenther@suse.de>
30394 * cfgloop.c (verify_loop_structure): Properly handle
30395 a loop exiting to another loop header.
30396 * ira-int.h (ira_loops): Remove.
30397 * ira.c (ira_loops): Remove.
30398 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
30399 (do_reload): Use loop_optimizer_finalize.
30400 * ira-build.c (create_loop_tree_nodes): Use get_loops and
30401 number_of_loops to access the loop tree.
30402 (more_one_region_p): Likewise.
30403 (finish_loop_tree_nodes): Likewise.
30404 (rebuild_regno_allocno_maps): Likewise.
30405 (mark_loops_for_removal): Likewise.
30406 (mark_all_loops_for_removal): Likewise.
30407 (remove_unnecessary_regions): Likewise.
30408 (ira_build): Likewise.
30409 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
30411 2013-02-08 Richard Biener <rguenther@suse.de>
30413 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
30414 * ipa-pure-const.c (analyze_function): Avoid calling
30415 mark_irreducible_loops twice.
30416 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
30418 2013-02-07 David S. Miller <davem@davemloft.net>
30420 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
30422 * var-tracking.c (vt_add_function_parameter): Test the presence of
30423 HAVE_window_save properly and do not remap argument registers when
30424 we have a leaf function.
30426 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
30429 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
30431 * config/i386/i386.c (ix86_print_operand): Ditto.
30433 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
30435 * lra-constraints.c (process_alt_operands): Fix recently added comment.
30437 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
30439 PR rtl-optimization/56225
30440 * lra-constraints.c (process_alt_operands): Check that reload hard
30441 reg can hold value for strict_low_part.
30443 2013-02-07 Jakub Jelinek <jakub@redhat.com>
30446 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
30447 dwarf2out_end_function.
30448 (in_first_function_p, maybe_at_text_label_p,
30449 first_loclabel_num_not_at_text_label): New variables.
30450 (dwarf2out_var_location): In the first function find out
30451 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
30452 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
30455 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
30457 PR rtl-optimization/56178
30458 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
30459 SUBREG of a register. Tidy up related block of code.
30460 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
30461 note if the source is a register or a SUBREG of a register.
30463 2013-02-07 Jakub Jelinek <jakub@redhat.com>
30466 * config/rs6000/rs6000.md (ptrm): New mode attr.
30467 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
30468 call_value_indirect_aix<pttrsize>,
30469 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
30472 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
30474 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
30475 if -bnortl. Convert to strcmp and strncmp.
30477 2013-02-07 Alan Modra <amodra@gmail.com>
30480 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
30481 addresses won't wrap when offsetting.
30482 (rs6000_secondary_reload): Provide secondary reloads needed for
30483 wrapping LO_SUM addresses.
30485 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
30487 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
30488 MACH, just __MACH__.
30490 2013-02-06 Richard Biener <rguenther@suse.de>
30492 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
30493 instead of calling fix_loop_structure.
30495 2013-02-06 Jakub Jelinek <jakub@redhat.com>
30497 PR middle-end/56217
30498 * omp-low.c (use_pointer_for_field): Return false if
30499 lower_send_shared_vars doesn't generate any copy-out code.
30501 2013-02-06 Tom de Vries <tom@codesourcery.com>
30503 PR rtl-optimization/56131
30504 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
30505 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
30506 of the label is NULL. Add comment.
30508 2013-02-05 Jakub Jelinek <jakub@redhat.com>
30510 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
30513 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
30514 (STATIC_LIBTSAN_LIBS): Likewise.
30515 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
30516 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
30517 is defined, don't add anything else beyond that.
30518 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
30519 (LINK_COMMAND_SPEC): Use them.
30521 PR tree-optimization/56205
30522 * tree-stdarg.c (check_all_va_list_escapes): Return true if
30523 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
30524 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
30526 2013-02-05 Richard Biener <rguenther@suse.de>
30528 PR tree-optimization/53342
30529 PR tree-optimization/53185
30530 * tree-vectorizer.h (vect_check_strided_load): Remove.
30531 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
30532 not disallow peeling for vectorized strided loads.
30533 (vect_check_strided_load): Make static and simplify.
30534 (vect_analyze_data_refs): Adjust.
30535 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
30536 correctly when vectorizing strided loads.
30538 2013-02-05 Richard Biener <rguenther@suse.de>
30540 * doc/install.texi: Refer to ISL, not PPL.
30542 2013-02-05 Jan Hubicka <jh@suse.cz>
30544 PR tree-optimization/55789
30545 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
30547 2013-02-05 Jan Hubicka <jh@suse.cz>
30549 PR tree-optimization/55789
30550 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
30551 the dead call anyway.
30553 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
30556 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
30558 2013-02-04 Alexander Potapenko <glider@google.com>
30559 Jack Howarth <howarth@bromo.med.uc.edu>
30560 Jakub Jelinek <jakub@redhat.com>
30563 * config/darwin.c (sort_ctor_records): Stabilized qsort
30564 on constructor priority by using original position.
30565 (finalize_ctors): New routine to sort constructors by
30566 priority before use in assemble_integer.
30567 (machopic_asm_out_constructor): Use finalize_ctors if needed.
30569 2013-02-04 Jakub Jelinek <jakub@redhat.com>
30572 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
30573 about visibility on artificial decls.
30574 * config/sol2.c (solaris_assemble_visibility): Likewise.
30576 2013-02-04 Kai Tietz <ktietz@redhat.com>
30579 * config/i386/i386.c (function_value_ms_64): Add additional valtype
30580 argument and improve checking of return-argument types for 16-byte
30582 (ix86_function_value_1): Add additional valtype argument on call
30583 of function_value_64.
30584 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
30585 handling infunction_value_64 function.
30587 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
30589 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
30591 2013-02-04 Richard Biener <rguenther@suse.de>
30593 PR tree-optimization/56188
30594 * tree-ssa-structalias.c (label_visit): Consider case with
30595 initially non-empty points-to set.
30596 (perform_var_substitution): Dump node mapping and clean up.
30598 2013-02-04 Richard Guenther <rguenther@suse.de>
30601 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
30602 node prevail as last resort.
30603 (lto_symtab_merge_decls): Remove guard on LTRANS here.
30604 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
30606 2013-02-04 Richard Biener <rguenther@suse.de>
30608 PR tree-optimization/56113
30609 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
30611 (equiv_class_lookup_or_add): ... this.
30612 (label_visit): Adjust and fix error in previous patch.
30613 (perform_var_substitution): Adjust.
30615 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
30617 * config/sh/divtab.c: Fix formatting and comments throughout the file.
30618 * config/sh/sh4-300.md: Likewise.
30619 * config/sh/sh4a.md: Likewise.
30620 * config/sh/constraints.md: Likewise.
30621 * config/sh/sh.md: Likewise.
30622 * config/sh/netbsd-elf.h: Likewise.
30623 * config/sh/predicates.md: Likewise.
30624 * config/sh/sh-protos.h: Likewise.
30625 * config/sh/ushmedia.h: Likewise.
30626 * config/sh/linux.h: Likewise.
30627 * config/sh/sh.c: Likewise.
30628 * config/sh/superh.h: Likewise.
30629 * config/sh/elf.h: Likewise.
30630 * config/sh/sh4.md: Likewise.
30631 * config/sh/sh.h: Likewise.
30633 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30635 * config/pa/constraints.md: Adjust unused letters. Change "T"
30636 constraint to match_test floating_point_store_memory_operand().
30637 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
30638 (base14_operand): New.
30639 (floating_point_store_memory_operand): New.
30640 (integer_store_memory_operand): Revise to use base14_operand and
30641 reg_plus_base_memory_operand.
30642 (move_dest_operand): Allow symbolic_memory_operands.
30643 (symbolic_memory_operand): Check for LO_SOM.
30644 (symbolic_operand): Change default case to break.
30645 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
30646 CONST_DOUBLE values to be reloaded by putting them into memory when
30647 the destination is a floating point register.
