1 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
3 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
6 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
8 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags
10 * configure: Regenerate.
12 2013-10-11 Jakub Jelinek <jakub@redhat.com>
14 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
15 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
16 formatting fixes, use pp_colon instead of pp_character (..., ':'),
17 similarly pp_right_paren.
18 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
19 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
20 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
21 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
23 (omp_declare_simd_clauses_equal,
24 omp_remove_redundant_declare_simd_attrs): New functions.
25 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
26 (walk_tree_1): Handle new OpenMP 4.0 clauses.
27 * tree.h (OMP_LOOP_CHECK): Define.
28 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
29 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
30 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
31 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
32 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
33 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
34 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
35 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
36 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
37 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
38 OMP_CLAUSE_SIMDLEN_EXPR): Define.
39 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
40 (omp_remove_redundant_declare_simd_attrs): New prototype.
41 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
42 GIMPLE_OMP_TEAMS): New codes.
43 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
44 * omp-low.c (struct omp_context): Add cancel_label and cancellable
46 (target_nesting_level): New variable.
47 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
48 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
49 for collapse > 1 static schedule unless ordered.
50 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
51 (determine_parallel_type): Adjust get_ws_args_for caller.
52 (install_var_field): Handle mask & 4 for double indirection.
53 (scan_sharing_clauses): Ignore shared clause on teams construct.
54 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
55 (create_omp_child_function): If inside target or declare target
56 constructs, set "omp declare target" attribute on the child
58 (find_combined_for): New function.
59 (scan_omp_parallel): Handle combined loops.
60 (scan_omp_target, scan_omp_teams): New functions.
61 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
62 restrictions and set ctx->cancellable for cancellable constructs.
63 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
64 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
65 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
66 (build_omp_barrier): Add lhs argument, return gimple rather than
68 (omp_clause_aligned_alignment): New function.
69 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
71 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
72 on teams constructs. Handle user defined reductions and new
74 (lower_reduction_clauses): Don't set placeholder to address of ref
75 if it has already the right type.
76 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
77 (expand_parallel_call): Use the new non-_start suffixed builtins,
78 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
79 and GOMP_parallel_end after the call.
80 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
81 (expand_omp_for_init_counts): Handle combined loops.
82 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
84 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
85 end of cancellable loops.
86 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
87 Likewise. Handle collapse > 1 loops.
88 (expand_omp_simd): Handle combined loops.
89 (expand_omp_for): Add inner_stmt argument, adjust callers of
90 expand_omp_for* functions, use expand_omp_for_static*chunk even
91 for collapse > 1 unless ordered.
92 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
93 of cancellable sections.
94 (expand_omp_single): Remove need_barrier variable, just rely on
95 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
96 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
97 (expand_omp_atomic_load, expand_omp_atomic_store,
98 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
99 (expand_omp_target): New function.
100 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
101 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
102 (build_omp_regions_1): Immediately close region for
103 GF_OMP_TARGET_KIND_UPDATE.
104 (maybe_add_implicit_barrier_cancel): New function.
105 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
107 (lower_omp_single): Likewise. Add clobber after the barrier.
108 (lower_omp_taskgroup): New function.
109 (lower_omp_for): Handle combined loops. Adjust
110 lower_rec_input_clauses caller. Handle cancellation.
111 (lower_depend_clauses): New function.
112 (lower_omp_taskreg): Lower depend clauses. Adjust
113 lower_rec_input_clauses caller. Add clobber after the call. Handle
115 (lower_omp_target, lower_omp_teams): New functions.
116 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
117 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
118 and GOMP_cancellation_point calls.
119 (lower_omp): Fold stmts inside of target region.
120 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
121 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
122 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
123 (BT_FN_VOID_OMPFN_PTR_UINT,
124 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
125 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
126 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
127 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
128 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
129 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
130 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
131 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
132 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
133 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
134 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
135 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
136 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
137 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
138 BUILT_IN_GOMP_PARALLEL_END.
139 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
140 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
141 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
142 GF_OMP_FOR_KIND_DISTRIBUTE.
143 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
144 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
145 (dump_gimple_omp_return): Print lhs if it has any.
146 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
147 gimple_omp_atomic_seq_cst_p.
148 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
149 and GIMPLE_OMP_TEAMS.
150 * langhooks.c (lhd_omp_mappable_type): New function.
151 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
152 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
154 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
155 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
156 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
158 (struct gimplify_omp_ctx): Add combined_loop field.
159 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
160 on stmts inside of target region.
161 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
162 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
163 ORT_TARGET and ORT_TARGET_DATA.
164 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
166 (omp_notice_threadprivate_variable): Complain about threadprivate
167 variables in target region.
168 (omp_notice_variable): Complain about vars with non-mappable type
169 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
170 (omp_check_private): Ignore ORT_TARGET* regions.
171 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
172 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
173 (find_combined_omp_for): New function.
174 (gimplify_omp_for): Handle gimplification of combined loops.
175 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
177 (gimplify_omp_target_update): New function.
178 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
179 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
180 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
181 (gimplify_body): If fndecl has "omp declare target" attribute, add
182 implicit ORT_TARGET context around it.
183 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
184 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
185 * tree-nested.c (convert_nonlocal_reference_stmt,
186 convert_local_reference_stmt, convert_gimple_call): Handle
187 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
188 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
189 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
190 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
191 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
192 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
193 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
194 BUILT_IN_GOMP_LOOP_END_CANCEL,
195 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
196 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
197 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
198 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
199 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
200 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
201 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
202 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
203 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
204 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
205 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
206 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
207 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
208 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
209 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
210 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
211 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
212 gimple_build_omp_teams): New functions.
213 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
214 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
215 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
216 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
217 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
218 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
219 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
220 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
221 GF_OMP_ATOMIC_SEQ_CST): New.
222 (gimple_build_omp_taskgroup, gimple_build_omp_target,
223 gimple_build_omp_teams): New prototypes.
224 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
225 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
226 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
227 GIMPLE_OMP_SINGLE as end of range.
228 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
229 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
230 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
231 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
232 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
233 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
234 gimple_omp_target_kind, gimple_omp_target_set_kind,
235 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
236 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
237 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
238 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
239 gimple_omp_teams_set_clauses): New inlines.
240 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
241 and GIMPLE_OMP_TASKGROUP.
242 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause
244 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
245 enum omp_clause_proc_bind_kind): New.
246 (union omp_clause_subcode): Add depend_kind, map_kind and
247 proc_bind_kind fields.
248 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
249 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
250 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
251 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
252 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
254 2013-10-10 Teresa Johnson <tejohnson@google.com>
256 * predict.c (tree_estimate_probability): Add new parameter
257 for estimate_bb_frequencies.
258 (estimate_bb_frequencies): Add new parameter to force estimation.
259 (rebuild_frequencies): When max frequency in function is small,
260 recompute counts from frequencies.
261 * predict.h (estimate_bb_frequencies): New parameter.
263 2013-10-10 David Malcolm <dmalcolm@redhat.com>
265 * ipa-inline.c (ipa_inline): Fix leak of "order" when
266 optimizations are disabled.
268 2013-10-10 David Malcolm <dmalcolm@redhat.com>
270 * coverage.c (coverage_finish): Fix leak of da_file_name.
272 2013-10-10 Jan Hubicka <jh@suse.cz>
274 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
275 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
278 2013-10-10 Jakub Jelinek <jakub@redhat.com>
281 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
282 if any labels are in FALLTHRU_BB, use a special label emitted
283 immediately after the asm goto insn rather than label_rtx
285 (expand_asm_stmt): Adjust caller.
286 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
287 edge if the last insn in predecessor is a jump with single successor,
288 but it isn't simplejump_p.
290 2013-10-10 Richard Biener <rguenther@suse.de>
292 PR tree-optimization/58656
293 * tree-ssa-pre.c (phi_translate): Do not cache failed
296 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
298 * gimplify.c: Include expr.h and tm_p.h for targets with special va-arg
299 padding requirements.
301 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
303 * tree-flow.h: Move some prototypes to gimple.h.
304 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
305 * gimple.h: Relocate some prototypes from tree-flow.h
306 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
308 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
309 (build_va_arg_indirect_ref): Relocate and make static.
310 (std_gimplify_va_arg_expr): Relocate here.
311 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
312 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
314 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
316 * doc/md.texi: Document the mnemonic attribute.
318 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
321 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
322 (set_attr_alternative x ...) when searching for user defined
325 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
327 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
329 2013-10-09 Easwaran Raman <eraman@google.com>
331 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
332 * cfgexpand.c (defer_stack_allocation): ...use here
333 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
335 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
337 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
338 (operand_equal_for_value_replacement): New function, extracted from
339 value_replacement and enhanced to catch more cases.
340 (value_replacement): Use operand_equal_for_value_replacement.
342 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
344 * loop-doloop.c (doloop_modify, doloop_optimize): Use
345 get_max_loop_iterations.
347 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
349 * config/arm/aarch-common.c (arm_early_load_addr_dep):
350 Place comment above function.
352 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
354 * tree-flow.h: Remove all remaining prototypes, enums and structs that
355 are not related to tree-cfg.c.
356 * tree-ssa-address.h: New file. Relocate prototypes.
357 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
358 (addr_for_mem_ref): New. Combine call to get_address_description and
359 return addr_for_mem_ref.
360 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
361 * tree-ssa-live.h: Adjust prototypes.
362 * passes.c: Include tree-ssa-live.h.
363 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
364 * graphite.c (graphite_transform_loops): Make static.
365 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
366 make_pass_graphite, pass_data_graphite_transforms,
367 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
368 * ipa-pure-const.c (warn_function_noreturn): Make static.
369 (execute_warn_function_noreturn, gate_warn_function_noreturn,
370 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
371 Relocate from tree-cfg.c
372 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
374 (execute_warn_function_noreturn, gate_warn_function_noreturn,
375 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
376 Move to ipa-pure-const.c.
377 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg): Relocate
378 from tree-optimize.c.
379 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
380 make_pass_fixup_cfg): Move to tree-cfg.c.
381 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
382 Relocate some prototypes.
383 * tree-data-ref.h (tree_check_data_deps) Add prototype.
384 * tree-dump.c (dump_function_to_file): Remove prototype.
385 Add tree-flow.h to the include file.
386 * tree-dump.h: Remove prototype.
387 * tree-parloops.h: New File. Add prototypes.
388 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
389 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
390 from tree-ssa-loop.c.
391 * tree-predcom.c (run_tree_predictive_commoning,
392 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
393 Relocate here from tree-ssa-loop.c.
394 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
395 ssa_name_values.release ().
396 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
397 (ssa_name_values): Relocate from tree-flow.h.
398 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
399 * tree-ssa-loop.c (run_tree_predictive_commoning,
400 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
401 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
402 make_pass_graphite, pass_data_graphite_transforms,
403 make_pass_graphite_transforms, gate_tree_parallelize_loops,
404 tree_parallelize_loops, pass_data_parallelize_loops,
405 make_pass_parallelize_loops): Move to other files.
406 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
408 * tree.h: Remove prototypes from tree-address.c.
410 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
412 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
413 (struct int_tree_map): Move to tree-hasher.h
414 (SCALE, LABEL, PERCENT): Move to gimple.h
415 * tree-flow-inline.h: Delete. Move functions to other files.
416 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
417 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
418 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
419 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
420 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
421 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
422 * tree-hasher.h: Don't include tree-flow.h.
423 (struct int_tree_map): Relocate from tree-flow.h.
424 * tree-sra.c (contains_view_convert_expr_p): Relocate from
425 tree-flow-inline.h and make static.
426 * tree-ssa-alias.h (ranges_overlap_p): Relocate from tree-flow-inline.h.
427 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
428 tree-flow-inline.h and make static.
429 * tree.h (is_global_var, may_be_aliased): Relocate from
431 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
432 * value-prof.c: No longer include tree-flow-inline.h.
433 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
435 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
437 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
438 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
439 (enum move_pos): Move to tree-ssa-loop-im.h
440 * cfgloop.h: Move some prototypes.
441 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
442 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
443 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
444 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
445 (loop_containing_stmt): Relocate from tree-flow-inline.h.
446 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
447 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
448 (enum move_pos): Relocate here.
449 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
451 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
452 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim, make_pass_lim):
453 Relocate here from tree-ssa-loop.c.
454 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
456 (loop_edge_to_cancel, unloop_loops): Make static.
457 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
458 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
459 (tree_complete_unroll, gate_tree_complete_unroll,
460 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
461 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
462 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
463 * tree-ssa-loop-ivopts.c: Remove local prototypes.
464 (stmt_invariant_in_loop_p): Remove unused function.
465 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
466 * tree-ssa-loop-manip.h: New file. Add prototypes.
467 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
468 (gcov_type_to_double_int): Move to cfgloop.h.
469 (double_int_cmp, bound_index,
470 estimate_numbers_of_iterations_loop): Make static.
471 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
472 (max_loop_iterations): Factor out get_max_loop_iterations.
473 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
475 * tree-ssa-loop-niter.h: New file. Add prototypes.
476 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
477 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
478 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
479 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
480 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
481 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
482 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
483 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
484 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
485 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
486 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
487 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
488 pass_data_complete_unroll, make_pass_complete_unroll,
489 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
490 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
491 tree-ssa-loop-ivcanon.c.
492 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
493 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
494 tree-ssa-loop-prefetch.c.
495 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
497 (get_lsm_tmp_name): Relocate and add suffix parameter.
498 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
499 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
500 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
501 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
502 (get_estimated_loop_iterations): Factor out accessor from
503 estimated_loop_iterations in tree-ssa-loop-niter.c.
504 (get_max_loop_iterations): Factor out accessor from _max_loop_iterations
506 * loop-unroll.c (decide_unroll_constant_iterations,
507 decide_unroll_runtime_iterations, decide_peel_simple,
508 decide_unroll_stupid): Use new get_* accessors.
510 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
512 PR tree-optimization/20318
513 * doc/extend.texi (returns_nonnull): New function attribute.
514 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
516 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
517 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
519 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
522 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
523 false if both components are bitfields.
525 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
527 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
528 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
529 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
530 * config/aarch64/aarch64.h
531 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
532 * config/aarch64/aarch64-simd-builtins.def
533 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
535 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
537 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
539 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
541 * config/aarch64/arm_neon.h (vdiv_f64): Added.
543 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
545 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
546 (vneg_f64): New intrinsic.
547 (vneg_s8): Asm replaced with C.
548 (vneg_s16): Likewise.
549 (vneg_s32): Likewise.
550 (vneg_s64): New intrinsic.
551 (vnegq_f32): Asm replaced with C.
552 (vnegq_f64): Likewise.
553 (vnegq_s8): Likewise.
554 (vnegq_s16): Likewise.
555 (vnegq_s32): Likewise.
556 (vnegq_s64): Likewise.
558 2013-10-09 Renlin Li <Renlin.Li@arm.com>
560 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
562 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
564 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
565 packed stack special handling.
566 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
567 back to fixed stack slots for FPRs saved due to stdarg.
569 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
571 * config/s390/s390.c (s390_frame_info): Restructure function.
573 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
575 * config/s390/s390.c (struct s390_frame_layout): New field
577 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
578 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
579 regs_ever_clobbered to char*.
580 (s390_regs_ever_clobbered): Check crtl->saves_all_registers
581 instead of cfun->has_nonlocal_label. Ignore frame related restore
583 (s390_register_info): Enable FPR save slots. Move/Copy some
584 functionality into ...
585 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
586 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
588 (s390_frame_info): Do gpr slot allocation here now. stdarg does
589 not imply a stack frame.
590 (s390_init_frame_layout): Remove variable clobbered_regs.
591 (s390_update_register_info): Remove function.
592 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
593 cannot be used for register renaming.
594 (s390_hard_regno_scratch_ok): New function.
595 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
596 (s390_initial_elimination_offset): Change offset calculation of
597 the return address pointer.
598 (save_gprs): Deal with only r6 being saved from the call-saved
600 (restore_gprs): Set frame related flag.
601 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New
603 (s390_emit_prologue): Call s390_register_info instead of
604 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
605 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
606 (s390_optimize_prologue): Call s390_optimize_register_info.
607 Try to remove also FPR slot save/restore INSNs. Remove frame
608 related flags from restore INSNs.