30648 (movdf): Remove code to handle CONST_DOUBLE.
30650 (reload_indf_r1): New.
30651 (reload_insf_r1): New.
30652 Consistently use "Q" and "T" constraints with integer and floating
30653 point move instructions, respectively.
30654 (movdi): Remove FAIL.
30655 Change predicate for source operand unamed DImode move from
30656 general_operand to move_src_operand.
30657 (umulsidi3): Change predicate for destination operand to
30659 Likewise for similar unamed patterns.
30660 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
30661 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
30662 (hppa_legitimize_address): Simplify mask calculation.
30663 (pa_emit_move_sequence): Revised handling of secondary reloads from
30664 REG+D addresses for floating point loads and stores. Directly handle
30665 loading CONST0_RTX (mode) to a floating point register.
30666 (pa_secondary_reload): Handle reloading DF and SFmode constant values
30667 to floating point registers. Don't restrict secondary reloads to
30668 floating point registers to integer modes. Revise some comments and
30670 (TARGET_LEGITIMATE_ADDRESS_P): Define.
30671 (pa_legitimate_address_p): New.
30672 (pa_legitimize_reload_address): New.
30673 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
30674 (STRICT_REG_OK_FOR_BASE_P): New.
30675 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
30676 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
30678 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
30679 Andrew Dixie <andrewd@gentrack.com>
30681 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
30684 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
30686 * expmed.c (extract_bit_field_1): Pass the full width of the
30687 structure to get_best_reg_extraction_insn.
30689 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
30692 * configure.ac (use_cxa_atexit): Add AIX.
30693 * configure: Regenerate.
30695 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
30697 2013-02-01 Jakub Jelinek <jakub@redhat.com>
30700 * final.c (need_profile_function): New variable.
30701 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
30702 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
30703 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
30704 notes, targetm.asm_out.function_prologue doesn't emit anything,
30705 HAVE_prologue and profiler should be emitted before prologue,
30706 set need_profile_function instead of emitting it.
30707 (final_scan_insn): If need_profile_function, emit
30708 profile_function on the first NOTE_INSN_BASIC_BLOCK or
30709 NOTE_INSN_FUNCTION_BEG note.
30711 2013-02-01 Richard Henderson <rth@redhat.com>
30713 * config/rs6000/rs6000.md (smulditi3): New.
30716 * config/alpha/alpha.md (umulditi3): New.
30718 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
30720 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
30721 (ASM_OUTPUT_ALIGNED_LOCAL): New.
30723 2013-02-01 Richard Biener <rguenther@suse.de>
30725 PR tree-optimization/56113
30726 * tree-ssa-structalias.c (label_visit): Reduce work for
30727 single-predecessor nodes.
30729 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
30731 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
30732 range isn't testing for zero.
30734 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
30736 PR middle-end/56113
30737 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
30739 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
30740 Nick Clifton <nickc@redhat.com>
30742 * config/v850/constraints.md (Q): Define as a memory constraint.
30743 * config/v850/predicates.md (label_ref_operand): New predicate.
30744 (e3v5_shift_operand): New predicate.
30745 (ior_operator): New predicate.
30746 * config/v850/t-v850: Add e3v5 multilib.
30747 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
30748 (v850_gen_movdi): Prototype.
30749 * config/v850/v850.c: Add support for e3v5 architecture.
30750 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
30752 (construct_save_jarl): Add e3v5 long JARL support.
30753 (v850_adjust_insn_length): New function. Adjust length of call
30754 insns when using e3v5 instructions.
30755 (v850_gen_movdi): New function: Generate instructions to move a
30757 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
30758 (CPP_SPEC): Define __v850e3v5__ as appropriate.
30759 (TARGET_USE_FPU): Enable for e3v5.
30760 (CONST_OK_FOR_W): New macro.
30761 (ADJUST_INSN_LENGTH): Define.
30762 * config/v850/v850.md (UNSPEC_LOOP): Define.
30763 (attr cpu): Add v850e3v5.
30764 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
30765 (movdi): New pattern.
30766 (movdi_internal): New pattern.
30767 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
30768 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
30769 (cstoresf4): Likewise.
30770 (cstoredf4): Likewise.
30771 (insv): New pattern.
30772 (rotlso3_a): New pattern.
30773 (rotlsi3_b): New pattern
30774 (rotlsi3_v850e3v5): New pattern.
30775 (doloop_begin): New pattern.
30776 (fix_loop_counter): New pattern.
30777 (doloop_end): New pattern.
30778 (branch_normal): Add e3v5 long branch support.
30779 (branch_invert): Likewise.
30780 (branch_z_normal): Likewise.
30781 (branch_z_invert): Likewise.
30782 (branch_nz_normal): Likewise.
30783 (branch_nz_invert): Likewise.
30784 (call_internal_short): Add e3v5 register-indirect JARL support.
30785 (call_internal_long): Likewise.
30786 (call_value_internal_short): Likewise.
30787 (call_value_internal_long): Likewise.
30788 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
30789 (mloop): New option.
30790 * config.gcc: Add support for configuring v840e3v5 target.
30791 * doc/invoke.texi: Document new v850 specific command line options.
30793 2013-01-31 Paul Koning <ni1d@arrl.net>
30797 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
30798 children if parent is a class.
30799 (prune_unused_types_prune): Don't add DW_AT_declaration.
30801 2013-01-31 Richard Biener <rguenther@suse.de>
30803 PR tree-optimization/56157
30804 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
30805 match up operand with SLP child.
30807 2013-01-31 Jason Merrill <jason@redhat.com>
30810 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
30811 parameters the first time.
30812 (gen_scheduled_generic_parms_dies): Check completeness here.
30814 2013-01-31 Richard Biener <rguenther@suse.de>
30816 PR middle-end/53073
30817 * common.opt (faggressive-loop-optimizations): New flag,
30818 enabled by default.
30819 * doc/invoke.texi (faggressive-loop-optimizations): Document.
30820 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
30821 infer_loop_bounds_from_undefined by it.
30823 2013-01-31 Richard Biener <rguenther@suse.de>
30825 PR tree-optimization/56150
30826 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
30827 visit virtual operands.
30828 (find_uses_to_rename_bb): Likewise.
30830 2013-01-31 Richard Biener <rguenther@suse.de>
30832 PR tree-optimization/56150
30833 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
30834 mixed store non-store stmts.
30836 2013-01-30 Jakub Jelinek <jakub@redhat.com>
30839 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
30840 LIBASAN_EARLY_SPEC is defined.
30841 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
30842 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
30844 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
30847 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
30848 invalid args instead of ICEing on it.
30849 (ix86_valid_target_attribute_tree): Return error_mark_node if
30850 ix86_valid_target_attribute_inner_p failed.
30851 (ix86_valid_target_attribute_p): Return false only if
30852 ix86_valid_target_attribute_tree returned error_mark_node. Allow
30853 target("default") attribute.
30854 (sorted_attr_string): Change argument from const char * to tree,
30855 merge in all target attribute arguments rather than just one.
30856 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
30857 instead of free. Avoid using strcat.
30858 (ix86_mangle_function_version_assembler_name): Mangle
30859 target("default") as if no target attribute is present. Adjust
30860 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
30861 instead of xmalloc and XDELETEVEC instead of free.
30862 (ix86_function_versions): Don't return true if one of the decls
30863 doesn't have target attribute. If they don't and one of the decls
30864 is DECL_FUNCTION_VERSIONED, report an error. Adjust
30865 sorted_attr_string caller. Use XDELETEVEC instead of free.
30866 (ix86_supports_function_versions): Remove.
30867 (make_name): Fix up formatting.
30868 (make_dispatcher_decl): Remove resolver_name and its initialization.
30869 Avoid leaking memory.
30870 (is_function_default_version): Return true if there is
30871 target("default") attribute rather than no target attribute at all.
30872 (make_resolver_func): Avoid leaking memory.
30873 (ix86_generate_version_dispatcher_body): Likewise.