610 2013-10-08 DJ Delorie <dj@redhat.com>
612 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
615 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
616 avoid conflict with the MI use of %c.
617 * config/rl78/rl78-real.md: change %c to %C throughout.
618 * config/rl78/rl78-virt.md: Likewise.
620 2013-10-08 Jan Hubicka <jh@suse.cz>
622 * config/i386/i386.c (ix86_option_override_internal): Switch
623 to SSE math for -ffast-math when target ISA supports SSE2.
625 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
627 * tree-flow.h: Remove some prototypes.
628 * tree.h: Remove some protypes, add a couple.
629 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
630 using_eh_for_cleanups_p): Add interface routines for front ends.
631 * tree-eh.h: New file. Add protoptyes.
632 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
633 (add_stmt_to_eh_lp_fn): Make static.
634 (lower_try_finally): Use new using_eh_for_cleanups_p.
635 * emit-rtl.c: Include tree-eh.h.
636 * gimple.h: Include tree-eh.h.
638 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
640 PR tree-optimization/58480
641 * tree-vrp.c (infer_nonnull_range): New function.
642 (infer_value_range): Call infer_nonnull_range.
644 2013-10-08 Dehao Chen <dehao@google.com>
646 PR tree-optimization/58619
647 * tree-inline.c (copy_phis_for_bb): Combine location data
650 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
653 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
654 RTX_FRAME_RELATED_P on INSN.
656 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
658 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
659 (altivec_expand_vec_perm_const): Call it.
661 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
663 * config/rs6000/vector.md (mov<mode>): Emit permuted move
664 sequences for LE VSX loads and stores at expand time.
665 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
667 * config/rs6000/rs6000.c (rs6000_const_vec): New.
668 (rs6000_gen_le_vsx_permute): New.
669 (rs6000_gen_le_vsx_load): New.
670 (rs6000_gen_le_vsx_store): New.
671 (rs6000_gen_le_vsx_move): New.
672 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
673 (*vsx_le_perm_load_v4si): New.
674 (*vsx_le_perm_load_v8hi): New.
675 (*vsx_le_perm_load_v16qi): New.
676 (*vsx_le_perm_store_v2di): New.
677 (*vsx_le_perm_store_v4si): New.
678 (*vsx_le_perm_store_v8hi): New.
679 (*vsx_le_perm_store_v16qi): New.
680 (*vsx_xxpermdi2_le_<mode>): New.
681 (*vsx_xxpermdi4_le_<mode>): New.
682 (*vsx_xxpermdi8_le_V8HI): New.
683 (*vsx_xxpermdi16_le_V16QI): New.
684 (*vsx_lxvd2x2_le_<mode>): New.
685 (*vsx_lxvd2x4_le_<mode>): New.
686 (*vsx_lxvd2x8_le_V8HI): New.
687 (*vsx_lxvd2x16_le_V16QI): New.
688 (*vsx_stxvd2x2_le_<mode>): New.
689 (*vsx_stxvd2x4_le_<mode>): New.
690 (*vsx_stxvd2x8_le_V8HI): New.
691 (*vsx_stxvd2x16_le_V16QI): New.
693 2013-10-07 Renlin Li <Renlin.Li@arm.com>
695 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
697 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
699 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
700 loop to work also for 31bit ABI.
701 Save the stack pointer for frame_size > 0.
703 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
705 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
706 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
707 constraint letters from expanders.
708 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
709 retry count to general_operand.
710 ("tabort"): Give operand 0 a mode.
711 ("tabort_1"): Add mode and constraint letter for operand 0.
712 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
714 2013-10-04 Jeff Law <law@redhat.com>
716 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
718 * tree-ssa-threadedge.c (thread_through_normal_block): Broken out of ...
719 (thread_across_edge): Here. Call it.
721 2013-10-04 Cary Coutant <ccoutant@google.com>
723 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
724 discarding a location list expression (or a piece of one).
726 2013-10-03 Jan Hubicka <jh@suse.cz>
728 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
729 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
730 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
732 2013-10-03 Jan Hubicka <jh@suse.cz>
734 * config/i386/i386.c (ix86_option_override_internal): Do not enable
735 accumulate-outgoing-args when producing unwind info.
737 2013-10-03 Wei Mi <wmi@google.com>
739 * lra-constraints.c (insert_move_for_subreg): New function
740 extracted from simplify_operand_subreg.
741 (simplify_operand_subreg): Add reload for paradoxical subreg.
743 2013-10-03 Rong Xu <xur@google.com>
745 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
746 the candidate of builtin_expect such that we should fix the
747 size/time estimation.
748 (estimate_function_body_sizes): Do the acutally size/time fix-up
751 2013-10-03 Rong Xu <xur@google.com>
753 * predict.c (tree_predict_by_opcode): Get the probability
754 for builtin_expect from param builtin_expect_probability.
755 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
756 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
757 * doc/invoke.texi: Add documentation for builtin-expect-probability.
759 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
762 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
763 * calls.c (alloca_call_p): Use get_callee_fndecl.
764 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
765 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
767 (vrp_visit_stmt): Remove duplicated code.
769 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
771 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
772 ceildf2, btruncdf2, instead of vsx_* name.
774 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
775 iterators to only do V2DF and V4SF here. Move the DF code to
776 rs6000.md where it is combined with SF mode. Replace <VSv> with
777 just 'v' since only vector operations are handled with these insns
778 after moving the DF support to rs6000.md.
779 (vsx_sub<mode>3): Likewise.
780 (vsx_mul<mode>3): Likewise.
781 (vsx_div<mode>3): Likewise.
782 (vsx_fre<mode>2): Likewise.
783 (vsx_neg<mode>2): Likewise.
784 (vsx_abs<mode>2): Likewise.
785 (vsx_nabs<mode>2): Likewise.
786 (vsx_smax<mode>3): Likewise.
787 (vsx_smin<mode>3): Likewise.
788 (vsx_sqrt<mode>2): Likewise.
789 (vsx_rsqrte<mode>2): Likewise.
790 (vsx_fms<mode>4): Likewise.
791 (vsx_nfma<mode>4): Likewise.
792 (vsx_copysign<mode>3): Likewise.
793 (vsx_btrunc<mode>2): Likewise.
794 (vsx_floor<mode>2): Likewise.
795 (vsx_ceil<mode>2): Likewise.
796 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
797 (vsx_sminsf3): Likewise.
798 (vsx_fmadf4): Likewise.
799 (vsx_fmsdf4): Likewise.
800 (vsx_nfmadf4): Likewise.
801 (vsx_nfmsdf4): Likewise.
802 (vsx_cmpdf_internal1): Likewise.
804 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
805 simpler to select whether a target has SPE or traditional floating
806 point support in iterators.
807 (TARGET_DF_SPE): Likewise.
808 (TARGET_SF_FPR): Likewise.
809 (TARGET_DF_FPR): Likewise.
810 (TARGET_SF_INSN): Macros to say whether floating point support
811 exists for a given operation for expanders.
812 (TARGET_DF_INSN): Likewise.
814 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
815 combining of SF/DF mode operations, using both traditional and VSX
823 (abs<mode>2): Combine SF/DF modes using traditional floating point
824 instructions. Add support for using the upper DF registers with
825 VSX support, and SF registers with power8-vector support. Update
826 expanders for operations supported by both the SPE and traditional
827 floating point units.
828 (abs<mode>2_fpr): Likewise.
829 (nabs<mode>2): Likewise.
830 (nabs<mode>2_fpr): Likewise.
831 (neg<mode>2): Likewise.
832 (neg<mode>2_fpr): Likewise.
833 (add<mode>3): Likewise.
834 (add<mode>3_fpr): Likewise.
835 (sub<mode>3): Likewise.
836 (sub<mode>3_fpr): Likewise.
837 (mul<mode>3): Likewise.
838 (mul<mode>3_fpr): Likewise.
839 (div<mode>3): Likewise.
840 (div<mode>3_fpr): Likewise.
841 (sqrt<mode>3): Likewise.
842 (sqrt<mode>3_fpr): Likewise.
844 (rsqrt<mode>2): Likewise.
845 (cmp<mode>_fpr): Likewise.
846 (smax<mode>3): Likewise.
847 (smin<mode>3): Likewise.
848 (smax<mode>3_vsx): Likewise.
849 (smin<mode>3_vsx): Likewise.
850 (negsf2): Delete SF operations that are merged with DF.
857 (fmasf4_fpr): Likewise.
858 (fmssf4_fpr): Likewise.
859 (nfmasf4_fpr): Likewise.
860 (nfmssf4_fpr): Likewise.
862 (rsqrtsf_internal1): Likewise.
865 (cmpsf_internal1): Likewise.
866 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
867 (negdf2): Delete DF operations that are merged with SF.
875 (rsqrtdf_internal1): Likewise.
876 (fmadf4_fpr): Likewise.
877 (fmsdf4_fpr): Likewise.
878 (nfmadf4_fpr): Likewise.
879 (nfmsdf4_fpr): Likewise.
883 (cmpdf_internal1): Likewise.
884 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
885 (btrunc<mode>2): Delete separate expander, and combine with the
886 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
887 (btrunc<mode>2_fpr): Likewise.
888 (ceil<mode>2): Likewise.
889 (ceil<mode>2_fpr): Likewise.
890 (floor<mode>2): Likewise.
891 (floor<mode>2_fpr): Likewise.
892 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
893 Add support for using the upper registers with VSX and
894 power8-vector. Move insns to be closer to the define_expands. On
895 VSX systems, prefer the traditional form of FMA over the VSX
896 version, since the traditional form allows the target not to
897 overlap with the inputs.
898 (fms<mode>4_fpr): Likewise.
899 (nfma<mode>4_fpr): Likewise.
900 (nfms<mode>4_fpr): Likewise.
902 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
903 Richard Earnshaw <richard.earnshaw@arm.com>
905 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
906 (struct mult_cost_table): Likewise.
907 (struct mem_cost_table): Likewise.
908 (struct fp_cost_table): Likewise.
909 (struct vector_cost_table): Likewise.
910 (cpu_cost_table): Likewise.
911 * config/arm/arm.opt (mold-rts-costs): New option.
912 (mnew-generic-costs): Likewise.
913 * config/arm/arm.c (generic_extra_costs): New table.
914 (cortexa15_extra_costs): Likewise.
915 (arm_slowmul_tune): Use NULL as new costs.
916 (arm_fastmul_tune): Likewise.
917 (arm_strongarm_tune): Likewise.
918 (arm_xscale_tune): Likewise.
919 (arm_9e_tune): Likewise.
920 (arm_v6t2_tune): Likewise.
921 (arm_cortex_a5_tune): Likewise.
922 (arm_cortex_a9_tune): Likewise.
923 (arm_v6m_tune): Likewise.
924 (arm_fa726te_tune): Likewise.
925 (arm_cortex_a15_tune): Use cortex15_extra_costs.
926 (arm_cortex_tune): Use generict_extra_costs.
927 (shifter_op_p): New function.
928 (arm_unspec_cost): Likewise.
929 (LIBCALL_COST): Define.
930 (arm_new_rtx_costs): New function.
931 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
932 table is available. Use old costs otherwise unless mnew-generic-costs
934 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
935 (cpu_cost_table): Declare.
937 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
940 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
941 (*subs_mul_imm_<mode>)
942 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
943 (*sub_<shift>_<mode>)
944 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
947 2013-10-03 Ian Bolton <ian.bolton@arm.com>
949 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
951 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
953 2013-10-02 Teresa Johnson <tejohnson@google.com>
955 * predict.c (probably_never_executed): New function.
956 (probably_never_executed_bb_p): Invoke probably_never_executed.
957 (probably_never_executed_edge_p): Ditto.
958 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
959 Treat profile insanities conservatively.
961 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
963 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
965 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
967 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
968 use it. Use smaller increase for scratch. Don't increase reject
969 for early clobber scratch.
970 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
971 setting eliminated regs except setting fp from hfp.
972 (lra_eliminate): Check lra_insn_recog_data on NULL.
974 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
977 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
978 setting -mvsx-timode by default until the underlying problem is fixed.
979 (RS6000_CPU, power7 defaults): Likewise.
981 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
983 * config/x-linux (host-linux.o): Remove header dependencies.
984 Use $(COMPILE) and $(POSTCOMPILE).
985 * config/t-linux-android (linux-android.o): Ditto.
987 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
989 * Makefile.in (expmed.o-warn): Remove.
991 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
993 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
994 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
996 2013-10-02 Teresa Johnson <tejohnson@google.com>
998 * dojump.c (do_jump_1): Divide probability between
999 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
1001 2013-10-02 Tom Tromey <tromey@redhat.com>
1003 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
1005 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1007 * tree-flow.h: Remove some prototypes.
1008 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
1009 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
1010 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
1011 mark_virtual_phi_result_for_renaming): Relocate here.
1012 * tree-into-ssa.h: Add prototypes.
1013 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
1014 single_pred_before_succ_order.
1015 (blocks_in_phiopt_order): Rename and move to cfganal.c.
1016 (nonfreeing_call_p) Move to gimple.c.
1017 * cfganal.c (single_pred_before_succ_order): Move and renamed from
1019 * basic-block.h (single_pred_before_succ_order): Add prototype.
1020 * gimple.c (nonfreeing_call_p): Relocate here.
1021 * gimple.h: Add prototype.
1022 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
1023 * tree-ssa-dom.h: New file. Relocate prototypes here.
1024 * tree-ssa.h: Include tree-ssa-dom.h.
1026 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
1028 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
1029 Use $(COMPILE) and $(POSTCOMPILE).
1031 * config/alpha/x-alpha (driver-alpha.o): Ditto.
1033 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1035 * tree-flow.h: Remove some prototypes.
1036 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
1037 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
1038 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
1039 propagate_tree_value*): Move from here to...
1040 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
1041 propagate_tree_value*): Relocate here.
1042 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
1043 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
1044 * gimple-fold.c: Remove gimple-fold.h from include list.
1045 * tree-vrp.c: Remove gimple-fold.h from include list.
1046 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
1047 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
1048 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
1049 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
1050 * sese.c: Add tree-ssa-propagate.h to include list.
1052 2013-10-02 Richard Biener <rguenther@suse.de>
1054 * tree-loop-distribution.c: Include tree-vectorizer.h for
1056 (enum partition_kind): Remove PKIND_REDUCTION.
1057 (struct partition_s): Remove has_writes member, add reduction_p member.
1058 (partition_alloc): Adjust.
1059 (partition_builtin_p): Likewise.
1060 (partition_has_writes): Remove.
1061 (partition_reduction_p): New function.
1062 (partition_merge_into): Likewise.
1063 (generate_code_for_partition): Commonize builtin partition
1065 (rdg_cannot_recompute_vertex_p): Remove.
1066 (already_processed_vertex_p): Likewise.
1067 (rdg_flag_vertex): Do not set has_writes.
1068 (classify_partition): Adjust.
1069 (rdg_build_partitions): Do not set has_writes, treat all
1070 partitions as useful.
1071 (distribute_loop): Record number of library calls generated. Adjust.
1072 (tree_loop_distribution): Report number of loops and library
1073 calls generated as opt-info.
1075 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1077 * tree-flow.h: Include new .h files. Move prototypes.
1078 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
1079 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
1080 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
1081 * tree-pretty-print.h: Add prototypes from tree-flow.h.
1082 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
1083 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
1084 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
1085 * tree.h (get_ref_base_and_extent): Move prototype out.
1086 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
1088 * gimple-low.h: New File. Add prototypes from tree-flow.h.
1089 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
1090 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
1091 * tree-scalar-evolution.c: Include tree.h.
1092 * sese.c: Include tree.h.
1093 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
1094 * dwarf2out.c: Include tree-dfa.h.
1095 * tree-chrec.c: Include tree.h.
1096 * tree-data-ref.c: Include tree.h.
1098 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
1100 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
1103 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1105 * config/t-sol2 (sol2-c.o): Remove header dependencies.
1106 Use $(COMPILE) and $(POSTCOMPILE).
1107 (sol2-cxx.o): Likewise.
1108 (sol2-stubs.o): Likewise.
1110 * config/x-solaris (host-solaris.o): Likewise.
1112 * config/sparc/t-sparc (sparc.o): Remove.
1113 (sparc-c.o): Remove header dependencies.
1114 Use $(COMPILE) and $(POSTCOMPILE).
1115 * config/sparc/x-sparc: Likewise.
1117 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
1119 * config/arc/arc-opts.h: Add 2013 to Copyright years.
1120 * config/arc/arc700.md: Likewise.
1121 * config/arc/arc-modes.def: Likewise.
1122 * config/arc/arc-simd.h: Likewise.
1123 * config/arc/t-arc-uClibc: Likewise.
1124 * config/arc/t-arc-newlib: Likewise.
1126 2013-10-02 Renlin Li <renlin.li@arm.com>
1128 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
1130 (aarch64_expand_epilogue): Likewise.
1132 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1133 Yufeng Zhang <yufeng.zhang@arm.com>
1135 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
1137 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
1138 'base_in' represent a conversion and legal_cast_p_1 holds; set
1139 'base_in' with the returned value from get_unwidened.
1141 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1143 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
1144 plus_constant is not used.
1146 2013-10-01 Wei Mi <wmi@google.com>
1148 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
1149 * config/i386/i386.md: Add define_peephole2 to
1150 break partial reg stall for cvtss2sd/cvtsd2ss.
1152 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1154 * config/arc/arc.c (pass_arc_ifcvt::clone):
1155 Update for ctxt_ -> m_ctxt change.
1157 2013-10-01 Jeff Law <law@redhat.com>
1159 * tree-ssa-threadupdate.c (struct redirection_data): Delete
1160 outgoing_edge and intermediate_edge fields. Instead store the path.
1161 (redirection_data::hash): Hash on the last edge's destination index.
1162 (redirection_data::equal): Check the entire thread path.
1163 (lookup_redirectio_data): Corresponding changes.
1164 (create_edge_and_update_destination_phis): Likewise.
1165 (thread_single_edge): Likewise.
1167 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1168 Diego Novillo <dnovillo@google.com>
1170 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
1171 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
1172 (vld32wh_insn, vld32wl_insn): Delete commented-out old
1173 versions of these patterns.
1175 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
1176 (__builtin_arc_aligned): Likewise.
1178 * config/arc/arc.md: Expand adc_0 comment stating the intended
1179 purpose and why it isn't ready.
1180 Replace commented out call_value_via_label_mixed with a
1181 plain comment about bl_s.
1183 * config/arc/arc.c (stdio.h): Don't include directly.
1184 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
1185 Remove if (1) condition.
1186 (arc_encode_section_info): Fix comment.
1188 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1190 * config/arc/arc.c (arc_conditional_register_usage):
1191 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
1192 Also set reg_alloc_order for DMA config regs.
1194 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1195 Jeremy Bennett <jeremy.bennett@embecosm.com>
1197 * doc/install.texi (--with-cpu): Mention ARC.
1198 (arc-*-elf32): New paragraph.
1199 (arc-linux-uclibc): Likewise.
1200 * doc/md.texi (Machine Constraints): Add ARC part.
1201 * doc/invoke.texi: (menu): Add ARC Options.
1202 (Machine Dependent Options) <ARC Options>: Add synopsis.
1203 (node ARC Options): Add.
1204 * doc/extend.texi (long_call / short_call attribute): Add ARC.
1205 (ARC Built-in Functions): New section defining
1206 generic ARC built-in functions.
1207 (ARC SIMD Built-in Functions): New section defining SIMD specific
1209 (Declaring Attributes of Functions): Extended
1210 description of short_call and long_call attributes for ARC and
1211 added index entries.
1213 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
1214 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
1215 Joern Rennecke <joern.rennecke@embecosm.com>
1216 Muhammad Khurram Riaz <khurram.riaz@arc.com>
1217 Brendan Kehoe <brendan@zen.org>
1218 Michael Eager <eager@eagercon.com>
1219 Simon Cook <simon.cook@embecosm.com>
1220 Jeremy Bennett <jeremy.bennett@embecosm.com>
1222 * config/arc, common/config/arc: New directories.
1224 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1225 Brendan Kehoe <brendan@zen.org>
1226 Simon Cook <simon.cook@embecosm.com>
1228 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
1230 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
1232 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
1233 * tree-ssa-coalesce.h: New. Move prototype to here.
1234 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
1235 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
1236 (gimple_can_coalesce_p): Move to...
1237 * gimple.c (gimple_can_coalesce_p): Here.
1239 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
1241 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
1242 (dump_decl_set): Move to gimple.c.
1243 * gimple.h: Don't include tree-ssa-operands.h.
1244 (dump_decl_set): Add prototype.
1245 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
1246 Move to gimple-ssa.h.
1247 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
1248 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
1249 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
1250 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
1251 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
1253 (dump_decl_set): Relocate here.
1254 * gimple-ssa.h: New file.
1255 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
1256 Relocate from gimple.h.
1257 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
1258 * tree-ssa-operands.c (swap_ssa_operands): Rename from
1259 swap_tree_operands and remove non-ssa path.
1260 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
1261 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
1263 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
1264 vect_is_simple_reduction_1): Use swap_ssa_operands.
1265 * tree-flow.h: Move various prototypes to tree-phinodes.h.
1266 (enum need_phi_state): Move to tree-into-ssa.c.
1267 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
1268 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
1269 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
1270 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
1271 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
1272 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
1273 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
1274 num_imm_uses): Move to ssa-iterators.h.
1275 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
1276 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
1278 (op_iter_done, op_iter_next_def, op_iter_next_tree,
1279 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
1280 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
1281 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
1282 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
1283 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
1284 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
1285 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
1286 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
1287 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
1288 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
1289 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
1290 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
1291 (set_phi_nodes): Move to tree-phinodes.h.
1292 * tree-ssa-operands.h (enum ssa_op_iter_type,
1293 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
1294 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
1295 (dump_decl_set): Remove prototype.
1296 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
1297 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
1298 (set_phi_nodes): Relocate from tree-flow-inline.h.
1299 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
1301 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
1302 include list. Temporarily add gimple.h to include list.
1303 * ssa-iterators.h: New file.
1304 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
1305 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
1306 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
1307 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
1308 Relocate from tree-ssa-operands.h.
1309 (delink_imm_use, link_imm_use_to_list, link_imm_use,
1310 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
1311 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
1312 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
1313 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
1314 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
1315 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
1316 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
1317 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
1318 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
1319 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
1320 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
1321 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
1322 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
1323 Relocate from tree-flow-inline.h.
1324 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
1326 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
1329 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
1330 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
1331 Insert '\t' to output template.
1332 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
1333 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
1334 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
1335 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
1337 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
1339 * doc/install.texi (Host/target specific installation notes for GCC):
1340 Put @anchor before @heading.
1341 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
1342 Use @email for email addresses.
1344 2013-10-01 Jeff Law <law@redhat.com>
1346 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
1347 a vec. Only delete the path if we create one without successfully
1348 registering a jump thread.
1349 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
1351 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
1352 (paths): New vector of jump threading paths.
1353 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
1354 (THREAD_PATH): New accessor macro for the entire thread path.
1355 (lookup_redirection_data): Get intermediate and final outgoing edge
1356 from the thread path.
1357 (create_edge_and_update_destination_phis): Copy the threading path.
1358 (ssa_fix_duplicate_block_edges): Get edges and block types from the
1359 jump threading path.
1360 (ssa_redirect_edges): Get edges and block types from the jump threading
1361 path. Free the path vector.
1362 (thread_block): Get edges from the jump threading path. Look at the
1363 entire path to see if we thread to a loop exit. If we cancel a jump
1364 thread request, then free the path vector.
1365 (thread_single_edge): Get edges and block types from the jump threading
1366 path. Free the path vector.
1367 (thread_through_loop_header): Get edges and block types from the jump
1368 threading path. Free the path vector.
1369 (mark_threaded_blocks): Iterate over the vector of paths and store
1370 the path on the appropriate edge. Get edges and block types from the
1371 jump threading path.
1372 (mark_threaded_blocks): Get edges and block types from the jump
1373 threading path. Free the path vector.
1374 (thread_through_all_blocks): Use the vector of paths rather than
1375 a vector of 3-edge sets.
1376 (register_jump_thread): Accept pointer to a path vector rather
1377 than the path vector itself. Store the path vector for later use.
1380 2013-10-01 Jakub Jelinek <jakub@redhat.com>
1381 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1384 * config/s390/s390.c (s390_split_branches): Modify check for table
1386 (s390_chunkify_start): Rearrange table jump insn check in order to
1387 deal with compare and branch insns correctly.
1389 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
1393 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
1394 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
1395 define_insn_and_split.
1396 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
1397 (shiftsi3_compare): New pattern.
1399 * config/arm/unspecs.md (UNSPEC_RRX): New.
1401 2013-10-01 Alan Modra <amodra@gmail.com>
1403 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
1404 casing inout operands.
1406 2013-10-01 Richard Biener <rguenther@suse.de>
1408 PR tree-optimization/58553
1409 * tree-loop-distribution.c (struct partition_s): Add niter member.
1410 (classify_partition): Populate niter member for the partition
1411 and properly identify whether the relevant store happens before
1412 or after the loop exit.
1413 (generate_memset_builtin): Use niter member from the partition.
1414 (generate_memcpy_builtin): Likewise.
1416 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
1418 * vec.h (vec_prefix, vec): Prefix member names with "m_".
1419 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
1421 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
1423 * basic-block.h (edge_list): Prefix member names with "m_".
1424 * context.h (context): Likewise.
1425 * domwalk.h (dom_walker): Likewise.
1426 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
1427 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1428 * hash-table.h (hash_table): Likewise.
1429 * machmode.h (bit_field_mode_iterator): Likewise.
1430 * pass_manager.h (pass_list): Likewise.
1431 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
1432 * tree-pass.h (pass_data): Likewise.
1433 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
1434 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
1435 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1436 * asan.c (pass_data_asan): Update accordingly.
1437 * cfganal.c (control_dependences::find_control_dependence): Likewise.
1438 (control_dependences::control_dependences): Likewise.
1439 (control_dependences::~control_dependences): Likewise.
1440 (control_dependences::~control_dependences): Likewise.
1441 (control_dependences::get_edges_dependent_on): Likewise.
1442 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
1443 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
1444 * context.c (gcc::context::context): Likewise.
1445 * cprop.c (pass_rtl_cprop::clone): Likewise.
1446 * domwalk.c (dom_walker::walk): Likewise.
1447 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
1448 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
1449 * mode-switching.c (pass_mode_switching::clone): Likewise.
1450 * passes.c (opt_pass::opt_pass): Likewise.
1451 (pass_manager::pass_manager): Likewise.
1452 * predict.c (pass_strip_predict_hints::clone): Likewise.
1453 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
1454 (pass_split_all_insns::clone): Likewise.
1455 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
1457 (bit_field_mode_iterator::next_mode): Likewise.
1458 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
1459 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
1460 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
1461 * tree-complex.c (pass_lower_complex::clone): Likewise.
1462 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
1463 * tree-object-size.c (pass_object_sizes::clone): Likewise.
1464 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
1465 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
1466 (pass_fold_builtins::clone): Likewise.
1467 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
1468 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
1469 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
1470 (pass_cd_dce::clone): Likewise.
1471 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
1472 (pass_phi_only_cprop::clone): Likewise.
1473 * tree-ssa-dse.c (pass_dse::clone): Likewise.
1474 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
1475 * tree-ssa-loop.c (pass_lim::clone): Likewise.
1476 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
1477 * tree-ssa-pre.c (pass_fre::clone): Likewise.
1478 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
1479 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
1480 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
1481 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
1482 * tree-vrp.c (pass_vrp::clone): Likewise.
1483 * tsan.c (pass_tsan::clone): Likewise.
1485 2013-09-30 Jakub Jelinek <jakub@redhat.com>
1488 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
1489 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
1492 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
1493 optimization, punt if sign_bit_p looked through any zero extension.
1495 2013-09-30 Teresa Johnson <tejohnson@google.com>
1497 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
1498 Update redirected out edge count in joiner case.
1499 (ssa_redirect_edges): Common the joiner and non-joiner cases
1500 so that joiner case gets profile updates.
1502 2013-09-30 Richard Biener <rguenther@suse.de>
1504 PR tree-optimization/58554
1505 * tree-loop-distribution.c (classify_partition): Require
1506 unconditionally executed stores for memcpy and memset recognition.
1507 (tree_loop_distribution): Calculate dominance info.
1509 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1511 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
1512 (NO_PROFILE_COUNTERS): Likewise.
1513 (PROFILE_HOOK): Likewise.
1514 (FUNCTION_PROFILER): Likewise.
1515 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
1517 2013-09-30 Iain Sandoe <iain@codesourcery.com>
1519 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
1520 calls and defines in TARGET_MACHO conditional.
1521 (load_macho_picbase_di): Likewise.
1522 (reload_macho_picbase): Likewise.
1523 (reload_macho_picbase_si): Likewise.
1524 (reload_macho_picbase_di): Likewise.
1525 (nonlocal_goto_receiver): Likewise.
1527 2013-09-30 Nick Clifton <nickc@redhat.com>
1529 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
1530 that use the MSP430X ISA.
1531 (msp430_option_override): Scan -mmcu command line option for any
1532 MCU name that supports the MSP430X ISA.
1533 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
1534 -mmcu options which enable the MSP430X ISA.
1536 2013-09-30 Richard Biener <rguenther@suse.de>
1539 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
1540 before looking for setjmp-like calls.
1542 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1545 * config/darwin-protos.h (machopic_get_function_picbase): New.
1546 * config/darwin.c (machopic_get_function_picbase): New.
1547 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
1548 label for a new func. (load_macho_picbase_di): Likewise.
1549 (reload_macho_picbase): New expand.
1550 (reload_macho_picbase_si): New insn.
1551 (reload_macho_picbase_di): New insn.
1552 (nonlocal_goto_receiver): New define and split.
1553 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
1554 (unspecv enum): Add UNSPECV_NLGR.
1556 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1558 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
1559 that altivec registers are correctly sized on Darwin.
1561 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1563 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
1564 darwin-driver.o): Use COMPILE and POSTCOMPILE.
1565 * config/x-darwin (host-darwin.o): Likewise.
1566 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
1567 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
1568 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
1570 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
1572 * doc/invoke.texi: Fix usage of @tie{} command.
1574 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
1576 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
1578 2013-09-28 Jan Hubicka <jh@suse.cz>
1580 * config/i386/x86-tune.def: Add documentation for each of the options;
1583 2013-09-28 Jan Hubicka <jh@suse.cz>
1585 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
1587 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
1588 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
1589 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
1590 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
1591 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
1593 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
1595 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
1596 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
1597 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
1598 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
1599 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
1600 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
1601 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
1602 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
1603 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
1604 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
1605 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
1606 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
1607 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
1608 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
1609 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
1610 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
1611 gimple.h, godump.c, graphite-clast-to-gimple.c,
1612 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
1613 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
1614 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
1615 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1616 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
1617 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
1618 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
1619 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
1620 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
1621 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
1622 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
1623 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
1624 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
1625 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
1626 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
1627 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
1628 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
1629 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
1630 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
1631 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
1632 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
1633 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
1634 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
1635 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
1636 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1637 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1638 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
1639 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
1640 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
1641 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1642 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1643 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
1644 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
1645 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1646 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1647 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
1648 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
1649 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
1650 whitespace before "(".
1652 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
1654 * expr.h (extract_bit_field): Remove packedp parameter.
1655 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
1656 from forward declaration.
1657 (store_split_bit_field): Remove packedp arg from calls to
1658 extract_fixed_bit_field.
1659 (extract_bit_field_1): Remove packedp parameter and packedp
1660 argument from recursive calls and calls to extract_fixed_bit_field.
1661 (extract_bit_field): Remove packedp parameter and corresponding
1662 arg to extract_bit_field_1.
1663 (extract_fixed_bit_field): Remove packedp parameter. Remove code
1665 (extract_split_bit_field): Remove packedp arg from call to
1666 extract_fixed_bit_field.
1667 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
1668 (copy_blkmode_from_reg): Likewise.
1669 (copy_blkmode_to_reg): Likewise.
1670 (read_complex_part): Likewise.
1671 (store_field): Likewise.
1672 (expand_expr_real_1): Likewise.
1673 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
1674 to extract_bit_field.
1675 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
1676 call to extract_bit_field.
1677 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
1678 call to extract_bit_field.