30874 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
30875 * target.def (supports_function_versions): Remove.
30876 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
30877 * doc/tm.texi: Regenerated.
30879 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
30881 PR rtl-optimization/56144
30882 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
30883 for values with side effects.
30885 2013-01-30 Richard Biener <rguenther@suse.de>
30887 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
30888 (sparseset_pop): Likewise.
30889 * cfganal.c (compute_idf): Likewise. Increase work-stack size
30890 to be able to use quick_push in the worker loop.
30892 2013-01-30 Marek Polacek <polacek@redhat.com>
30894 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
30896 2013-01-30 Richard Biener <rguenther@suse.de>
30899 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
30901 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
30903 PR tree-optimization/56064
30904 * fixed-value.c (fixed_from_double_int): New function.
30905 * fixed-value.h (fixed_from_double_int): New prototype.
30906 (const_fixed_from_double_int): New static inline function.
30907 * fold-const.c (native_interpret_fixed): New static function.
30908 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
30909 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
30910 (native_encode_fixed): New static function.
30911 (native_encode_expr) <FIXED_CST>: Use it.
30912 (native_interpret_int): Move double_int worker code to...
30913 * double-int.c (double_int::from_buffer): ...this new static method.
30914 * double-int.h (double_int::from_buffer): Prototype it.
30916 2013-01-30 Richard Biener <rguenther@suse.de>
30918 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
30919 New pointer-map and obstack.
30920 (init_alias_vars): Allocate pointer-map and obstack.
30921 (delete_points_to_sets): Free them.
30922 (find_what_var_points_to): Cache result.
30923 (find_what_p_points_to): Adjust for changed interface of
30924 find_what_var_points_to.
30925 (compute_points_to_sets): Likewise.
30926 (ipa_pta_execute): Likewise.
30928 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
30930 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
30931 * configure: Regenerate.
30932 * config.in: Regenerate.
30933 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
30934 #nobits/#progbits if supported.
30936 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
30939 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
30940 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
30941 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
30943 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30945 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
30946 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
30948 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30950 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
30951 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
30953 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30955 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
30957 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
30958 * config/arm/cortex-a7.md: New bypasses using
30959 arm_mac_accumulator_is_result.
30961 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30963 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
30964 (cortex_a7_neon_mla): Likewise.
30965 (cortex_a7_fpfmad): New reservation.
30966 (cortex_a7_fpmacs): Use ffmas and update required units.
30967 (cortex_a7_fpmuld): Update required units and latency.
30968 (cortex_a7_fpmacd): Likewise.
30969 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
30970 (cortex_a7_neon). Likewise.
30971 (bypass) Update participating units.
30973 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30975 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
30976 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
30978 * config/arm/vfp11.md (vfp_farith): Use ffmas.
30979 (vfp_fmul): Use ffmad.
30980 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
30981 (cortex_r4_fmacd): Use ffmad.
30982 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
30983 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
30984 (cortex_a9_fmacd): Use ffmad.
30985 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
30986 (cortex_a8_vfp_macd): Use ffmad.
30987 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
30988 (cortex_a5_fpmacd): Use ffmad.
30989 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
30990 (cortex_a15_vfp_macd): Use ffmad.
30991 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
30993 2013-01-29 Jason Merrill <jason@redhat.com>
30996 * varasm.c (default_assemble_visibility): Don't warn about
30997 visibility on artificial decls.
30999 2013-01-29 Richard Biener <rguenther@suse.de>
31001 PR tree-optimization/56113
31002 * tree-ssa-structalias.c (equiv_class_lookup): Also return
31004 (label_visit): Free duplicate bitmaps and record the leader instead.
31005 (perform_var_substitution): Adjust.
31007 2013-01-29 Richard Biener <rguenther@suse.de>
31009 PR tree-optimization/55270
31010 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
31011 the CFG, schedule loops for fixup.
31013 2013-01-29 Nick Clifton <nickc@redhat.com>
31015 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
31018 2013-01-28 Leif Ekblad <leif@rdos.net>
31020 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
31021 * config/i386/i386.h (TARGET_RDOS): New macro.
31022 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
31023 * config/i386/i386.c (ix86_option_override_internal): For 64bit
31024 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
31025 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
31026 DEFAULT_LARGE_SECTION_THRESHOLD.
31027 * config/i386/i386.md (R14_REG, R15_REG): New constants.
31028 * config/i386/rdos.h: New file.
31029 * config/i386/rdos64.h: New file.
31031 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
31034 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
31037 2013-01-28 Jakub Jelinek <jakub@redhat.com>
31039 PR rtl-optimization/56117
31040 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
31041 call cselib_lookup_from_insn on the MEM before calling
31042 add_insn_mem_dependence.
31044 2013-01-28 Richard Biener <rguenther@suse.de>
31046 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
31047 to a stmt that didn't have one.
31048 (copy_phis_for_bb): Likewise for PHI arguments.
31049 (copy_debug_stmt): Likewise for debug stmts.
31051 2013-01-28 Richard Biener <rguenther@suse.de>
31053 PR tree-optimization/56034
31054 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
31055 (partition_builtin_p): Adjust.
31056 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
31057 it is the last partition.
31058 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
31059 up the vertex for the definition.
31060 (classify_partition): Classify whether a partition is a
31061 PKIND_REDUCTION, thus has uses outside of the loop.
31062 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
31063 Merge all PKIND_REDUCTION partitions into the last partition.
31064 (tree_loop_distribution): Seed partitions from reductions as well.
31066 2013-01-28 Jakub Jelinek <jakub@redhat.com>
31068 PR tree-optimization/56125
31069 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
31070 pow(x,c) into sqrt(x) * powi(x, n/2) or
31071 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
31072 optimizing for size.
31073 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
31074 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
31077 PR tree-optimization/56094
31078 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
31079 to UNKNOWN_LOCATION while gimplifying expr.
31081 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
31084 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
31085 operand 0 in movabs insn template for -masm=intel asm alternative.
31086 (*movabs<mode>_2): Ditto for operand 1.
31088 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
31091 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
31094 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31096 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
31099 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
31101 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
31102 Undef to avoid warning.
31104 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
31106 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
31107 * configure: Regenerate.
31109 2013-01-25 Jakub Jelinek <jakub@redhat.com>
31111 PR tree-optimization/56098
31112 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
31113 for stmts with volatile ops.
31114 (cond_store_replacement): Don't optimize if assign has volatile ops.
31115 (cond_if_else_store_replacement_1): Don't optimize if either
31116 then_assign or else_assign have volatile ops.
31117 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
31120 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
31122 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
31124 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
31126 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
31127 missing ':' in asm example.
31129 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
31131 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
31132 entries into lane and laneq entries.
31133 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
31134 Remove AdvSIMD scalar modes.
31135 (aarch64_sq<r>dmulh_laneq<mode>): New.
31136 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
31138 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
31139 builtin implementations to relfect changes in RTL in aarch64-simd.md.
31140 * config/aarch64/iterators.md (VCOND): New.
31143 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
31146 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
31147 Add NULL LIBNAME argument to existing definitions.
31148 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
31149 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
31150 * config/avr/avr.c (DEF_BUILTIN): Same.
31151 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
31152 (avr_expand_builtin): Expand to a vanilla call if a libgcc
31153 implementation is available (DECL_ASSEMBLER_NAME is set).
31154 (avr_fold_absfx): New static function.
31155 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
31156 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
31157 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
31158 AVR_BUILTIN_ABSLLK.
31159 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
31160 (abshk, absk, abslk, absllk): Provide as static inline functions.
31162 2013-01-25 Marek Polacek <polacek@redhat.com>
31164 PR tree-optimization/56035
31165 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
31167 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
31169 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
31170 (*movtf_internal_rex64): Add (!o,C) alternative
31171 (*movxf_internal_rex64): Ditto.
31172 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
31174 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
31176 * doc/invoke.texi: fix typo.
31177 * doc/objc.texi: fix typo.
31179 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
31181 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
31182 for the first two alternatives.