1679 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
1680 and special packedp behavior from -fstrict-volatile-bitfields
1683 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1685 * lra-eliminations.c (init_elim_table): Guard value_p.
1687 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
1689 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
1690 DFmode, DImode, and SFmode in the upper VSX registers based on the
1691 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
1692 if -mpower8-vector. Combine -mvsx-timode handling with the rest
1693 of the VSX register handling.
1695 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
1697 (zero_extendsidi2_lfiwzx): Add support for loading into the
1698 Altivec registers with -mpower8-vector. Use wu/wv constraints to
1699 only do VSX memory options on Altivec registers.
1700 (extendsidi2_lfiwax): Likewise.
1701 (extendsfdf2_fpr): Likewise.
1702 (mov<mode>_hardfloat, SF/SD modes): Likewise.
1703 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
1704 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
1705 (movdi_internal64): Likewise.
1707 2013-09-27 Xinliang David Li <davidxl@google.com>
1709 * opts.c (finish_options): Adjust parameters
1710 according to vect cost model.
1711 (common_handle_option): Set dynamic vect cost
1713 targhooks.c (default_add_stmt_cost): Compute stmt cost
1715 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
1716 Use helper function.
1717 * tree-vectorizer.h (unlimited_cost_model): New function.
1718 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
1719 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
1721 (vect_enhance_data_refs_alignment): Ditto.
1722 * flag-types.h: New enum.
1723 * common/config/i386/i386-common.c (ix86_option_init_struct):
1724 No need to initialize vect_cost_model flag.
1725 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
1728 2013-09-27 Diego Novillo <dnovillo@google.com>
1730 * gimple.h (enum ssa_mode): Remove.
1732 2013-09-27 Paulo Matos <pmatos@broadcom.com>
1735 * gcc.dg/pr58463.c: New test.
1737 2013-09-27 Paulo Matos <pmatos@broadcom.com>
1739 * cfgloop.h (number_of_loops): Fix typo in check for null.
1741 2013-09-27 Jakub Jelinek <jakub@redhat.com>
1744 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
1745 are children of outermost saved_cfun's loop, and set it up to
1746 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
1747 if loop != loop0 and SESE region contains bbs that belong to loop0.
1749 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
1751 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
1752 (binary_scale_code_p, get_base_term, get_index_term): New functions.
1753 (set_address_segment, set_address_base, set_address_index)
1754 (set_address_disp): Accept the argument unconditionally.
1755 (baseness): Remove must_be_base_p and must_be_index_p checks.
1756 (decompose_normal_address): Classify as much as possible in the
1759 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
1761 * cse.c (count_reg_usage): Handle INT_LIST.
1762 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
1763 * reginfo.c (reg_scan_mark_refs): Likewise.
1764 * reload1.c (eliminate_regs_1): Likewise.
1766 2013-09-27 Iain Sandoe <iain@codesourcery.com>
1769 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
1772 2013-09-27 Richard Biener <rguenther@suse.de>
1774 PR tree-optimization/58459
1775 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
1776 restriction not propagating into loops.
1778 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
1780 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
1781 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
1782 * doc/tree-ssa.texi (Walking use-def chains): Delete.
1784 2013-09-26 Richard Biener <rguenther@suse.de>
1786 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
1788 2013-09-26 Richard Biener <rguenther@suse.de>
1790 * alias.h (component_uses_parent_alias_set): Rename to ...
1791 (component_uses_parent_alias_set_from): ... this.
1792 * alias.c (component_uses_parent_alias_set): Rename to ...
1793 (component_uses_parent_alias_set_from): ... this and return
1795 (reference_alias_ptr_type_1): Use the result from
1796 component_uses_parent_alias_set_from instead of stripping
1797 components one at a time.
1798 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
1800 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
1802 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
1803 Move prototypes to...
1804 * tree-ssa-ter.h: New File. Move prototypes here.
1805 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
1806 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
1807 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
1808 from is_replaceable_p.
1809 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
1810 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
1811 (process_replaceable): Use ter_is_replaceable_p.
1812 (find_replaceable_in_bb): Use ter_is_replaceable_p.
1813 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
1814 newly refactored ssa_is_replaceable_p.
1815 * cfgexpand.c: Include tree-outof-ssa.h.
1816 * ssaexpand.h: Delete.
1818 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
1820 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
1821 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
1823 (create_gimple_tmp): Delete.
1824 (get_expr_type, build_assign, build_type_cast): Move to...
1825 * gimple-builder.c: New File.
1826 (get_expr_type): Relocate from gimple.c.
1827 (build_assign, build_type_cast): Change to only create ssanames.
1828 * gimple.h: Move prototypes to...
1829 * gimple-builder.h: New File. Here.
1830 * tree-ssa.h: And here.
1831 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
1832 count_uses_and_derefs): Relocate from gimple.c.
1833 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
1834 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
1835 * tree-ssa-math-opts (execute_cse_reciprocals): Use
1836 gimple_replace_ssa_lhs.
1837 * asan.c: Include gimple-builder.h.
1838 * Makefile.in: Add gimple-builder.o.
1840 2013-09-26 Richard Biener <rguenther@suse.de>
1842 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
1843 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
1844 (loe_visit_block): Use gcc_checking_assert.
1845 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
1846 gimple_assign_ssa_name_copy_p.
1847 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
1850 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
1852 * config/rs6000/t-rs6000 (rs6000.o): Remove.
1853 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
1855 2013-09-26 Richard Biener <rguenther@suse.de>
1857 PR tree-optimization/58539
1858 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
1859 the fact that debug statements are not taking part in loop-closed
1862 2013-09-26 Nick Clifton <nickc@redhat.com>
1864 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
1865 time warning message.
1866 (msp430_print_operand_raw): Delete unused letter parameter.
1867 (TARGET_PRINT_OPERAND_ADDRESS): Define.
1868 (msp430_print_operand_address): New function.
1869 (msp430_print_operand): Move address printing code from here to
1871 * config/msp430/msp430.md (movsipsi2): Add comment in generated
1873 (zero_extendpsisi2): Likewise.
1874 (extendpsisi2): New pattern.
1875 (andneghi3): New pattern.
1877 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
1879 * config/aarch64/aarch64.opt (mlra): New option.
1880 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
1881 (TARGET_LRA_P): Define.
1883 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
1885 * expr.c (expand_assignment): Remove obsolete comment.
1887 2013-09-25 Jeff Law <law@redhat.com>
1889 * tree-flow.h (thread_through_all_blocks): Prototype moved into
1890 tree-ssa-threadupdate.h.
1891 (register_jump_thread): Similarly.
1892 * tree-ssa-threadupdate.h: New header file.
1893 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
1894 * tree-vrp.c: Likewise.
1895 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
1896 (thread_around_empty_blocks): Change type of path vector argument to
1897 an edge,type pair from just an edge. Initialize both elements when
1898 appending to a jump threading path. Tweak references to elements
1900 (thread_across_edge): Similarly. Release memory for the elements
1902 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
1903 (dump_jump_thread_path): New function broken out from
1904 register_jump_thread.
1905 (register_jump_thread): Use dump_jump_thread_path. Change type of
1906 path vector entries. Search the path for NULL edges and dump
1907 the path if one is found. Tweak the conversion of path to 3-edge
1908 form to use the block copy type information embedded in the path.
1910 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
1912 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
1914 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
1915 Vladimir Makarov <vmakarov@redhat.com>
1917 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
1918 from the least significant bit.
1919 (strip_address_mutations): Add bitfield operations handling.
1920 (must_be_index_p): Add shifting and rotate operations handling.
1921 (set_address_base): Use must_be_base_p predicate.
1922 (set_address_index): Use must_be_index_p predicate.
1924 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
1925 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1926 Sergey Lega <sergey.s.lega@intel.com>
1927 Anna Tikhonova <anna.tikhonova@intel.com>
1928 Ilya Tocar <ilya.tocar@intel.com>
1929 Andrey Turetskiy <andrey.turetskiy@intel.com>
1930 Ilya Verbin <ilya.verbin@intel.com>
1931 Kirill Yukhin <kirill.yukhin@intel.com>
1932 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1934 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
1936 (ix86_expand_vector_move_misalign): Support new unaligned load and
1937 stores and use new names.
1938 (CODE_FOR_sse2_storedqu): Rename to ...
1939 (CODE_FOR_sse2_storedquv16qi): ... this.
1940 (CODE_FOR_sse2_loaddqu): Rename to ...
1941 (CODE_FOR_sse2_loaddquv16qi): ... this.
1942 (CODE_FOR_avx_loaddqu256): Rename to ...
1943 (CODE_FOR_avx_loaddquv32qi): ... this.
1944 (CODE_FOR_avx_storedqu256): Rename to ...
1945 (CODE_FOR_avx_storedquv32qi): ... this.
1946 * config/i386/i386.md (fpint_logic): New.
1947 * config/i386/sse.md (VMOVE): Extend for AVX512.
1951 (VI_UNALIGNED_LOADSTORE): Ditto.
1952 (sse2_avx_avx512f): Ditto.
1953 (sse2_avx2): Extend for AVX512.
1954 (sse4_1_avx2): Ditto.
1955 (avx2_avx512f): New.
1956 (sse): Extend for AVX512.
1959 (avxsizesuffix): Ditto.
1960 (sseintvecmode): Ditto.
1962 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
1963 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
1964 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
1965 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
1966 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
1967 (<sse2_avx_avx512f>_storedqu<mode): ... this.
1968 (<sse>_movnt<mode>): Replace constraint "x" with "v".
1969 (STORENT_MODE): Extend for AVX512.
1970 (*absneg<mode>2): Replace constraint "x" with "v".
1971 (*mul<mode>3): Ditto.
1972 (*ieee_smin<mode>3): Ditto.
1973 (*ieee_smax<mode>3): Ditto.
1974 (avx_cmp<mode>3): Replace VF with VF_128_256.
1975 (*<sse>_maskcmp<mode>3_comm): Ditto.
1976 (<sse>_maskcmp<mode>3): Ditto.
1977 (<sse>_andnot<mode>3): Extend for AVX512.
1978 (<code><mode>3, anylogic): Replace VF with VF_128_256.
1979 (<code><mode>3, fpint_logic): New.
1980 (*<code><mode>3): Extend for AVX512.
1981 (avx512flogicsuff): New.
1982 (avx512f_<logic><mode>): Ditto.
1983 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
1985 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
1986 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
1987 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
1988 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
1989 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
1990 (xop_vpermil2<mode>3): Ditto.
1991 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
1992 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
1993 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
1994 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
1996 2013-09-25 Tom Tromey <tromey@redhat.com>
1998 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
1999 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
2000 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
2001 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
2002 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
2003 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
2004 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
2005 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
2006 (GRAPHITE_HTAB_H): Remove.
2008 2013-09-25 Tom Tromey <tromey@redhat.com>
2010 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
2012 2013-09-25 Tom Tromey <tromey@redhat.com>
2014 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
2016 2013-09-25 Tom Tromey <tromey@redhat.com>
2018 * config/i386/t-i386 (i386.o): Remove.
2019 (i386-c.o): Use COMPILE and POSTCOMPILE.
2021 2013-09-25 Tom Tromey <tromey@redhat.com>
2023 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
2025 2013-09-25 Tom Tromey <tromey@redhat.com>
2027 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
2028 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
2029 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
2030 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
2031 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
2032 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
2033 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
2034 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
2035 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
2036 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
2037 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
2038 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
2039 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
2040 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
2041 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
2042 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
2043 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
2044 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
2045 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
2046 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
2047 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
2048 (tree-ssa-pre.o, tree-ssa-sccvn.o)
2049 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
2050 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
2051 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
2052 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
2053 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
2054 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
2055 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
2056 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
2057 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
2058 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
2059 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
2060 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
2061 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
2062 (graphite-clast-to-gimple.o, graphite-dependences.o)
2063 (graphite-interchange.o, graphite-poly.o)
2064 (graphite-scop-detection.o, graphite-sese-to-poly.o)
2065 (graphite-optimize-isl.o, tree-vect-loop.o)
2066 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
2067 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
2068 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
2069 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
2070 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
2071 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
2072 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
2073 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
2074 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
2075 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
2076 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
2077 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
2078 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
2079 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
2080 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
2081 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
2082 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
2083 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
2084 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
2085 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
2086 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
2087 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
2088 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
2089 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
2090 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
2091 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
2092 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
2093 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
2094 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
2095 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
2096 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
2097 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
2098 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
2099 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
2100 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
2101 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
2102 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
2103 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
2104 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
2105 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
2106 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
2107 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
2108 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
2109 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
2110 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
2111 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
2112 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
2113 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
2114 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
2115 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
2116 (gcov-dump.o): Remove.
2117 (default-c.o): Use COMPILE and POSTCOMPILE.
2118 (CFLAGS-gcc.o): New variable.
2119 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
2121 2013-09-25 Tom Tromey <tromey@redhat.com>
2123 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
2124 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
2125 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
2126 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
2127 (c-family/c-pragma.o, c-family/c-pretty-print.o)
2128 (c-family/c-semantics.o, c-family/c-ada-spec.o)
2129 (c-family/array-notation-common.o, c-family/stub-objc.o)
2130 (c-family/c-ubsan.o): Remove.
2132 2013-09-25 Tom Tromey <tromey@redhat.com>
2134 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
2136 2013-09-25 Tom Tromey <tromey@redhat.com>
2138 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
2139 to add -DENABLE_SHARED_LIBGCC.
2140 (gcc.o): Don't use subshell.
2142 2013-09-25 Tom Tromey <tromey@redhat.com>
2144 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
2145 * configure.ac: Don't invoke AM_PROG_CC_C_O.
2146 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
2147 * configure, config.in: Rebuild.
2149 2013-09-25 Tom Tromey <tromey@redhat.com>
2151 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
2152 (COMPILE, POSTCOMPILE): New variables.
2153 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
2154 (DEPFILES): New variable.
2155 Include ".Po" files.
2156 * configure.ac: Add checks for dependency checking.
2157 * configure, aclocal.m4: Regenerate.
2159 2013-09-25 Tom Tromey <tromey@redhat.com>
2161 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
2162 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
2164 2013-09-25 Tom Tromey <tromey@redhat.com>
2166 * Makefile.in (generated_files): Add options.h,
2167 target-hooks-def.h, insn-opinit.h,
2168 common/common-target-hooks-def.h, pass-instances.def,
2169 c-family/c-target-hooks-def.h.
2171 2013-09-25 Jeff Law <law@redhat.com>
2173 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
2174 than foo[foo.length () - 1] to access last member in a vec.
2175 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
2177 2013-09-25 Richard Biener <rguenther@suse.de>
2180 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
2182 2013-09-25 Jan Hubicka <jh@suse.cz>
2184 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
2187 2013-09-25 Marek Polacek <polacek@redhat.com>
2190 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
2191 when determining the type name.
2193 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
2195 * config/sh/sh.md: Fix formatting.
2197 2013-09-24 Xinliang David Li <davidxl@google.com>
2199 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
2200 max peel iterations parameter.
2201 * param.def: New parameter.
2202 * doc/invoke.texi: Document New parameter.
2204 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
2206 * gimple-pretty-print.c: Various whitespace tweaks.
2207 * tree-core.h: Likewise.
2208 * tree-pretty-print.c: Likewise.
2209 * tree-ssa-alias.c: Likewise.
2210 * tree-ssa-copy.c: Likewise.
2211 * tree-ssanames.c: Likewise.
2212 * tree-ssanames.h: Likewise.
2213 * tree-vrp.c: Likewise.
2215 2013-09-24 Alan Modra <amodra@gmail.com>
2219 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
2220 for output operands that disallow regs. Don't use EXPAND_WRITE on
2223 2013-09-24 Richard Biener <rguenther@suse.de>
2226 * tree.c (reference_alias_ptr_type): Move ...
2227 * alias.c (reference_alias_ptr_type): ... here and implement
2228 in terms of the new reference_alias_ptr_type_1.
2229 (ref_all_alias_ptr_type_p): New helper.
2230 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
2231 use ref_all_alias_ptr_type_p.
2232 (get_deref_alias_set): Add flag_strict_aliasing check here.
2233 (reference_alias_ptr_type_1): New function, split out from ...
2234 (get_alias_set): ... here.
2235 (alias_ptr_types_compatible_p): New function.
2236 * alias.h (reference_alias_ptr_type): Declare.