31184 2013-01-24 Diego Novillo <dnovillo@google.com>
31186 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
31187 (ggc-zone.o): Remove.
31188 * configure.ac: Remove option --with-gc.
31189 * configure: Re-generate.
31190 * doc/install.texi: Remove documentation for --with-gc.
31191 * gengtype.c (write_enum_defn): Remove. Update all users.
31192 (write_Types_process_field): Remove generation of gt_e_* argument.
31193 (output_type_enum): Remove. Update all users.
31194 (write_enum_defn): Remove. Update all users.
31195 (enum alloc_zone): Remove. Update all users.
31196 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
31197 * ggc-common.c (ggc_splay_alloc): Remove first argument.
31198 Update all callers.
31199 (struct ptr_data): Remove field TYPE. Update all users.
31200 (gt_pch_note_object): Remove argument TYPE. Update all users.
31201 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
31203 * ggc-none.c (ggc_alloc_typed_stat): Remove.
31204 (struct alloc_zone): Remove.
31205 (ggc_internal_alloc_zone_stat): Remove.
31206 (ggc_internal_cleared_alloc_zone_stat): Remove.
31207 * ggc-page.c (ggc_alloc_typed_stat): Remove.
31208 (ggc_pch_count_object): Remove last argument. Update all users.
31209 (ggc_pch_alloc_object): Remove last argument. Update all users.
31210 (struct alloc_zone): Remove.
31211 * ggc-zone.c: Remove.
31212 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
31213 (struct alloc_zone): Remove.
31214 (ggc_alloc_typed_stat): Remove.
31215 (ggc_alloc_typed): Remove.
31216 (ggc_splay_alloc): Remove first argument.
31217 (rtl_zone): Remove. Update all users.
31218 (tree_zone): Remove. Update all users.
31219 (tree_id_zone): Remove. Update all users.
31220 (ggc_internal_zone_alloc_stat): Remove. Update all users.
31221 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
31222 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
31223 * tree-ssanames.c: Remove references to zone allocator in comments.
31225 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
31227 * config/avr/avr.c (avr_out_fract): Make register numbers that
31228 might be outside of source operand signed.
31230 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
31232 * config/i386/constraints.md (Yf): New constraint.
31233 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
31234 of f constraint to conditionaly disable x87 register preferences.
31235 (*movdf_internal): Ditto.
31236 (*movsf_internal): Ditto.
31238 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
31240 PR inline-asm/55934
31241 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
31242 that have operands with impossible constraints.
31243 Add a FIXME for a speed-up opportunity.
31244 * lra-constraints.c (process_alt_operands): Verify that a class
31245 selected from constraints on asms is valid for the operand mode.
31246 (curr_insn_transform): Remove incorrect comment.
31248 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
31250 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
31251 TOC operand is a valid symbol ref in the constant pool.
31253 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31255 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
31257 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
31260 * config/avr/stdfix.h: New file.
31261 * t-avr (stdfix-gcc.h): New rule to build it.
31262 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
31264 2013-01-23 Kostya Serebryany <kcc@google.com>
31266 * config/darwin.h: remove dependency on
31267 CoreFoundation (asan on Mac OS).
31269 2013-01-23 Jakub Jelinek <jakub@redhat.com>
31272 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
31273 instead of cmpdi_operand for first comparison operand.
31274 Don't assert that comparison operands aren't both constants.
31276 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
31278 * doc/install.texi (Downloading the Source): Update references to
31279 downloading separate components.
31281 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
31283 * doc/extend.texi (__int128): Improve grammar.
31285 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
31288 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
31289 alternative to (o,r).
31290 (*movdi_internal_rex64): Remove (!o,n) alternative.
31291 (DImode immediate->memory splitter): Remove.
31292 (DImode immediate->memory peephole2): Remove.
31293 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
31294 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
31295 alternative to (!o,*r).
31296 (*movtf_internal_sse): New pattern.
31297 (*movxf_internal_rex64): New pattern.
31298 (*movxf_internal): Disable for TARGET_64BIT.
31299 (*movdf_internal_rex64): Remove (!o,F) alternative.
31301 2013-01-22 Jakub Jelinek <jakub@redhat.com>
31303 PR middle-end/56074
31304 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
31305 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
31306 * tree-vect-loop-manip.c (find_loop_location): Also ignore
31307 stmt locations where LOCATION_LOCUS of the stmt location is
31308 UNKNOWN_LOCATION or BUILTINS_LOCATION.
31311 * config/i386/i386.md (UNSPEC_STOS): New.
31312 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
31313 *strsetqi_1): Add UNSPEC_STOS.
31315 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
31318 * doc/invoke.texi: Remove left over -Wsynth example.
31320 2013-01-21 Jakub Jelinek <jakub@redhat.com>
31322 PR tree-optimization/56051
31323 * fold-const.c (fold_binary_loc): Don't fold
31324 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
31325 a narrowing conversion, or widening conversion from signed
31328 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
31330 PR rtl-optimization/56023
31331 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
31332 dependent on debug instruction.
31334 2013-01-21 Martin Jambor <mjambor@suse.cz>
31336 PR middle-end/56022
31337 * function.c (allocate_struct_function): Call
31338 invoke_set_current_function_hook earlier.
31340 2013-01-21 Jakub Jelinek <jakub@redhat.com>
31342 * reload1.c (init_reload): Only initialize reload_obstack
31343 during the first call.
31345 2013-01-21 Marek Polacek <polacek@redhat.com>
31347 * cfgloop.c (verify_loop_structure): Fix up grammar.
31349 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
31351 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
31352 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
31354 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
31357 * config/arm/marvell-pj4.md: Update copyright year.
31358 Fix up use of alu to alu_reg and simple_alu_imm.
31360 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
31362 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
31364 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
31367 * lra-constraints.c (curr_insn_transform): Don't reuse original
31368 insn for secondary memory move when memory mode should be different.
31370 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31372 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
31373 atomic_storedi_1): New patterns.
31375 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
31377 btver2 pipeline descriptions.
31378 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
31380 * config/i386/i386.md (btver2_decode): New type attributes.
31381 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
31383 * config/i386/btver2.md: New file describing btver2 pipelines.
31385 2013-01-19 Andrew Pinski <apinski@cavium.com>
31387 PR tree-optimization/52631
31388 * tree-ssa-sccvn (visit_use): Before looking up the original
31389 statement, try looking up the simplified expression.
31391 2013-01-19 Anthony Green <green@moxielogic.com>
31393 * config/moxie/moxie.c (moxie_expand_prologue): Set
31394 current_function_static_stack_size.
31396 2013-01-18 Jakub Jelinek <jakub@redhat.com>
31398 PR tree-optimization/56029
31399 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
31400 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
31402 2013-01-18 Sharad Singhai <singhai@google.com>
31404 PR tree-optimization/55995
31405 * dumpfile.c (dump_loc): Print location only if available.
31406 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
31408 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
31411 * lra-constraints.c (curr_insn_transform): Reuse original insn for
31412 secondary memory move.
31413 (inherit_reload_reg): Use rclass instead of cl for
31414 check_secondary_memory_needed_p.
31416 2013-01-18 Jakub Jelinek <jakub@redhat.com>
31418 PR middle-end/56015
31419 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
31420 the case where writing real complex part of target modifies op1.
31422 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
31424 * config/aarch64/aarch64-simd.md
31425 (aarch64_vcond_internal<mode>): Handle unordered cases.
31426 * config/aarch64/iterators.md (v_cmp_result): New.
31428 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
31429 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
31431 * config/arm/marvell-pj4.md: New file.
31432 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
31433 * config/arm/arm.md (generic_sched): Add marvell_pj4.
31434 (generic_vfp): Likewise.
31435 * config/arm/arm-cores.def: Add marvell-pj4.
31436 * config/arm/arm-tune.md: Regenerate.
31437 * config/arm/arm-tables.opt: Regenerate.
31438 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
31439 * doc/invoke.texi: Document marvell-pj4.