2237 (alias_ptr_types_compatible_p): Likewise.
2238 * tree.h (reference_alias_ptr_type): Remove.
2239 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
2240 to compare MEM_REF alias types.
2242 2013-09-24 Richard Biener <rguenther@suse.de>
2244 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
2246 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2248 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
2249 reload helper function arrays into a single array reg_addr.
2250 (reload_fpr_gpr): Likewise.
2251 (reload_gpr_vsx): Likewise.
2252 (reload_vsx_gpr): Likewise.
2253 (struct rs6000_reg_addr): Likewise.
2254 (reg_addr): Likewise.
2255 (rs6000_debug_reg_global): Change rs6000_vector_reload,
2256 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
2257 (rs6000_init_hard_regno_mode_ok): Likewise.
2258 (rs6000_secondary_reload_direct_move): Likewise.
2259 (rs6000_secondary_reload): Likewise.
2261 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
2262 constraints: wu, ww, and wy. Repurpose wv constraint added during
2263 power8 changes. Put wg constraint in alphabetical order.
2265 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
2266 for future work to add ISA 2.07 VSX single precision support.
2267 (-mvsx-scalar-double): Change default from -1 to 1, update
2268 documentation comment.
2269 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
2270 (-mupper-regs-df): New debug switch to control whether DF values
2271 can go in the traditional Altivec registers.
2272 (-mupper-regs-sf): New debug switch to control whether SF values
2273 can go in the traditional Altivec registers.
2275 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
2277 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
2278 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
2279 Add new constraints, wu/ww/wy. Repurpose wv constraint.
2280 (rs6000_debug_legitimate_address_p): Print if we are running
2281 before, during, or after reload.
2282 (rs6000_secondary_reload): Add a comment.
2283 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
2285 * config/rs6000/constraints.md (wa constraint): Sort w<x>
2286 constraints. Update documentation string.
2287 (wd constraint): Likewise.
2288 (wf constraint): Likewise.
2289 (wg constraint): Likewise.
2290 (wn constraint): Likewise.
2291 (ws constraint): Likewise.
2292 (wt constraint): Likewise.
2293 (wx constraint): Likewise.
2294 (wz constraint): Likewise.
2295 (wu constraint): New constraint for ISA 2.07 SFmode scalar
2297 (ww constraint): Likewise.
2298 (wy constraint): Likewise.
2299 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
2300 the previous submissions.
2301 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
2303 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
2305 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
2306 (REG_BR_PROB): Say that the probability is stored in an int_list.
2307 * reg-notes.def: Update commentary to mention INT_LIST.
2308 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
2309 (INT_LIST): New rtx.
2310 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
2311 * rtlanal.c (int_reg_note_p): New function.
2312 (alloc_reg_note): Assert that the note does not have an int argument.
2313 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
2314 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
2315 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
2316 rather than an INSN_LIST. Handle INT_LIST.
2317 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
2318 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
2319 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
2320 Manipulate them as ints rather than rtxes.
2321 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
2322 * regmove.c (copy_src_to_dest): Likewise.
2323 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
2325 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
2326 into the cases that need it.
2327 * config/arm/arm.c (arm_unwind_emit): Likewise.
2329 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
2330 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
2331 * loop-doloop.c (add_test, doloop_modify): Likewise.
2332 * loop-unswitch.c (compare_and_jump_seq): Likewise.
2333 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
2334 * predict.c (combine_predictions_for_insn): Likewise.
2335 * print-rtl.c (print_rtx): Handle INT_LIST.
2336 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
2337 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
2338 * config/arm/arm.c (emit_unlikely_jump): Likewise.
2339 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
2340 (ix86_split_fp_branch, predict_jump): Likewise.
2341 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
2342 * config/sh/sh.c (expand_cbranchsi4): Likewise.
2343 * config/spu/spu.c (ea_load_store_inline): Likewise.
2345 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
2346 value of a REG_BR_PROB note.
2347 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
2348 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
2349 * emit-rtl.c (try_split): Likewise.
2350 * predict.c (br_prob_note_reliable_p): Likewise.
2351 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
2352 * reorg.c (mostly_true_jump): Likewise.
2353 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
2354 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
2355 * config/i386/i386.c (ix86_print_operand): Likewise.
2356 * config/ia64/ia64.c (ia64_print_operand): Likewise.
2357 * config/mmix/mmix.c (mmix_print_operand): Likewise.
2358 * config/rs6000/rs6000.c (output_cbranch): Likewise.
2359 * config/s390/s390.c (s390_expand_tbegin): Likewise.
2360 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
2361 * config/sparc/sparc.c (output_cbranch): Likewise.
2362 * config/spu/spu.c (get_branch_target): Likewise.
2363 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
2364 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
2366 2013-09-23 Jan Hubicka <jh@suse.cz>
2368 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
2370 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
2371 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
2373 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
2374 (gimple_fold_call): Dump inconsistent devirtualizations; add
2375 sanity check for type based devirtualizations.
2376 * ipa-prop.c: Include ipa-utils.h
2377 (ipa_intraprocedural_devirtualization): Add sanity check.
2378 (try_make_edge_direct_virtual_call): Likewise.
2380 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
2382 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
2383 assignment statements.
2385 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
2387 * gimple-pretty-print.c (dump_ssaname_info): New function.
2388 (dump_gimple_phi): Call it.
2389 (pp_gimple_stmt_1): Likewise.
2390 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
2391 (range_info_def): Declare.
2392 * tree-pretty-print.c (pp_double_int): New function.
2393 (dump_generic_node): Call it.
2394 * tree-pretty-print.h (pp_double_int): Declare.
2395 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
2396 * tree-ssanames.h (range_info_def): New structure.
2397 (value_range_type): Move definition here.
2398 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
2400 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
2402 (set_range_info): New function.
2403 (get_range_info): Likewise.
2404 (duplicate_ssa_name_range_info): Likewise.
2405 (duplicate_ssa_name_fn): Check pointer type and call
2406 duplicate_ssa_name_range_info.
2407 * tree-ssa-copy.c (fini_copy_prop): Likewise.
2408 * tree-vrp.c (value_range_type): Remove definition, now in
2410 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
2411 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
2412 (SSA_NAME_RANGE_INFO): New macro.
2414 2013-09-23 Richard Biener <rguenther@suse.de>
2416 PR tree-optimization/58464
2417 * tree-ssa-pre.c (phi_trans_lookup): Remove.
2418 (phi_trans_add): Change to add conditionally on being not present.
2419 (phi_translate_1): Remove recursion detection here.
2420 (phi_translate): Pre-seed the cache with NULL to catch
2421 recursion here in a more generic way.
2422 (bitmap_find_leader): Adjust comment.
2423 (get_representative_for): Dump value-numbers.
2424 (create_expression_by_pieces): Likewise.
2425 (insert_into_preds_of_block): Likewise.
2427 2013-09-23 Christian Bruel <christian.bruel@st.com>
2430 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
2431 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
2433 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
2436 2013-09-20 Renlin Li <renlin.li@arm.com>
2438 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2440 (aarch64_expand_epilogue): Likewise.
2441 (aarch64_legitimize_reload_address): Likewise.
2443 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
2445 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
2446 NULL_TREE before pushing them onto the vector. Likewise for labels.
2448 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
2450 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
2451 comparison operators when -fno-trapping-math is in effect.
2452 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
2453 comparison operators in TFmode and assert that unsupported operators
2455 (ia64_print_operand): Likewise.
2457 2013-09-21 Jan Hubicka <jh@suse.cz>
2459 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
2460 (sse_typeless_stores): Enable for core
2461 (sse_load0_by_pxor): Likewise.
2462 (four_jump_limit): Disable for core.
2463 (pad_returns): Likewise.
2464 (avoid_vector_decode): Likewise.
2465 (fuse_cmp_and_branch): Enable for cores.
2466 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
2468 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
2471 * config/pa/pa.c (pa_option_override): Disable auto increment and
2472 decrement instructions until reload is completed.
2474 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
2475 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
2476 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
2478 2013-09-20 DJ Delorie <dj@redhat.com>
2479 Nick Clifton <nickc@redhat.com>
2481 * config/rl78/rl78.c: Various whitespace and comment tweaks.
2482 (need_to_save): Save bank 0 on interrupts.
2483 (characterize_address): Strip far address wrappers.
2484 (rl78_as_legitimate_address): Likewise.
2485 (transcode_memory_rtx): Likewise.
2486 (rl78_peep_movhi_p): Disable this peephole after devirt.
2487 (rl78_propogate_register_origins): Forget all origins when a
2489 * config/rl78/rl78-virt.md: Various whitespace tweaks.
2490 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
2492 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
2493 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
2494 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
2496 * config/rl78/constraints.md: Various whitespace and paren tweaks.
2498 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
2500 * config/pa/pa.md: In "scc" insn patterns, change output template to
2501 handle const0_rtx in reg_or_0_operand operands.
2503 2013-09-20 Martin Husemann <martin@NetBSD.org>
2506 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
2507 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
2509 2013-09-20 Richard Biener <rguenther@suse.de>
2512 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
2513 remembering SSA name version and block index.
2514 (new_scev_info_str): Adjust.
2515 (hash_scev_info): Likewise. Also hash the block index.
2516 (eq_scev_info): Adjust.
2517 (find_var_scev_info): Likewise.
2518 (struct instantiate_cache_entry): Remove.
2519 (struct instantiate_cache_type): Use a htab to map name, block
2521 (instantiate_cache_type::~instantiate_cache_type): Adjust.
2522 (get_instantiated_value_entry): Likewise.
2523 (hash_idx_scev_info, eq_idx_scev_info): New functions.
2524 (instantiate_scev_name): Adjust.
2526 2013-09-20 Jeff Law <law@redhat.com>
2528 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
2530 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
2532 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
2533 Call aarch64_simd_expand_args to update op[argc].
2535 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
2537 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
2540 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
2542 * gengtype.c (file_rules): Added rule for *.cc files.
2543 (get_output_file_with_visibility): Give fatal message when no
2546 2013-09-20 Renlin Li <renlin.li@arm.com>
2548 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
2549 (aarch64_expand_epilogue): Likewise.
2550 (aarch64_legitimize_reload_address): Likewise.
2552 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
2555 * expr.c (expand_assignment): Remove misalignp code path.
2557 2013-09-20 Marek Polacek <polacek@redhat.com>
2560 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
2561 TYPE_PRECISION. Add asserts.
2563 2013-09-20 Richard Biener <rguenther@suse.de>
2565 PR tree-optimization/58453
2566 * tree-loop-distribution.c (distribute_loop): Apply the cost
2567 model for -ftree-loop-distribute-patterns, too.
2569 2013-09-20 Richard Biener <rguenther@suse.de>
2572 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
2573 make type comparison less strict.
2575 2013-09-20 Alan Modra <amodra@gmail.com>
2577 * configure: Regenerate.
2578 * aclocal.m4: Regenerate.
2580 2013-09-20 Marek Polacek <polacek@redhat.com>
2583 * doc/extend.texi: Document that attribute used is meant to be used
2584 on variables with static storage duration.
2586 2013-09-19 Jakub Jelinek <jakub@redhat.com>
2588 PR tree-optimization/58472
2589 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
2590 simd_lane_access set inv_p = false.
2591 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
2592 the simduid magic VAR_DECL.
2594 2013-09-19 Jan Hubicka <jh@suse.cz>
2596 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
2598 2013-09-17 Jeff Law <law@redhat.com>
2600 * tree-ssa-dom.c (record_temporary_equivalences): New function
2601 split out of dom_opt_dom_walker::after_dom_children.
2602 (dom_opt_dom_walker::thread_across_edge): Move common code
2603 in here from dom_opt_dom_walker::after_dom_children.
2604 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
2606 2013-09-19 Jan Hubicka <jh@suse.cz>
2608 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
2609 (TARGET_GENERIC): Use PROCESOR_GENERIC
2610 (enum processor_type): Unify generic32 and 64.
2611 * i386.md (cpu): Likewise.
2612 * x86-tune.def (use_leave): Enable for generic32.
2613 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
2614 * athlon.md: Change generic64 to generic in all occurences.
2615 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
2616 (ix86_target_macros_internal): Likewise.
2617 * driver-i386.c (host_detect_local_cpu): Likewise.
2618 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
2620 (generic_memcpy, generic_memset, generic_cost): This one.
2621 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
2622 (m_GENERIC32, m_GENERIC64): Remove.
2623 (m_GENERIC): Turn into one flag.
2624 (processor_target): Unify generic tunnings.
2625 (ix86_option_override_internal): Replace generic32/64 by generic.
2626 (ix86_issue_rate): Likewise.
2627 (ix86_adjust_cost): Likewise.
2629 2013-09-19 Jan Hubicka <jh@suse.cz>
2631 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
2632 of speculative flag.
2634 2013-09-19 Jakub Jelinek <jakub@redhat.com>
2636 * omp-low.c (expand_omp_sections): Always pass len - 1 to
2637 GOMP_sections_start, even if !exit_reachable.
2639 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
2641 * lra-constraints.c (need_for_all_save_p): Use macro
2642 HARD_REGNO_CALL_PART_CLOBBERED.
2643 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
2644 set up pseudo conflict hard regs.
2646 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
2649 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
2652 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
2654 PR rtl-optimization/58438
2655 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
2656 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
2657 from previous subpasses.
2659 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
2661 * arm.c (arm_get_frame_offsets): Validate architecture supports
2662 LDRD/STRD before accepting the tuning preference.
2663 (arm_expand_prologue): Likewise.
2664 (arm_expand_epilogue): Likewise.
2666 2013-09-18 Richard Biener <rguenther@suse.de>
2668 PR tree-optimization/58417
2669 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
2670 have chrecs with symbols defined in the loop as operands.
2671 (chrec_fold_multiply): Likewise.
2672 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
2673 parameters before folding binary operations.
2674 (struct instantiate_cache_entry_hasher): Remove.
2675 (struct instantiate_cache_type): Use a pointer-map.
2676 (instantiate_cache_type::instantiate_cache_type): New function.
2677 (instantiate_cache_type::get): Likewise.
2678 (instantiate_cache_type::set): Likewise.
2679 (instantiate_cache_type::~instantiate_cache_type): Adjust.
2680 (get_instantiated_value_entry): Likewise.
2681 (global_cache): New global.
2682 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
2683 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
2684 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
2685 (instantiate_scev_name): Adjust.
2686 (instantiate_scev): Construct global instead of local cache.
2687 (resolve_mixers): Likewise.
2689 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
2690 Paolo Carlini <paolo.carlini@oracle.com>
2693 * doc/implement-cxx.texi: Fix references to the C++ standards.
2695 2013-09-18 Jakub Jelinek <jakub@redhat.com>
2697 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
2698 * tree-vect-data-refs.c (vect_analyze_data_refs): For
2699 simd_lane_access drs, update also DR_ALIGNED_TO.
2701 2013-09-18 Marek Polacek <polacek@redhat.com>
2704 * doc/extend.texi: Document no_sanitize_undefined attribute.
2705 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
2706 no_sanitize_undefined attribute.
2708 2013-09-18 Nick Clifton <nickc@redhat.com>
2710 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
2711 (ASM_DECLARE_FUNCTION_NAME): Define.
2713 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
2715 * compare-elim.c (find_comparison_dom_walker): New class
2716 (find_comparisons_in_bb): Rename to
2717 find_comparison_dom_walker::before_dom_children
2718 (find_comparisons): Adjust
2719 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
2721 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
2722 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
2723 (init_walk_dominator_tree): Remove declaration.
2724 (fini_walk_dominator_tree): Remove declaration.
2725 * fwprop.c (single_def_use_dom_walker): New class
2726 (single_def_use_enter_block): Convert to
2727 single_def_use_dom_walker::before_dom_children.
2728 (single_def_use_leave_block): Convert to
2729 single_def_use_dom_walker::after_dom_children.
2730 (build_single_def_use_links): Adjust.
2731 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
2733 (find_candidates_in_block): Convert to
2734 find_candidates_dom_walker::before_dom_children.
2735 (execute_strength_reduction): Adjust.
2736 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
2737 (sese_dom_walker): New class.
2738 (sese_dom_walker::sese_dom_walker): New constructor.
2739 (sese_dom_walker::~sese_dom_walker): New destructor.
2740 (build_sese_conditions_before): Convert to
2741 sese_dom_walker::before_dom_children.
2742 (build_sese_conditions_after): Convert to
2743 sese_dom_walker::after_dom_children.
2744 (build_poly_scop): Adjust
2745 * tree-into-ssa.c (rewrite_dom_walker): New class
2746 (rewrite_enter_block): Convert to
2747 rewrite_dom_walker::before_dom_children.
2748 (rewrite_leave_block): Convert to
2749 rewrite_dom_walker::after_dom_children.
2750 (rewrite_update_dom_walker): New class.
2751 (rewrite_update_enter_block): Convert to
2752 rewrite_update_dom_walker::before_dom_children.
2753 (rewrite_update_leave_block): Convert to
2754 rewrite_update_dom_walker::after_dom_children.
2755 (rewrite_blocks, rewrite_into_ssa): Adjust.
2756 (mark_def_dom_walker): New class.
2757 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
2758 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
2759 (mark_def_sites_blocks): Convert to
2760 mark_def_dom_walker::before_dom_children.
2761 (mark_def_site_blocks): Remove.
2762 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
2763 (tree_ssa_dominator_optimize): Adjust.
2764 (dom_thread_across_edge): Convert to method
2765 dom_opt_dom_walker::thread_across_edge.
2766 (dom_opt_enter_block): Convert to member function
2767 dom_opt_dom_walker::before_dom_children.
2768 (dom_opt_leave_block): Convert to member function
2769 dom_opt_dom_walker::after_dom_children.
2770 * tree-ssa-dse.c (dse_dom_walker): New class.
2771 (dse_enter_block): Convert to member function
2772 dse_dom_walker::before_dom_children.
2773 (tree_ssa_dse): Adjust.
2774 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
2775 (determine_invariantness_stmt): Convert to method
2776 invariantness_dom_walker::before_dom_children.
2777 (determine_invariantness): Remove
2778 (move_computations_dom_walker): New class.
2779 (move_computations_stmt): Convert to method
2780 move_computations_dom_walker::before_dom_children.
2781 (move_computations, tree_ssa_lim): Adjust.
2782 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
2783 (nt_init_block): Make method
2784 notrappping_dom_walker::before_dom_children.
2785 (nt_fini_block): Make
2786 method nontrapping_dom_walker::after_dom_children.
2787 (get_non_trapping): Adjust.
2788 * tree-ssa-pre.c (eliminate_dom_walker): New class.
2789 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
2790 (eliminate_leave_block): Make method.
2791 eliminate_dom_walker::after_dom_children.
2793 * tree-ssa-strlen.c (strlen_dom_walker): New class.
2794 (strlen_enter_block): Make method
2795 strlen_dom_walker::before_dom_children.
2796 (strlen_leave_block): Make
2797 method strlen_dom_walker::after_dom_children.
2798 (tree_ssa_strlen): Adjust.
2799 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
2800 (tree_ssa_uncprop): Adjust.
2801 (uncprop_leave_block): Make method
2802 uncprop_dom_walker::after_dom_children.
2803 (uncprop_leave_block): Make method
2804 uncprop_dom_walker::before_dom_children.
2806 2013-09-18 Bin Cheng <bin.cheng@arm.com>
2808 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
2809 branch in same basic block. Check both src and dest of the move insn.
2811 2013-09-17 Nick Clifton <nickc@redhat.com>
2813 * config/rl78/rl78-real.md (bf): New pattern.
2815 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
2816 (rl78_print_operand): Do not put a # before a %B.
2817 * config/rl78/rl78.opt: Tweak doc strings.
2819 2013-09-17 DJ Delorie <dj@redhat.com>
2821 * config/rl78/constraints.md (Wcv): Allow up to $r31.
2822 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
2823 (rl78_option_override): Likewise, if -mallregs.
2824 (is_virtual_register): Likewise.
2825 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
2826 (REGNO_OK_FOR_BASE_P): Likewise.
2827 * config/rl78/rl78.opt (-mallregs): New.
2829 2013-09-17 Nick Clifton <nickc@redhat.com>
2831 * config/rl78/rl78.c (need_to_save): Change return type to bool.
2832 For interrupt functions: save all call clobbered registers if the
2833 interrupt handler is not a leaf function.
2834 (rl78_expand_prologue): Always recompute the frame information.
2835 For interrupt functions: only select bank 0 if one of the bank 0
2836 registers is going to be psuhed.
2838 2013-09-17 DJ Delorie <dj@redhat.com>
2840 * config/rl78/constraints.md: For each W* constraint, rename to C*
2841 and create a W* constraint that checks for an optional ES: prefix
2843 * config/rl78/rl78.md (UNS_ES_ADDR): New.
2844 (es_addr): New. Used to wrap far addresses.
2845 * config/rl78/rl78-protos.h (rl78_es_addr): New.
2846 (rl78_es_base): New.
2847 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
2848 wrapped far addresses.
2849 (rl78_print_operand_1): Unwrap far addresses before processing.
2850 (rl78_lo16): Wrap far addresses in unspecs.
2851 (rl78_es_addr): New.
2852 (rl78_es_base): New.
2853 (insn_ok_now): Check for not-yet-wrapped far addresses.
2854 (transcode_memory_rtx): Properly re-wrap far addresses.
2856 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
2858 * config/sparc/t-rtems: Add leon3 multilibs.
2860 2013-09-17 Cong Hou <congh@google.com>
2862 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
2863 when checking the dot production pattern. The type of rhs operand
2864 of multiply is now checked correctly.
2866 2013-09-17 Jeff Law <law@redhat.com>
2868 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2869 edge implied equivalences into successor phis.
2870 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
2871 here from tree-ssa-threadedge.c.
2872 (mark_threaded_blocks): When threading through a joiner, if both
2873 successors of the joiner's clone reach the same block, verify the
2874 PHI arguments are equal. If not, cancel the jump threading request.
2875 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
2876 tree-ssa-threadupdate.c
2877 (thread_across_edge): Don't check PHI argument equality when
2878 threading through joiner block here.
2880 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
2882 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
2883 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
2884 (warn_uninit, warn_uninitialized_vars,
2885 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
2886 Move to tree-ssa-uninit.c.
2887 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
2888 (has_undefined_value_p): New. Pass dependant parts of
2889 ssa_undefined_value_p.
2890 (uninit_undefined_value_p): Use has_undefined_value_p.
2891 (warn_uninit, warn_uninitialized_vars,
2892 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
2893 Move from tree-ssa.c.
2894 * tree-ssa.h: Adjust prototypes.
2896 2013-09-17 Jan Hubicka <jh@suse.cz>
2899 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
2900 * ipa-inline.c (can_inline_edge_p): Do not downgrade
2901 FUNCTION_NOT_OPTIMIZED.
2902 * ipa-inline-analysis.c (compute_inline_parameters): Function
2903 not optimized is not inlinable unless it is alwaysinline.
2904 (inline_analyze_function): Force calls in not optimized
2905 function not inlinable.
2907 2013-09-17 Jan Hubicka <jh@suse.cz>
2910 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
2912 * ipa.c (function_and_variable_visibility): Likewise.
2913 * ipa-profile.c (ipa_profile): Likewise.
2915 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
2918 * cgraph.c (cgraph_function_body_availability): Check for ifunc
2919 attribute, and don't inline the resolver in this case.
2921 2013-09-17 Teresa Johnson <tejohnson@google.com>
2923 * coverage.c (get_coverage_counts): Add missing newline.
2925 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2927 PR tree-optimization/58088
2928 * fold-const.c (mask_with_trailing_zeros): New function.
2929 (fold_binary_loc): Make sure we don't recurse infinitely
2930 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
2931 Use mask_with_trailing_zeros where appropriate.
2933 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
2935 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
2936 of 'prev' var to get better distance estimation.
2938 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
2940 * tree-inline.h (struct copy_body_data): Add transform_parameter.
2941 * tree-inline.c (is_parameter_of): New predicate.
2942 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
2943 a parameter has been remapped.
2944 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
2945 (optimize_inline_calls): Initialize transform_parameter.
2946 (copy_gimple_seq_and_replace_locals): Likewise.
2947 (tree_function_versioning): Likewise.
2948 (maybe_inline_call_in_expr): Likewise.
2950 2013-09-17 Nick Clifton <nickc@redhat.com>
2952 * config/msp430/msp430-protos.h: Add prototypes for new functions.
2953 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
2955 (is_attr_func): New function.
2956 (msp430_is_interrupt_func): New function.
2957 (is_naked_func): New function.
2958 (is_reentrant_func): New function.
2959 (is_critical_func): New function.
2960 (msp430_start_function): Add annotations for function attributes.
2961 (msp430_attr): New function.
2962 (msp430_attribute_table): New.
2963 (msp430_function_section): New function.
2964 (TARGET_ASM_FUNCTION_SECTION): Define.
2965 (msp430_builtin): New enum.
2966 (msp430_init_builtins): New function.
2967 (msp430_builtin_devl): New function.
2968 (msp430_expand_builtin): New function.
2969 (TARGET_INIT_BUILTINS): Define.
2970 (TARGET_EXPAND_BUILTINS): Define.
2971 (TARGET_BUILTIN_DECL): Define.
2972 (msp430_expand_prologue): Add support for naked, interrupt,
2973 critical and reentrant functions.
2974 (msp430_expand_epilogue): Likewise.
2975 (msp430_print_operand): Handle 'O' character.
2976 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
2978 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
2979 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
2980 (pushm): Use a 'n' rather than an 'i' constraint.
2981 (msp_return): Add generation of the interrupt return instruction.
2982 (disable_interrupts): New pattern.
2983 (enable_interrupts): New pattern.
2984 (push_intr_state): New pattern.
2985 (pop_intr_state): New pattern.
2986 (bic_SR): New pattern.
2987 (bis_SR): New pattern.
2988 * doc/extend.texi: Document MSP430 function attributes and builtin
2991 2013-09-17 Richard Biener <rguenther@suse.de>
2993 PR tree-optimization/58432
2994 * tree-loop-distribution.c (tree_loop_distribution): Also
2995 scan PHIs for outside loop uses and seed a partition from them.
2997 2013-09-17 Bin Cheng <bin.cheng@arm.com>
2999 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
3000 (restructure_reference): Call backtrace_base_for_ref.
3002 2013-09-17 Alan Modra <amodra@gmail.com>
3005 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
3008 2013-09-16 DJ Delorie <dj@redhat.com>
3010 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
3011 vregs location for RL78/G10.
3012 (rl78_expand_prologue): Avoid SEL on G10.
3013 (rl78_expand_epilogue): Likewise.
3014 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
3015 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
3016 __RL78_G10__ when appropriate.
3017 (ASM_SPEC): Pass -mg10 along to the assembler.
3018 * config/rl78/rl78.md (sel_rb): Disable for G10.
3019 * config/rl78/rl78.opt: Add -mg10 option.
3020 * config/rl78/t-rl78: Add -mg10 multilib.
3022 2013-09-16 Xinliang David Li <davidxl@google.com>
3024 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
3025 * omp-low.c (omp_max_vf): Ditto.
3026 (expand_omp_simd): Ditto.
3027 * tree-vectorizer.c (vectorize_loops): Ditto.
3028 (gate_vect_slp): Ditto.
3029 (gate_increase_alignment): Ditto.
3030 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
3031 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
3032 (gate_tree_vectorize): Name change.
3033 (tree_vectorize): Ditto.
3034 (pass_vectorize::gate): Call new function.
3035 (pass_vectorize::execute): Ditto.
3036 * opts.c: O3 default setting change.
3037 (finish_options): Check new flag.
3038 * doc/invoke.texi: Document new flags.
3039 * common.opt: New flags.
3041 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
3043 * doc/tm.texi.in (Cond Exec Macros): Remove node.
3044 (Condition Code): Don't reference it.
3045 * doc/tm.texi: Regenerate.
3047 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
3050 * lra-constraints.c (undo_optional_reloads): Consider all optional
3051 reload even if it did not get a hard reg.
3053 2013-09-16 Teresa Johnson <tejohnson@google.com>
3055 * dumpfile.c (dump_loc): Remove newline emission.
3056 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
3057 emission to dump_printf_loc calls where missing.
3058 (vect_mark_for_runtime_alias_test): Ditto.
3059 (vect_analyze_data_ref_dependence): Ditto.
3060 (vect_analyze_data_ref_dependences): Ditto.
3061 (vect_slp_analyze_data_ref_dependence): Ditto.
3062 (vect_slp_analyze_data_ref_dependences): Ditto.
3063 (vect_compute_data_ref_alignment): Ditto.
3064 (vect_update_misalignment_for_peel): Ditto.
3065 (vect_verify_datarefs_alignment): Ditto.
3066 (vector_alignment_reachable_p): Ditto.
3067 (vect_get_data_access_cost): Ditto.
3068 (vect_enhance_data_refs_alignment): Ditto.
3069 (vect_find_same_alignment_drs): Ditto.
3070 (vect_analyze_data_refs_alignment): Ditto.
3071 (vect_analyze_group_access): Ditto.
3072 (vect_analyze_data_ref_access): Ditto.
3073 (vect_analyze_data_ref_accesses): Ditto.
3074 (vect_prune_runtime_alias_test_list): Ditto.
3075 (vect_analyze_data_refs): Ditto.
3076 (vect_create_addr_base_for_vector_ref): Ditto.
3077 (vect_create_data_ref_ptr): Ditto.
3078 (vect_grouped_store_supported): Ditto.
3079 (vect_grouped_load_supported): Ditto.
3080 * value-prof.c (check_counter): Ditto.
3081 (check_ic_target): Ditto.
3082 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
3083 (vect_recog_widen_mult_pattern): Ditto.
3084 (vect_recog_widen_sum_pattern): Ditto.
3085 (vect_recog_over_widening_pattern): Ditto.
3086 (vect_recog_widen_shift_pattern): Ditto.
3087 (vect_recog_rotate_pattern): Ditto.
3088 (vect_recog_vector_vector_shift_pattern): Ditto.
3089 (vect_recog_divmod_pattern): Ditto.
3090 (vect_recog_mixed_size_cond_pattern): Ditto.
3091 (vect_recog_bool_pattern): Ditto.
3092 (vect_pattern_recog_1): Ditto.
3093 (vect_pattern_recog): Ditto.
3094 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
3095 (vect_is_simple_iv_evolution): Ditto.
3096 (vect_analyze_scalar_cycles_1): Ditto.
3097 (vect_get_loop_niters): Ditto.
3098 (vect_analyze_loop_1): Ditto.
3099 (vect_analyze_loop_form): Ditto.
3100 (vect_analyze_loop_operations): Ditto.
3101 (vect_analyze_loop_2): Ditto.
3102 (vect_analyze_loop): Ditto.
3103 (report_vect_op): Ditto.
3104 (vect_is_slp_reduction): Ditto.
3105 (vect_is_simple_reduction_1): Ditto.
3106 (vect_get_known_peeling_cost): Ditto.
3107 (vect_estimate_min_profitable_iters): Ditto.
3108 (vect_model_reduction_cost): Ditto.
3109 (vect_model_induction_cost): Ditto.
3110 (get_initial_def_for_induction): Ditto.
3111 (vect_create_epilog_for_reduction): Ditto.
3112 (vectorizable_reduction): Ditto.
3113 (vectorizable_induction): Ditto.
3114 (vectorizable_live_operation): Ditto.
3115 (vect_loop_kill_debug_uses): Ditto.
3116 (vect_transform_loop): Ditto.
3117 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
3118 (vect_stmt_relevant_p): Ditto.
3119 (process_use): Ditto.
3120 (vect_mark_stmts_to_be_vectorized): Ditto.
3121 (vect_model_simple_cost): Ditto.
3122 (vect_model_promotion_demotion_cost): Ditto.
3123 (vect_model_store_cost): Ditto.
3124 (vect_get_store_cost): Ditto.
3125 (vect_model_load_cost): Ditto.
3126 (vect_get_load_cost): Ditto.
3127 (vect_init_vector_1): Ditto.
3128 (vect_get_vec_def_for_operand): Ditto.
3129 (vect_finish_stmt_generation): Ditto.
3130 (vectorizable_call): Ditto.
3131 (vectorizable_conversion): Ditto.
3132 (vectorizable_assignment): Ditto.
3133 (vectorizable_shift): Ditto.