31441 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
31443 * config/aarch64/arm_neon.h: Map scalar types to standard types.
31445 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31450 * var-tracking.c (negative_power_of_two_p): New.
31451 (global_get_addr_cache, local_get_addr_cache): New.
31452 (get_addr_from_global_cache, get_addr_from_local_cache): New.
31453 (vt_canonicalize_addr): Rewrite using the above. Adjust the
31455 (vt_stack_offset_p): Remove.
31456 (vt_canon_true_dep): Always canonicalize loc's address.
31457 (clobber_overlapping_mems): Make sure we have a MEM.
31458 (local_get_addr_clear_given_value): New.
31459 (val_reset): Clear local cached entries.
31460 (compute_bb_dataflow): Create and release the local cache.
31461 Disable duplicate MEMs clobbering.
31462 (emit_notes_in_bb): Clobber MEMs likewise.
31463 (vt_emit_notes): Create and release the local cache.
31464 (vt_initialize, vt_finalize): Create and release the global
31465 cache, respectively.
31466 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
31468 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31470 PR libmudflap/53359
31471 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
31472 not found in the symtab.
31474 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31477 PR rtl-optimization/55547
31478 PR rtl-optimization/53827
31481 * alias.c (offset_overlap_p): New, factored out of...
31482 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
31483 the conservative special case for symbolic constants. Don't
31484 adjust zero sizes on alignment.
31486 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
31488 PR rtl-optimization/52573
31489 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
31490 REG_UNUSED for the same register.
31492 2013-01-17 Richard Biener <rguenther@suse.de>
31493 Marek Polacek <polacek@redhat.com>
31495 PR rtl-optimization/55833
31496 * loop-unswitch.c (unswitch_loops): Move loop verification...
31497 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
31498 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
31499 Set it to true when we're removing a loop from hierarchy tree in
31500 an irreducible region.
31501 (fix_bb_placements): Adjust caller.
31502 (fix_loop_placements): Likewise.
31504 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
31506 * config/avr/builtins.def (DEF_BUILTIN): Factor out
31507 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
31508 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
31509 Remove ID. Adjust comments.
31510 * config/avr/avr-c.c (avr_builtin_name): Remove.
31511 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
31512 * config/avr/avr.c (avr_tolower): New static function.
31513 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
31514 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
31515 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
31518 2013-01-17 Jan Hubicka <jh@suse.cz>
31520 PR tree-optimization/55273
31521 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
31523 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
31526 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
31527 store through atomic_store<mode>_1.
31528 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
31530 2013-01-17 Martin Jambor <mjambor@suse.cz>
31532 PR tree-optimizations/55264
31533 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
31534 for virtual methods.
31535 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
31536 virtual methods before inlining is over.
31537 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
31539 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
31542 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
31544 PR rtl-optimization/56005
31545 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
31546 pending reads for prefetch.
31548 2013-01-16 Ian Bolton <ian.bolton@arm.com>
31550 * config/aarch64/aarch64.md
31551 (*cstoresi_neg_uxtw): New pattern.
31552 (*cmovsi_insn_uxtw): New pattern.
31553 (*<optab>si3_uxtw): New pattern.
31554 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
31555 (*<optab>si3_insn_uxtw): New pattern.
31556 (*bswapsi2_uxtw): New pattern.
31558 2013-01-16 Richard Biener <rguenther@suse.de>
31560 * tree-inline.c (tree_function_versioning): Remove set but
31561 never used variable.
31563 2013-01-16 Richard Biener <rguenther@suse.de>
31565 PR tree-optimization/55964
31566 * tree-flow.h (rename_variables_in_loop): Remove.
31567 (rename_variables_in_bb): Likewise.
31568 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
31569 (copy_loop_before): Adjust and delete update-ssa status.
31570 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
31571 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
31572 (rename_variables_in_loop): Remove.
31573 (slpeel_update_phis_for_duplicate_loop): Likewise.
31574 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
31575 use available cfg machinery instead of duplicating it.
31576 Update PHI nodes and perform poor-mans SSA update here.
31577 (slpeel_tree_peel_loop_to_edge): Adjust.
31579 2013-01-16 Richard Biener <rguenther@suse.de>
31581 PR tree-optimization/54767
31582 PR tree-optimization/53465
31583 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
31584 (vrp_visit_phi_node): For PHI arguments coming via backedges
31585 drop all symbolical range information.
31586 (execute_vrp): Compute backedges.
31588 2013-01-16 Richard Biener <rguenther@suse.de>
31590 * doc/install.texi: Update CLooG and ISL requirements to
31593 2013-01-16 Christian Bruel <christian.bruel@st.com>
31596 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
31597 (broken_move): Handle UNSPECV_SP_SWITCH_B.
31598 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
31600 2013-01-16 DJ Delorie <dj@redhat.com>
31602 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
31603 (UNSPECV_SP_SWITCH_E): New.
31604 (sp_switch_1): Change to an unspec.
31605 (sp_switch_2): Change to an unspec. Don't use post-inc when we
31608 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
31610 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
31611 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
31612 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
31613 (expand_mem_thread_fence): Ditto.
31614 (expand_mem_signal_fence): Ditto.
31615 (expand_atomic_load): Ditto.
31616 (expand_atomic_store): Ditto.
31618 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
31620 PR rtl-optimization/55547
31621 PR rtl-optimization/53827
31624 * alias.c (memrefs_conflict_p): Set sizes to negative after
31627 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31630 * function.c (thread_prologue_and_epilogue_insns): Always
31631 add crtl->drap_reg to set_up_by_prologue.set, even if
31632 stack_realign_drap is false.
31634 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
31636 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
31637 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
31638 *call): Fix indention.
31640 2013-01-15 Tom de Vries <tom@codesourcery.com>
31643 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
31646 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
31648 PR rtl-optimization/55153
31649 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
31651 2013-01-15 Martin Jambor <mjambor@suse.cz>
31653 PR tree-optimization/55920
31654 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
31655 accesses as grp_to_be_debug_replaced.
31657 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31659 PR tree-optimization/55920
31660 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
31661 there is non-useless type conversion needed from debug rhs to lhs,
31662 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
31664 2013-01-15 Joseph Myers <joseph@codesourcery.com>
31665 Mikael Pettersson <mikpe@it.uu.se>
31668 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
31670 (ASM_OUTPUT_CASE_LABEL): Remove.
31671 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
31672 * final.c (shorten_branches): Update alignment of labels before
31673 jump tables if CASE_VECTOR_SHORTEN_MODE.
31675 2013-01-15 Richard Biener <rguenther@suse.de>
31678 * system.h: Do not include gmp.h for building host tools.
31680 2013-01-15 Richard Biener <rguenther@suse.de>
31682 PR middle-end/55882
31683 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
31684 account for bitpos when computing alignment.
31686 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
31688 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
31689 (ix86_target_macros_internal): Likewise.
31691 * config/i386/i386.c (m_CORE2I7): Removed.
31692 (m_CORE_HASWELL): New macro.
31693 (m_CORE_ALL): Likewise.
31694 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
31695 (initial_ix86_arch_features): Likewise.
31696 (processor_target_table): Initializations for Core avx2.
31697 (cpu_names): New names "core-avx2".
31698 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
31699 PROCESSOR_CORE_HASWELL.
31700 (ix86_issue_rate): New case.
31701 (ia32_multipass_dfa_lookahead): Likewise.
31702 (ix86_sched_init_global): Likewise.
31704 * config/i386/i386.h (TARGET_HASWELL): New macro.
31705 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
31706 (processor_type): New PROCESSOR_HASWELL.
31708 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31710 PR tree-optimization/55955
31711 * tree-vect-loop.c (vectorizable_reduction): Give up early on
31712 *SHIFT_EXPR and *ROTATE_EXPR codes.
31714 PR tree-optimization/48766
31715 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
31716 -ftrapv disable -fwrapv.
31718 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31721 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
31722 etc. to 1 and not to __flash.
31723 Use LL suffix for __INT24_MAX__ with -mint8.
31724 Use ULL suffix for __UINT24_MAX__ with -mint8.
31726 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31728 * config/avr/avr-arch.h
31729 (struct base_arch_s): Use typedef avr_arch_t instead.
31730 (struct arch_info_s): Use typedef avr_arch_info_t instead.
31731 (struct mcu_type_s): Use typedef avr_mcu_t instead.
31732 * config/avr/avr.c: Same.
31733 * config/avr/avr-devices.c: Same.
31734 * config/avr/driver-avr.c: Same.
31735 * config/avr/gen-avr-mmcu-texi.c: Same.
31736 * config/avr/avr-mcus.def: Adjust comment.
31738 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
31740 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
31741 * config/aarch64/iterators.md (VALLDI): New.
31743 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
31744 Andi Kleen <ak@linux.intel.com>
31747 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
31748 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
31751 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31753 * config/avr/avr-stdint.h: Remove trailing blanks.
31754 * config/avr/avr-log.h: Same.
31755 * config/avr/avr-arch.h: Same.
31756 * config/avr/avr-devices.c: Same.
31757 * config/avr/avr-dimode.md: Same.
31758 * config/avr/predicates.md: Same.
31759 * config/avr/avr-c.c: Same. And fix typo.
31761 * config/avr/avr-protos.h: Same. And:
31762 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
31763 (init_cumulative_args): Rename to avr_init_cumulative_args.
31764 (expand_prologue): Rename to avr_expand_prologue.
31765 (expand_epilogue): Rename to avr_expand_epilogue.
31766 (adjust_insn_length): Rename to avr_adjust_insn_length.
31767 (notice_update_cc): Rename to avr_notice_update_cc.
31768 (final_prescan_insn): Rename to avr_final_prescan_insn.
31769 * config/avr/avr.c: Same.
31770 * config/avr/avr.h: Same.
31771 * config/avr/avr.md: Remove trailing blanks.
31772 (prologue): Use avr_expand_prologue.
31773 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
31775 2013-01-14 Richard Biener <rguenther@suse.de>
31777 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
31778 verify_location, collect_subblocks): New functions.
31779 (verify_gimple_in_cfg): Verify that locations only reference
31780 BLOCKs in the functions BLOCK tree.
31782 2013-01-14 Richard Biener <rguenther@suse.de>
31784 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
31786 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
31788 (insert_out_of_ssa_copy_on_edge): Likewise.
31789 (rewrite_close_phi_out_of_ssa): Likewise.
31790 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
31792 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
31793 propagated constants.
31794 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
31797 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31799 * config/avr/avr-modes.def: Add GPL copyright notice.
31801 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
31803 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
31804 MEMMODEL_MASK to determine memory model.
31805 (atomic_store<mode>): Ditto from operands[2].
31806 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
31808 2013-01-13 Jakub Jelinek <jakub@redhat.com>
31811 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
31812 (fold_gimple_assign): Don't call unshare_expr here.
31813 (fold_ctor_reference): Call unshare_expr.
31815 2013-01-13 Terry Guo <terry.guo@arm.com>
31817 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
31818 * doc/fragments.texi: Document MULTILIB_REUSE.
31819 * gcc.c (multilib_reuse): New internal spec.
31820 (set_multilib_dir): Also search multilib from multilib_reuse.
31821 * genmultilib (tmpmultilib3): Refactor code.
31822 (tmpmultilib4): Ditto.
31823 (multilib_reuse): New multilib argument.
31825 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
31827 * Makefile.in: Update copyright.
31829 2013-01-12 Tom de Vries <tom@codesourcery.com>
31831 PR middle-end/55890
31832 * calls.c (expand_call): Check if arg_nr is valid.
31834 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
31836 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
31837 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
31838 documentation. Add missing '__' in front of
31839 __builtin_ia32_packssdw256.
31841 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31844 * config/s390/s390.c (s390_preferred_reload_class): Do not return
31845 NO_REGS for larl operands.
31846 (s390_reload_larl_operand): Use s390_load_address instead of
31849 2013-01-11 Richard Biener <rguenther@suse.de>
31851 * tree-cfg.c (verify_node_sharing_1): Split out from ...
31852 (verify_node_sharing): ... here.
31853 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
31855 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
31857 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
31858 Substitute TREECHECKING.
31859 * configure: Regenerate.
31860 * Makefile.in (TREECHECKING): New.
31862 2013-01-11 Richard Guenther <rguenther@suse.de>
31864 PR tree-optimization/44061
31865 * tree-vrp.c (extract_range_basic): Compute zero as
31866 value-range for __builtin_constant_p of function parameters.
31868 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
31870 Update copyright years.
31872 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
31874 PR rtl-optimization/55672
31875 * lra-eliminations.c (mark_not_eliminable): Permit addition with
31876 const to be eliminable.
31878 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
31880 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
31881 * configure: Regenerate.
31883 2013-01-10 Richard Biener <rguenther@suse.de>
31885 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
31887 2013-01-10 Richard Biener <rguenther@suse.de>
31890 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
31891 locations for virtual PHI arguments.
31892 (rewrite_update_phi_arguments): Likewise.
31894 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
31896 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
31899 2013-01-10 Jakub Jelinek <jakub@redhat.com>
31901 PR tree-optimization/55921
31902 * tree-complex.c (expand_complex_asm): New function.
31903 (expand_complex_operations_1): Call it for GIMPLE_ASM.
31905 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31908 * config/s390/s390.c (s390_symref_operand_p)
31909 (s390_loadrelative_operand_p): Merge the two functions.
31910 (s390_check_qrst_address, print_operand_address): Add parameters
31911 to s390_loadrelative_operand_p invokation.
31912 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
31913 (s390_reload_larl_operand, s390_secondary_reload): Use
31914 s390_loadrelative_operand_p instead of s390_symref_operand_p.
31915 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
31917 2013-01-09 Mike Stump <mikestump@comcast.net>
31919 * dse.c (record_store): Remove unnecessary assert.
31921 2013-01-09 Jan Hubicka <jh@suse.cz>
31923 PR tree-optimization/55569
31924 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
31925 * cfgloop.h (scale_loop_profile): Likewise.
31927 2013-01-09 Jan Hubicka <jh@suse.cz>
31930 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
31932 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
31934 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
31936 PR middle-end/55114
31937 * expr.h (maybe_emit_group_store): Declare.
31938 * expr.c (maybe_emit_group_store): New function.
31939 * builtins.c (expand_builtin_int_roundingfn): Call it.
31940 (expand_builtin_int_roundingfn_2): Likewise.
31942 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
31944 PR rtl-optimization/55829
31945 * lra-constraints.c (match_reload): Add code for absent output.
31946 (curr_insn_transform): Add code for reloads of matched inputs
31949 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
31951 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
31952 attribute of movddup insn to DF.
31953 (*vec_interleave_lowv2df): Ditto.
31954 (vec_dupv2df): Ditto.
31956 2013-01-09 Jan Hubicka <jh@suse.cz>
31958 PR tree-optimiation/55875
31959 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
31960 EVERY_ITERATION parameter.
31961 (number_of_iterations_exit): Check if exit is executed every iteration.
31962 (idx_infer_loop_bounds): Similarly here.
31963 (n_of_executions_at_most): Simplify
31964 to only test for cases where statement is dominated by the
31965 particular bound; handle correctly the "postdominance" test.
31966 (scev_probably_wraps_p): Use max loop iterations info
31967 as a global bound first.
31969 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
31970 Nick Clifton <nickc@redhat.com>
31972 * config/v850/v850.md (cbranchsf4): New pattern.
31973 (cstoresf4): New pattern.
31974 (cbranchdf4): New pattern.
31975 (cstoredf4): New pattern.
31976 (movsicc): Disallow floating point comparisons.
31977 (cmpsf_le_insn): Fix order of operators.
31978 (cmpsf_lt_insn): Likewise.