3134 (vectorizable_operation): Ditto.
3135 (vectorizable_store): Ditto.
3136 (vectorizable_load): Ditto.
3137 (vectorizable_condition): Ditto.
3138 (vect_analyze_stmt): Ditto.
3139 (vect_transform_stmt): Ditto.
3140 (vect_is_simple_use): Ditto.
3141 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
3142 (vect_can_advance_ivs_p): Ditto.
3143 (vect_update_ivs_after_vectorizer): Ditto.
3144 (vect_do_peeling_for_loop_bound): Ditto.
3145 (vect_gen_niters_for_prolog_loop): Ditto.
3146 (vect_update_inits_of_drs): Ditto.
3147 (vect_create_cond_for_alias_checks): Ditto.
3148 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
3149 (vect_build_slp_tree_1): Ditto.
3150 (vect_supported_load_permutation_p): Ditto.
3151 (vect_analyze_slp_instance): Ditto.
3152 (vect_analyze_slp): Ditto.
3153 (vect_make_slp_decision): Ditto.
3154 (vect_detect_hybrid_slp): Ditto.
3155 (vect_bb_vectorization_profitable_p): Ditto.
3156 (vect_slp_analyze_bb_1): Ditto.
3157 (vect_update_slp_costs_according_to_vf): Ditto.
3158 (vect_get_mask_element): Ditto.
3159 (vect_transform_slp_perm_load): Ditto.
3160 (vect_schedule_slp_instance): Ditto.
3161 (vect_schedule_slp): Ditto.
3162 (vect_slp_transform_bb): Ditto.
3163 * profile.c (read_profile_edge_counts): Ditto.
3164 (compute_branch_probabilities): Ditto.
3165 * coverage.c (get_coverage_counts): Ditto.
3167 2013-09-16 Diego Novillo <dnovillo@google.com>
3169 * tree-core.h: Add missing comment lines from refactoring of tree.h.
3171 2013-09-16 Jan Hubicka <jh@suse.cz>
3173 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
3174 abstract functions; for static functions check the presence of body.
3176 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3178 * config/aarch64/aarch64-simd-builtins.def (fma): New.
3179 * config/aarch64/aarch64-simd.md
3180 (aarch64_mla_elt<mode>): New.
3181 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3182 (aarch64_mls_elt<mode>): Likewise.
3183 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3184 (aarch64_fma4_elt<mode>): Likewise.
3185 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3186 (aarch64_fma4_elt_to_128v2df): Likewise.
3187 (aarch64_fma4_elt_to_64df): Likewise.
3188 (fnma<mode>4): Likewise.
3189 (aarch64_fnma4_elt<mode>): Likewise.
3190 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
3191 (aarch64_fnma4_elt_to_128v2df): Likewise.
3192 (aarch64_fnma4_elt_to_64df): Likewise.
3193 * config/aarch64/iterators.md (VDQSF): New.
3194 * config/aarch64/arm_neon.h
3195 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
3196 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
3198 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3200 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
3201 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3202 (aarch64_mul3_elt_to_128df): Likewise.
3203 (aarch64_mul3_elt_to_64v2df): Likewise.
3204 * config/aarch64/iterators.md (VEL): Also handle DFmode.
3206 (VMUL_CHANGE_NLANES) Likewise.
3209 * config/aarch64/arm_neon.h
3210 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
3212 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3214 * config/aarch64/arm_neon.h
3215 (vcvtx_high_f32_f64): Fix parameters.
3217 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3218 Uros Bizjak <ubizjak@gmail.com>
3220 * config/alpha.c: Include tree-ssanames.h.
3222 2013-09-16 Richard Biener <rguenther@suse.de>
3224 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
3225 (dot_rdg_1): Handle control_dd.
3226 (create_edge_for_control_dependence): New function.
3227 (create_rdg_edges): Add control dependences if asked for.
3228 (build_rdg): Likewise.
3229 (generate_loops_for_partition): If there are not necessary
3230 control stmts remove all their dependencies.
3231 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
3232 (distribute_loop): Pass on control dependences.
3233 (tree_loop_distribution): Compute control dependences and remove
3234 restriction on number of loop nodes.
3236 2013-09-16 Jakub Jelinek <jakub@redhat.com>
3238 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
3241 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
3243 * cse.c (try_const_anchors): Punt on CC modes.
3245 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3247 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
3249 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
3252 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
3255 2013-09-14 Iain Sandoe <iain@codesourcery.com>
3258 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
3260 (darwin_objc1_section): Likewise.
3261 (darwin_file_end): Emit Image Info section when required.
3263 2013-09-14 Jan Hubicka <jh@suse.cz>
3265 * tree-into-ssa.c (gate_into_ssa): New.
3266 (pass_data_build_ssa): Use it.
3267 * cgraph.h (expand_thunk): Update prototype.
3268 * cgraphunit.c (analyze_function): Expand thunks early.
3269 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
3270 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
3271 set lowered flag; do not add new function.
3272 (assemble_thunks_and_aliases): Update.
3273 * tree-ssa.c (gate_init_datastructures): New gate.
3274 (pass_data_init_datastructures): Use it.
3276 2013-09-14 Iain Sandoe <iain@codesourcery.com>
3279 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
3280 xmm register set described in the psABI.
3282 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
3284 * dwarf2out.c (should_emit_struct_debug): Add check
3285 for type_decl variable is not NULL.
3287 2013-09-13 Jacek Caban <jacek@codeweavers.com>
3289 * config.gcc: Use new winnt-c.c target hooks
3290 * config/t-winnt: New file
3291 * config/winnt-c.c: New file
3292 * doc/tm.texi.in: Document new hook
3293 * doc/tm.texi: Regenerated
3295 2013-09-13 Jan Hubicka <jh@suse.cz>
3298 * ipa-inline.c (check_callers): New function.
3299 (check_caller_edge): Remove.
3300 (want_inline_function_to_all_callers_p): Also permit alises that are
3302 (inline_to_all_callers): Terminate the walk when devirtualization
3303 introduce new calls.
3305 2013-09-13 Jan Hubicka <jh@suse.cz>
3307 * ipa-inline-analysis.c (struct growth_data): Add node.
3308 (do_estimate_growth_1): Fix detection of recursion.
3310 2013-09-13 Jakub Jelinek <jakub@redhat.com>
3312 PR tree-optimization/58392
3313 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
3314 to avoid shadowing of outer loop variable. If
3315 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
3316 replace_by_duplicate_decl simduid of loops that have it set and
3317 set dest_cfun->has_simduid_loops and/or
3318 dest_cfun->has_force_vect_loops.
3319 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
3320 instead of maybe_lookup_decl.
3321 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
3322 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
3323 call. Copy over force_vect and copy and remap simduid. Set
3324 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
3325 (copy_cfg_body): Remove blocks_to_copy argument. Use
3326 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
3327 caller. Don't set cfun->has_simduid_loops and/or
3328 cfun->has_force_vect_loops here.
3329 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
3331 (expand_call_inline, tree_function_versioning): Adjust copy_body
3334 2013-09-13 Martin Jambor <mjambor@suse.cz>
3337 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
3338 the assert if the edge was a speculative one.
3340 2013-09-13 Richard Biener <rguenther@suse.de>
3342 * tree-data-ref.h (known_dependences_p): Move here ...
3343 * tree-loop-distribution.c (known_dependences_p): ... from here.
3344 (dump_rdg_component, debug_rdg_component): Remove.
3346 (generate_loops_for_partition): Use gimple_uid instead of
3347 relying on matching stmt visit order.
3348 (rdg_build_partitions): Take starting stmt vector.
3349 (ldist_gen): Merge into ...
3350 (distribute_loop): ... this function. Do not compute starting
3352 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
3354 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3356 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
3357 Set type attribute correctly. Set predicable_short_it attribute.
3358 (cmpsi_shiftsi): Remove %? from output template.
3360 2013-09-13 Richard Biener <rguenther@suse.de>
3362 * tree-loop-distribution.c (struct rdg_component,
3363 rdg_defs_used_in_other_loops_p, free_rdg_components,
3364 rdg_build_components): Remove.
3365 (stmts_from_loop): Do not record virtual PHIs.
3366 (generate_loops_for_partition): Skip virtual PHIs.
3367 (build_rdg_partition_for_component): Rename to ...
3368 (build_rdg_partition_for_vertex): ... this and adjust.
3369 (rdg_build_partitions): Take a vector of starting vertices
3370 instead of components. Remove unnecessary leftover handling.
3371 (ldist_gen): Do not build components or record other stores.
3372 (distribute_loop): Do not distribute loops containing stmts
3375 2013-09-13 Christian Bruel <christian.bruel@st.com>
3378 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
3380 2013-09-13 Kai Tietz <ktietz@redhat.com>
3382 * config.gcc: Separate cases for mingw and cygwin targets,
3383 and add 64-bit cygwin target case.
3385 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
3386 dll-export inline-functions.
3387 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
3390 2013-09-13 Jeff Law <law@redhat.com>
3394 2013-09-06 Jeff Law <law@redhat.com>
3396 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3397 edge implied equivalences into successor phis.
3399 2013-09-12 DJ Delorie <dj@redhat.com>
3401 * config/rl78/rl78-virt.md: Change from | to \; for asm line
3404 2013-09-12 Brooks Moses <bmoses@google.com>
3407 * Makefile.in: Do not install driver binaries in $(target)/bin.
3409 2013-09-12 DJ Delorie <dj@redhat.com>
3411 * config/rl78/rl78.opt (mrelax): New.
3412 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
3413 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
3415 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
3417 (rl78_expand_epilogue): Likewise.
3419 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
3422 * lra-eliminations.c (remove_reg_equal_offset_note): New.
3423 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
3424 pointer elimination with using remove_reg_equal_offset_note.
3426 2013-09-12 DJ Delorie <dj@redhat.com>
3428 * config/msp430/: New port.
3429 * config.gcc (msp430): Added.
3430 * doc/invoke.texi: Document MSP430 options.
3431 * doc/install.texi: Document msp430-elf
3432 * doc/md.texi: Document msp430-elf
3433 * doc/contrib.texi: Document msp430-elf
3435 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
3438 2013-09-12 Martin Jambor <mjambor@suse.cz>
3441 * ipa-prop.c (remove_described_reference): Give up if the edge in the
3442 reference descriptor is NULL.
3443 (ipa_edge_removal_hook): If owning a reference descriptor, set its
3446 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3448 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
3449 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
3450 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
3451 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
3452 make_temp_ssa_name): move to tree-ssanames.h
3453 * tree-ssa-alias.h: Move prototype.
3454 * tree-ssa.h: Include tree-ssanames.h.
3455 * tree-ssanames.c (FREE_SSANAMES): Move to here.
3456 * tree-ssanames.h: New. Move items from tree-flow*.h
3457 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
3459 2013-09-12 Richard Biener <rguenther@suse.de>
3461 PR tree-optimization/58404
3462 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
3463 propagate non-invariant addresses into dereferences wrapped
3464 in component references.
3466 2013-09-12 Richard Biener <rguenther@suse.de>
3468 PR tree-optimization/58402
3469 * passes.def: Move pass_late_warn_uninitialized later.
3471 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3473 * tree-ssa.h: New. Move content from tree-flow.h and
3475 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
3476 Move prototypes belonging to tree-ssa.c.
3477 * tree-flow-inline.h (redirect_edge_var_map_def,
3478 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
3480 * gimple.h: Adjust prototypes.
3481 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
3483 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
3484 * tree.h: Move prototype to tree-ssa.h.
3485 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
3486 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
3487 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
3488 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
3489 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
3490 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
3491 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
3492 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
3493 graphite-blocking.c, graphite-clast-to-gimple.c,
3494 graphite-dependences.c, graphite-interchange.c,
3495 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
3496 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
3497 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
3498 ipa-reference.c, ipa-split.c, ipa-utils.c,
3499 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
3500 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
3501 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
3502 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3503 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
3504 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
3505 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
3506 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
3507 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
3508 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
3509 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
3510 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3511 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
3512 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
3513 value-prof.c, var-tracking.c,
3514 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
3516 2013-09-12 Richard Biener <rguenther@suse.de>
3518 PR tree-optimization/58396
3519 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
3520 (build_rdg): Take a loop-nest parameter, fix memleaks.
3521 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
3523 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3525 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
3528 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
3530 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
3531 The FMA instruction names should have a 'v' prefix.
3533 2013-09-12 Richard Biener <rguenther@suse.de>
3535 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
3536 (dot_rdg): Use popen instead of system in optional code.
3537 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
3538 (already_processed_vertex_p): Adjust.
3539 (has_anti_or_output_dependence, predecessor_has_mem_write,
3540 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
3541 rdg_flag_uses): Remove.
3542 (rdg_flag_vertex): Simplify.
3543 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
3545 (build_rdg_partition_for_component): Process the first vertex
3546 of a component only.
3547 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
3549 2013-09-12 Alan Modra <amodra@gmail.com>
3551 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
3553 2013-09-11 DJ Delorie <dj@redhat.com>
3554 Nick Clifton <nickc@redhat.com>
3556 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
3557 (rl78_stack_based_mem): New.
3558 * config/rl78/constraints.md (Iv08): New.
3567 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
3569 (movsi): Change from expand to insn-and-split.
3570 (ashrsi3): Clobber AX.
3574 * config/rl78/rl78.md (CC_REG): Fix.
3575 (addsi3): Allow memory and immediate operands.
3576 (addsi3_internal): Split into...
3577 (addsi3_internal_virt): ...new, and ...
3578 (addsi3_internal_real): ...new.
3580 (subsi3_internal_virt): New.
3581 (subsi3_internal_real): New.
3582 (mulsi3): Add memory operand.
3583 (mulsi3_rl78): Likewise.
3584 (mulsi3_g13): Likewise.
3585 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
3586 (cbranchqi4_real): Add more constraint options.
3587 (cbranchhi4_real): Expand pattern.
3588 (cbranchhi4_real_signed): New.
3589 (cbranchhi4_real_inverted): New.
3590 (cbranchsi4_real_lt): New.
3591 (cbranchsi4_real_ge): New.
3592 (cbranchsi4_real_signed): New.
3593 (cbranchsi4_real): New.
3595 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
3597 (lshrsi3_virt): Likewise.
3598 (ashlsi3_virt): Likewise.
3599 (cbranchqi4_virt_signed): New.
3600 (cbranchhi4_virt_signed): New.
3601 (cbranchsi4_virt): New.
3602 * config/rl78/rl78.c: Whitespace fixes throughout.
3603 (move_elim_pass): New.
3604 (pass_data_rl78_move_elim): New.
3605 (pass_rl78_move_elim): New.
3606 (make_pass_rl78_move_elim): New.
3607 (rl78_devirt_info): Run devirt earlier.
3608 (rl78_move_elim_info): New.
3609 (rl78_asm_file_start): Register it.
3610 (rl78_split_movsi): New.
3611 (rl78_as_legitimate_address): Allow virtual base registers when
3613 (rl78_addr_space_convert): Remove spurious debug stuff.
3614 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
3615 (rl78_print_operand): More cases for not printing '#'.
3616 (rl78_expand_compare): Remove most of the logic.
3617 (content_memory): New.
3618 (clear_content_memory): New.
3619 (get_content_index): New.
3620 (get_content_name): New.
3621 (display_content_memory): New.
3622 (update_content): New.
3623 (record_content): New.
3624 (already_contains): New.
3625 (insn_ok_now): Re-recog insns with virtual registers.
3626 (add_postponed_content_update): New.
3627 (process_postponed_content_update): New.
3628 (gen_and_emit_move): New.
3629 (transcode_memory_rtx): Record new location content.
3630 Use gen_and_emit_move.
3631 (force_into_acc): New.
3632 (move_to_acc): Use gen_and_emit_move.
3633 (move_from_acc): Likewise.
3634 (move_acc_to_reg): Likewise.
3635 (move_to_x): Likewise.
3636 (move_to_hl): Likewise.
3637 (move_to_de): Likewise.
3638 (rl78_alloc_physical_registers_op1): Record location content.
3639 (has_constraint): New.
3640 (rl78_alloc_physical_registers_op2): Record location content.
3642 (rl78_alloc_physical_registers_ro1): Likewise.
3643 (rl78_alloc_physical_registers_cmp): Likewise.