31979 (cmpsf_eq_insn): Likewise.
31980 (cmpdf_le_insn): Likewise.
31981 (cmpdf_lt_insn): Likewise.
31982 (cmpdf_eq_insn): Likewise.
31983 (cmpsf_ge_insn): Use LE comparison.
31984 (cmpdf_ge_insn): Likewise.
31985 (cmpsf_gt_insn): Use LT comparison.
31986 (cmpdf_gt_insn): Likewise.
31987 (cmpsf_ne_insn): Delete pattern.
31988 (cmpdf_ne_insn): Delete pattern.
31989 * config/v850/v850.c (v850_gen_float_compare): Use
31990 gen_cmpdf_eq_insn for NE comparison.
31991 (v850_float_z_comparison_operator)
31992 (v850_float_nz_comparison_operator): Move from here ...
31993 * config/v850/predicates.md: ... to here. Move GT and GE
31994 comparisons into v850_float_z_comparison_operator.
31995 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
31997 (v850_float_nz_comparison_operator): Likewise.
31999 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32001 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
32002 with calls to gen_insvsi/gen_insvdi.
32004 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
32006 * config/i386/i386.c (initial_ix86_tune_features): Set up
32007 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
32009 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
32010 Jakub Jelinek <jakub@redhat.com>
32012 PR tree-optimization/48189
32013 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
32014 If nitercst is 0, don't predict the exit edge.
32016 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
32018 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
32019 in asm_fprintf with reg_names.
32020 (aarch64_print_operand_address): Likewise.
32021 (aarch64_return_addr): Likewise.
32022 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
32024 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32026 * config/pa/pa.h (VAL_U6_BITS_P): Define.
32027 (INT_U6_BITS): Likewise.
32028 * config/pa/predicates.md (uint6_operand): New predicate.
32029 (shift5_operand, shift6_operand): Likewise.
32030 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
32032 (lshrdi3): Use shift6_operand.
32033 (shrpsi4, shrpdi4): New insn patterns.
32034 (extzv): Delete expander.
32035 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
32036 predicates in unamed zero extract patterns. Tighten common constraint.
32037 (extv): Delete expander.
32038 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
32039 predicates in unamed sign extract patterns. Tighten common constraint.
32040 (insv): Delete expander.
32041 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
32042 predicates in unamed insert patterns. Tighten common constraint.
32043 Change uint32_operand predicate to uint6_operand predicate in unamed
32044 DImode pattern to insert constant values of type 1...1xxxx.
32046 2013-01-04 Jan Hubicka <jh@suse.cz>
32048 PR tree-optimization/55823
32049 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
32052 2013-01-08 Jakub Jelinek <jakub@redhat.com>
32053 Uros Bizjak <ubizjak@gmail.com>
32055 PR rtl-optimization/55845
32056 * df-problems.c (can_move_insns_across): Stop scanning at
32057 volatile_insn_p source instruction or give up if
32058 across_from .. across_to range contains any volatile_insn_p
32061 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
32063 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
32064 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
32066 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
32067 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
32069 2013-01-08 Jakub Jelinek <jakub@redhat.com>
32072 * asan.c (asan_clear_shadow): New function.
32073 (asan_emit_stack_protection): Use it.
32075 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
32077 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
32078 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
32079 with tab instead of space.
32081 2013-01-08 Nick Clifton <nickc@redhat.com>
32083 * config/rl78/rl78.c (rl78_expand_prologue): Always select
32084 register bank 0 at the start of an interrupt handler.
32085 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
32088 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
32090 * config/aarch64/aarch64-simd.md
32091 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
32092 (aarch64_simd_bsl): Likewise.
32093 (aarch64_vcond_internal<mode>): Likewise.
32094 (vcond<mode><mode>): Likewise.
32095 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
32096 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
32098 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
32100 * config/aarch64/aarch64-builtins.c
32101 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
32103 2013-01-08 Martin Jambor <mjambor@suse.cz>
32106 * tree-sra.c (analyze_access_subtree): Return true also after
32107 potentially creating a debug-only replacement.
32109 2013-01-08 Jakub Jelinek <jakub@redhat.com>
32111 PR middle-end/55890
32112 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
32114 PR tree-optimization/54120
32115 * tree-vrp.c (range_fits_type_p): Don't allow
32116 src_precision < precision from signed vr to unsigned_p
32117 if vr->min or vr->max is negative.
32118 (simplify_float_conversion_using_ranges): Test can_float_p
32119 against CODE_FOR_nothing.
32121 2013-01-08 Jakub Jelinek <jakub@redhat.com>
32122 Richard Biener <rguenther@suse.de>
32124 PR middle-end/55851
32125 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
32126 types instead of just INTEGER_TYPE types.
32128 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
32130 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
32133 2013-01-07 Steve Ellcey <sellcey@mips.com>
32136 * config/mips/mips.opt: Change mad to mmad to match documentation.
32138 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
32141 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
32142 .progmemx.data now.
32144 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
32147 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
32148 (avr_addrspace_t): Add .section_name field.
32149 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
32151 (avr_addrspace): Same. Initialize .section_name. Remove last
32152 NULL entry. Put __memx into .progmemx.data.
32153 (progmem_section_prefix): Remove.
32154 (avr_asm_init_sections): No need to initialize progmem_section.
32155 (avr_asm_named_section): Use avr_addrspace[].section_name to get
32156 section name prefix.
32157 (avr_asm_select_section): Ditto. And use get_unnamed_section to
32158 retrieve the progmem section.
32159 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
32160 boundary to run over avr_addrspace[].
32161 (avr_register_target_pragmas): Ditto.
32163 2013-01-06 Jakub Jelinek <jakub@redhat.com>
32165 * varasm.c (output_constant_def_contents): For asan_protect_global
32166 protected strings, adjust DECL_ALIGN if needed, before testing for
32168 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
32169 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
32171 (output_object_block): For asan protected decls, emit asan padding
32172 after their contents.
32173 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
32174 (asan_finish_file): Test it here instead.
32176 2013-01-07 Nick Clifton <nickc@redhat.com>
32177 Matthias Klose <doko@debian.org>
32178 Doug Kwan <dougkwan@google.com>
32179 H.J. Lu <hongjiu.lu@intel.com>
32182 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
32184 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
32186 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
32188 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
32190 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
32192 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
32195 * doc/install.texi (Cross-Compiler-Specific Options): Document
32198 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
32200 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
32201 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
32202 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
32203 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
32204 vqmovun_high_s64): Fix source operand number and update copyright.
32206 2013-01-07 Richard Biener <rguenther@suse.de>
32208 PR middle-end/55890
32209 * gimple.h (gimple_call_builtin_p): New overload.
32210 * gimple.c (validate_call): New function.
32211 (gimple_call_builtin_p): Likewise.
32212 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
32213 Use gimple_call_builtin_p.
32214 (find_func_clobbers): Likewise.
32215 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
32216 (strlen_optimize_stmt): Likewise.
32218 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
32220 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
32221 (vld1q_dup_*): Likewise.
32222 (vld1_*): Likewise.
32223 (vld1q_*): Likewise.
32224 (vld1_lane_*): Likewise.
32225 (vld1q_lane_*): Likewise.
32227 2013-01-07 Richard Biener <rguenther@suse.de>
32229 * lto-streamer.h (LTO_minor_version): Bump to 2.
32231 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
32233 * config/aarch64/aarch64-protos.h
32234 (aarch64_const_double_zero_rtx_p): Rename to...
32235 (aarch64_float_const_zero_rtx_p): ...this.
32236 (aarch64_float_const_representable_p): New.
32237 (aarch64_output_simd_mov_immediate): Likewise.
32238 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
32239 move immediate case.
32240 * config/aarch64/aarch64.c
32241 (aarch64_const_double_zero_rtx_p): Rename to...
32242 (aarch64_float_const_zero_rtx_p): ...this.
32243 (aarch64_print_operand): Allow printing of new constants.
32244 (aarch64_valid_floating_const): New.
32245 (aarch64_legitimate_constant_p): Check for valid floating-point
32247 (aarch64_simd_valid_immediate): Likewise.
32248 (aarch64_vect_float_const_representable_p): New.
32249 (aarch64_float_const_representable_p): Likewise.
32250 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
32251 (aarch64_output_simd_mov_immediate): New.
32252 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
32253 (*movdf_aarch64): Likewise.
32254 * config/aarch64/constraints.md (Ufc): New.
32255 (Y): call aarch64_float_const_zero_rtx.
32256 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
32258 2013-01-07 Richard Biener <rguenther@suse.de>
32260 PR tree-optimization/55888
32261 PR tree-optimization/55862
32262 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
32263 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
32264 not if it is contained therein.
32266 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
32268 * config/avr/t-avr: Typo.
32270 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
32273 * config/avr/t-avr: Don't automatically rebuild
32274 $(srcdir)/config/avr/t-multilib
32275 $(srcdir)/config/avr/avr-tables.opt
32276 $(srcdir)/doc/avr-mmcu.texi
32277 (avr-mcus): New phony target to build them on request.
32278 (s-avr-mlib, s-avr-mmcu-texi): Remove.
32279 * avr/avr-mcus.def: Adjust comments.
32281 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
32283 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
32285 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
32287 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
32289 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
32291 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
32293 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
32295 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
32296 to generate profiling.
32297 * config/rs6000/aix64.h (LIB_SPEC): Same.
32299 2013-01-04 Andrew Pinski <apinski@cavium.com>
32301 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
32303 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
32305 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
32307 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
32309 (ix86_expand_move): Ditto.
32310 (ix86_zero_extend_to_Pmode): Ditto.
32311 (ix86_expand_call): Ditto.
32312 (ix86_expand_special_args_builtin): Ditto.
32313 (ix86_expand_builtin): Ditto.
32315 2013-01-04 Richard Biener <rguenther@suse.de>
32317 PR tree-optimization/55862
32318 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
32319 translating them through PHI nodes.
32321 2013-01-04 Martin Jambor <mjambor@suse.cz>
32323 PR tree-optimization/55755
32324 * tree-sra.c (sra_modify_assign): Do not check that an access has no
32325 children when trying to avoid producing a VIEW_CONVERT_EXPR.
32327 2013-01-04 Marek Polacek <polacek@redhat.com>
32329 PR middle-end/55859
32330 * opts.c (default_options_optimization): Clarify error message.
32332 2013-01-04 Richard Biener <rguenther@suse.de>
32334 PR middle-end/55863
32335 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
32338 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32341 * config/pa/pa.md (movsi): Revert previous change.
32342 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
32345 2013-01-03 Richard Henderson <rth@redhat.com>
32347 * config/i386/i386.c (ix86_expand_move): Always assign to op1
32348 after eliminating TLS symbols.
32350 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
32353 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
32354 * graphite-poly.c (debug_gmp_value): Likewise.
32356 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
32359 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
32360 selected code model, define __code_mode_small__, __code_model_medium__,
32361 __code_model_large__, __code_model_32__ or __code_model_kernel__.
32362 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
32363 xchg temporary register with %k. Declare temporary register as
32365 [__x86_64__]: For medium and large code models, preserve %rbx register.
32367 2013-01-03 Richard Biener <rguenther@suse.de>
32369 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
32370 (dump_subscript): Adjust.
32371 (finalize_ddr_dependent): Do not dump redundant info.
32372 (analyze_siv_subscript): Adjust.
32373 (subscript_dependence_tester): Likewise.
32374 (compute_affine_dependence): Likewise.
32376 2013-01-03 Richard Biener <rguenther@suse.de>
32379 2013-01-03 Richard Biener <rguenther@suse.de>
32381 PR tree-optimization/55857
32382 * tree-vect-stmts.c (vectorizable_load): Do not setup
32383 re-alignment for invariant loads.
32385 2013-01-02 Richard Biener <rguenther@suse.de>
32387 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
32388 invariant load do not generate a vector load from the scalar location.
32390 2013-01-03 Richard Biener <rguenther@suse.de>
32392 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
32393 for not vectorizing.
32394 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
32395 not build INDIRECT_REFs, call get_name once only.
32396 (vect_create_data_ref_ptr): Likewise. Dump base object kind
32397 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
32399 2013-01-03 Richard Biener <rguenther@suse.de>
32401 PR tree-optimization/55857
32402 * tree-vect-stmts.c (vectorizable_load): Do not setup
32403 re-alignment for invariant loads.
32405 2013-01-03 Richard Biener <rguenther@suse.de>
32408 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
32409 prefer a built-in decl.
32411 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32413 * gcc.c (process_command): Update copyright notice dates.
32414 * gcov.c (print_version): Likewise.
32415 * gcov-dump.c (print_version): Likewise.
32417 PR rtl-optimization/55838
32418 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
32419 iv0.step, iv1.step and step.
32421 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32422 Marc Glisse <marc.glisse@inria.fr>
32424 PR tree-optimization/55832
32425 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
32426 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
32427 integer_{one,zero}_node.
32429 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32432 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
32433 * var-tracking.c (reverse_op): Don't add reverse ops to
32434 VALUEs that have already
32435 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
32437 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
32439 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
32441 2013-01-02 Teresa Johnson <tejohnson@google.com>
32443 * dumpfile.c (dump_loc): Print filename with location.
32444 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
32445 new location_t parameter to emit complete unroll message with
32446 new dump framework.
32447 (canonicalize_loop_induction_variables): Compute loops location
32448 and pass to try_unroll_loop_completely.
32449 * loop-unroll.c (report_unroll_peel): New function.
32450 (peel_loops_completely): Use new dump format with location
32451 for main dumpfile message, and invoke report_unroll_peel on success.
32452 (decide_unrolling_and_peeling): Ditto.
32453 (decide_peel_once_rolling): Remove old dumpfile message subsumed
32454 by report_unroll_peel.
32455 (decide_peel_completely): Ditto.
32456 (decide_unroll_constant_iterations): Ditto.
32457 (decide_unroll_runtime_iterations): Ditto.
32458 (decide_peel_simple): Ditto.
32459 (decide_unroll_stupid): Ditto.
32460 * cfgloop.c (get_loop_location): New function.
32461 * cfgloop.h (get_loop_location): Declare.
32463 2013-01-02 Sriraman Tallam <tmsriram@google.com>
32465 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
32468 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32470 PR middle-end/55198
32471 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
32472 BLKmode objects when EXPAND_MEMORY is specified.
32474 2013-01-02 Sriraman Tallam <tmsriram@google.com>
32476 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
32478 (fold_builtin_cpu): Do not share cpu model decls across statements.
32480 2013-01-02 Jason Merrill <jason@redhat.com>
32483 * tree.c (build_array_type_1): Revert earlier change.
32485 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
32487 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
32489 * config/aarch64/aarch64-tune.md: Re-generate.
32491 2013-01-02 Richard Biener <rguenther@suse.de>
32493 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
32494 invariant load do not generate a vector load from the scalar location.
32496 2013-01-02 Richard Biener <rguenther@suse.de>
32499 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
32500 * configure: Regenerate.
32502 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
32504 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
32505 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
32506 (expand_builtin_int_roundingfn_2): Keep the original target around
32507 for the fallback case.
32509 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
32511 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
32512 to be clear for sign changes.
32514 2013-01-01 Jan Hubicka <jh@suse.cz>
32516 * ipa-inline-analysis.c: Fix formatting.
32518 2013-01-01 Jakub Jelinek <jakub@redhat.com>
32520 PR tree-optimization/55831
32521 * tree-vect-loop.c (get_initial_def_for_induction): Use
32522 gsi_after_labels instead of gsi_start_bb.
32524 Copyright (C) 2013 Free Software Foundation, Inc.
32526 Copying and distribution of this file, with or without modification,
32527 are permitted in any medium without royalty provided the copyright
32528 notice and this notice are preserved.