3644 (rl78_alloc_physical_registers_umul): Likewise.
3645 (rl78_alloc_address_registers_macax): New.
3646 (rl78_alloc_physical_registers): Initialize and set location
3647 content memory as needed.
3648 (rl78_reorg): Make sure split2 is called.
3649 (rl78_rtx_costs): New.
3651 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
3653 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
3654 for (not (neg ...)) and (neg (not ...)) cases.
3656 2013-09-11 Richard Biener <rguenther@suse.de>
3659 * passes.def: Split critical edges before late uninit warning passes.
3660 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
3662 2013-09-11 Jakub Jelinek <jakub@redhat.com>
3664 PR tree-optimization/58385
3665 * fold-const.c (build_range_check): If both low and high are NULL,
3666 use omit_one_operand_loc to preserve exp side-effects.
3668 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3670 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
3672 2013-09-11 Richard Biener <rguenther@suse.de>
3674 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
3675 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
3676 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
3677 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
3678 stmts_from_loop, known_dependences_p, build_empty_rdg,
3679 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
3680 * tree-loop-distribution.c: ... here.
3681 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
3682 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
3683 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
3684 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
3685 * tree-loop-distribution.c: ... here.
3686 * tree-loop-distribution.c: Include gimple-pretty-print.h.
3687 (struct partition_s): Add loops member.
3688 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
3689 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
3690 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
3692 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3693 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3694 Sergey Lega <sergey.s.lega@intel.com>
3695 Anna Tikhonova <anna.tikhonova@intel.com>
3696 Ilya Tocar <ilya.tocar@intel.com>
3697 Andrey Turetskiy <andrey.turetskiy@intel.com>
3698 Ilya Verbin <ilya.verbin@intel.com>
3699 Kirill Yukhin <kirill.yukhin@intel.com>
3700 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3702 * config/i386/constraints.md (k): New.
3704 * config/i386/i386.c (const regclass_map): Add new mask registers.
3705 (dbx_register_map): Ditto.
3706 (dbx64_register_map): Ditto.
3707 (svr4_dbx_register_map): Ditto.
3708 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
3710 (ix86_preferred_reload_class): Disable constants for mask registers.
3711 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
3712 (ix86_hard_regno_mode_ok): Support new mask registers.
3713 (x86_order_regs_for_local_alloc): Ditto.
3714 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
3715 (FIXED_REGISTERS): Add new mask registers.
3716 (CALL_USED_REGISTERS): Ditto.
3717 (REG_ALLOC_ORDER): Ditto.
3718 (VALID_MASK_REG_MODE): New.
3719 (FIRST_MASK_REG): Ditto.
3720 (LAST_MASK_REG): Ditto.
3721 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
3722 (MAYBE_MASK_CLASS_P): New.
3723 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
3724 (REG_CLASS_CONTENTS): Ditto.
3725 (MASK_REGNO_P): New.
3726 (ANY_MASK_REG_P): Ditto.
3727 (HI_REGISTER_NAMES): Add new mask registers.
3728 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
3729 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
3731 (attribute "type"): Add mskmov, msklog.
3732 (attribute "length_immediate"): Support them.
3733 (attribute "memory"): Ditto.
3734 (attribute "prefix_0f"): Ditto.
3735 (*movhi_internal): Support new mask registers.
3736 (*movqi_internal): Ditto.
3737 (define_split): Split out clobber pattern is a logic
3738 insn on mask registers.
3739 (*k<logic><mode>): New.
3740 (*andhi_1): Extend to support mask regs.
3741 (*andqi_1): Extend to support mask regs.
3743 (define_split): Split and-not to and and not if operands
3745 (*<code><mode>_1): Separate HI mode to new pattern...
3746 (*<code>hi_1): This.
3747 (*<code>qi_1): Extend to support mask regs.
3749 (kortestzhi): Ditto.
3750 (kortestchi): Ditto.
3752 (*one_cmpl<mode>2_1): Remove HImode and handle it...
3753 (*one_cmplhi2_1): ...Here, now with mask registers support.
3754 (*one_cmplqi2_1): Support new mask registers.
3755 (HI/QImode arithmetics splitter): Don't split if mask registers
3757 (HI/QImode not splitter): Ditto.
3758 * config/i386/predicated.md (mask_reg_operand): New.
3759 (general_reg_operand): Ditto.
3761 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3763 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
3764 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
3766 2013-09-10 Jeff Law <law@redhat.com>
3768 PR tree-optimization/58380
3769 * tree-ssa-threadupdate.c (thread_block): Recognize another case
3770 of threading through a buried loop header.
3772 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
3773 return value for single successor case.
3775 2013-09-10 Jan Hubicka <jh@suse.cz>
3777 * ipa-devirt.c (ipa_devirt): Enable with LTO.
3779 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
3782 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
3783 support conditional execution.
3784 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
3786 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
3788 * lra.c (lra): Clear lra_optional_reload_pseudos before every
3790 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
3791 Check destination too to check move insn.
3792 (undo_optional_reloads): Add check that the original peudo did not
3793 changed its allocation and the optional reload was inherited on last
3794 inheritance pass. Break loop after deciding to keep optional reload.
3795 (lra_undo_inheritance): Add check that inherited pseudo still in
3798 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
3800 * config/aarch64/aarch64.md (generic_sched): New.
3801 * config/aarch64/aarch64-generic.md (load): Make conditional
3802 on generic_sched attribute.
3803 (nonload): Likewise.
3805 2013-09-10 Jan Hubicka <jh@suse.cz>
3807 * lto-cgraph.c: Include ipa-utils.h.
3808 (compute_ltrans_boundary): Also add possible targets into the boundary.
3810 2013-09-10 Jan Hubicka <jh@suse.cz>
3812 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
3813 VAR_DECL of vtable rather than full expression.
3815 2013-09-10 Jan Hubicka <jh@suse.cz>
3816 Paolo Carlini <paolo.carlini@oracle.com>
3818 * cgraphunit.c (analyze_functions): Save input_location, set it
3819 to UNKNOWN_LOCATION and restore it at the end.
3821 2013-09-10 Martin Jambor <mjambor@suse.cz>
3823 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
3824 represented by a thunk.
3826 2013-09-10 Jeff Law <law@redhat.com>
3828 PR tree-optimization/58343
3829 * tree-ssa-threadupdate.c (thread_block): Identify and disable
3830 jump threading requests through loop headers buried in the middle
3831 of a jump threading path.
3833 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
3834 in return value/type.
3836 2013-09-10 Jakub Jelinek <jakub@redhat.com>
3838 PR rtl-optimization/58365
3839 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
3840 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
3843 2013-09-10 Richard Biener <rguenther@suse.de>
3845 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
3846 * tree-data-ref.c (create_rdg_vertices): Collect all data
3847 references, signal failure to the caller, use data-ref API.
3848 (build_rdg): Compute data references only once. Maintain lifetime
3849 of data references and data dependences from within RDG.
3850 (free_rdg): Free dependence relations.
3851 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
3852 inventing extra dependences.
3853 (distribute_loop): Update for RDG API changes.
3855 2013-09-10 Kai Tietz <ktietz@redhat.com>
3857 * doc/invoke.texi (fms-extensions): Document changed
3858 behavior for ms-abi targets.
3859 * config/i386/i386.c (ix86_option_override_internal):
3860 Set default value of option -fms-extension for ms-abi targets.
3862 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3864 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
3866 2013-09-10 Alan Modra <amodra@gmail.com>
3869 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
3871 2013-09-10 Alan Modra <amodra@gmail.com>
3873 * config/rs6000/predicates.md (add_cint_operand): New.
3874 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
3875 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
3876 using add_cint_operand.
3877 (largetoc_high_plus_aix): Likewise.
3879 2013-09-09 Jakub Jelinek <jakub@redhat.com>
3881 PR tree-optimization/58364
3882 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
3883 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
3884 the current range can't be an unconditional true or false.
3886 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
3888 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
3890 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
3892 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
3894 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
3897 * doc/invoke.texi (-Wdelete-incomplete): Document it.
3899 2013-09-09 Ian Bolton <ian.bolton@arm.com>
3901 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
3902 NO_REGS for immediate that can't be moved directly into FP_REGS.
3904 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3906 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
3907 comparison with negated operand.
3908 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
3911 2013-09-09 Richard Biener <rguenther@suse.de>
3914 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
3915 of a subloop record all its block as affecting loop-closed SSA form.
3917 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3919 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
3920 of an rtx/bitpos pair.
3921 (store_fixed_bit_field): Update accordingly.
3923 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3925 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
3927 * builtins.c (expand_errno_check): Likewise.
3928 * dwarf2cfi.c (init_return_column_size): Likewise.
3929 * except.c (sjlj_mark_call_sites): Likewise.
3930 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
3931 * lra-constraints.c (emit_inc): Likewise.
3932 * ree.c (combine_set_extension): Likewise.
3933 * regmove.c (fixup_match_2): Likewise.
3934 * reload1.c (inc_for_reload): Likewise.
3936 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3938 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
3939 (force_to_mode, simplify_shift_const_1, simplify_comparison):
3940 Use gen_int_mode with the mode of the associated simplify_* call.
3941 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
3942 * expmed.c (expand_shift_1): Likewise.
3943 * function.c (instantiate_virtual_regs_in_insn): Likewise.
3944 * loop-iv.c (iv_number_of_iterations): Likewise.
3945 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
3946 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
3948 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3950 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
3951 of the associated expand_* call.
3952 (asan_emit_stack_protection): Likewise.
3953 * builtins.c (round_trampoline_addr): Likewise.
3954 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
3955 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
3956 (emit_store_flag): Likewise.
3957 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
3958 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
3960 * function.c (instantiate_virtual_regs_in_insn): Likewise.
3961 * ifcvt.c (noce_try_store_flag_constants): Likewise.
3962 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
3963 * modulo-sched.c (generate_prolog_epilog): Likewise.
3964 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
3965 (expand_ctz, expand_ffs, expand_unop): Likewise.
3967 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3969 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
3970 of the associated gen_rtx_* call.
3971 * caller-save.c (init_caller_save): Likewise.
3972 * combine.c (find_split_point, make_extraction): Likewise.
3973 (make_compound_operation): Likewise.
3974 * dwarf2out.c (mem_loc_descriptor): Likewise.
3975 * explow.c (plus_constant, probe_stack_range): Likewise.
3976 * expmed.c (expand_mult_const): Likewise.
3977 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
3978 * reload1.c (init_reload): Likewise.
3979 * valtrack.c (cleanup_auto_inc_dec): Likewise.
3980 * var-tracking.c (adjust_mems): Likewise.
3981 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
3982 rather than gen_rtx_fmt_ee.
3984 2013-09-09 Jan Hubicka <jh@suse.cz>
3987 * value-prof.c (gimple_ic): Copy also abnormal edges.
3989 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3991 * asan.c (asan_shadow_cst): Use gen_int_mode.
3993 2013-09-08 Jan Hubicka <jh@suse.cz>
3995 * ipa-profile.c: Add toplevel comment.
3996 (ipa_propagate_frequency_1): Be more conservative when profile is read.
3997 (contains_hot_call_p): New function.
3998 (ipa_propagate_frequency): Set frequencies based on counts when
4000 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
4001 profile; do not tamper with profile after inlining if it is read.
4003 2013-09-08 Jan Hubicka <jh@suse.cz>
4005 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
4008 2013-09-08 Jan Hubicka <jh@suse.cz>
4010 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
4013 2013-09-08 Jeff Law <law@redhat.com>
4016 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
4019 2013-09-08 Andi Kleen <ak@linux.intel.com>
4021 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
4023 2013-09-08 Jan Hubicka <jh@suse.cz>
4025 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
4027 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
4029 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
4030 for non-debug insns.
4031 * lra.c (new_insn_reg): Take the containing insn as a parameter.
4032 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
4033 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
4036 2013-09-08 Jan Hubicka <jh@suse.cz>
4038 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
4039 targets and devirtualize to BUILT_IN_UNREACHABLE.
4040 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
4041 * ipa.c (walk_polymorphic_call_targets): New function.
4042 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
4043 functions; use the new timevar.
4044 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
4045 was removed from the program.
4046 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
4047 not consider it in the walk when its vtable is dead.
4048 (possible_polymorphic_call_targets_1): Pass anonymous flag to
4050 (devirt_variable_node_removal_hook): New function.
4051 (possible_polymorphic_call_targets): Also register
4052 devirt_variable_node_removal_hook.
4053 (ipa_devirt): Do not do non-speculative devirtualization.
4054 (gate_ipa_devirt): One execute if devirtualizing speculatively.
4056 2013-09-08 Jan Hubicka <jh@suse.cz>
4058 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
4059 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
4060 varpool_remove_variable_insertion_hook): Declare.
4061 * varpool.c (varpool_node_hook_list): New structure.
4062 (first_varpool_node_removal_hook,
4063 first_varpool_variable_insertion_hook): New variables.
4064 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
4065 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
4066 varpool_remove_variable_insertion_hook,
4067 varpool_call_variable_insertion_hooks): New functions.
4068 (varpool_remove_node): Use it.
4070 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
4073 * diagnostic.c (diagnostic_build_prefix): When s.file is
4074 "<built-in>" don't output line and column numbers.
4076 2013-09-06 Jan Hubicka <jh@suse.cz>
4078 * cgraphunit.c (expand_thunk): Get body before touching arguments.
4079 * lto-streamer-out.c: Stream thunks, too.
4080 * lto-streamer-in.c (input_function): Pop cfun here
4081 (lto_read_body): Instead of here.
4083 2013-09-06 Caroline Tice <cmtice@google.com>
4085 * doc/install.texi: Add documentation for the --enable-vtable-verify
4086 and the --disable-libvtv configure options.
4088 2013-09-06 Jeff Law <law@redhat.com>
4090 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4091 edge implied equivalences into successor phis.
4093 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
4095 * resource.c (mark_referenced_resources): Handle COND_EXEC.
4097 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
4099 * resource.c (mark_target_live_regs): Compute resources taking
4100 into account if a call is predicated or not.
4102 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
4104 * toplev.c (output_stack_usage): Be prepared for suffixes created by
4105 the compiler in the function names.
4107 2013-09-06 Jan Hubicka <jh@suse.cz>
4110 * ipa-inline.c (has_caller_p): New function.
4111 (want_inline_function_to_all_callers_p): Use it.
4112 (sum_callers, inline_to_all_callers): Break out from ...
4113 (ipa_inline): ... here.
4115 2013-09-06 Jan Hubicka <jh@suse.cz>
4117 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
4118 only when AVX is enabled.
4120 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4122 * config/aarch64/aarch64.md
4123 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
4124 is fpsimd_<load/store>2.
4125 (load_pair<mode>): Likewise.
4126 (store_pair<mode>): Likewise.
4128 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4130 * config/arm/types.md (type): Add "mrs" type.
4131 * config/aarch64/aarch64.md
4132 (aarch64_load_tp_hard): Make type "mrs".
4134 (load_tp_hard): Make type "mrs".
4135 * config/arm/cortex-a15.md: Update with new attributes.
4136 * config/arm/cortex-a5.md: Update with new attributes.
4137 * config/arm/cortex-a53.md: Update with new attributes.
4138 * config/arm/cortex-a7.md: Update with new attributes.
4139 * config/arm/cortex-a8.md: Update with new attributes.
4140 * config/arm/cortex-a9.md: Update with new attributes.
4141 * config/arm/cortex-m4.md: Update with new attributes.
4142 * config/arm/cortex-r4.md: Update with new attributes.
4143 * config/arm/fa526.md: Update with new attributes.
4144 * config/arm/fa606te.md: Update with new attributes.
4145 * config/arm/fa626te.md: Update with new attributes.
4146 * config/arm/fa726te.md: Update with new attributes.
4148 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4150 * config/aarch64/aarch64.md
4151 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
4152 (*movtf_aarch64): Likewise.
4154 (thumb1_movdi_insn): Use "multiple" for type where more than one
4155 instruction is used for a move.
4156 (*arm32_movhf): Likewise.
4157 (*thumb_movdf_insn): Likewise.
4159 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4161 * config/arm/types.md (type): Rename fcpys to fmov.
4163 (*arm_movsi_vfp): Rename type fcpys as fmov.
4164 (*thumb2_movsi_vfp): Likewise
4165 (*movhf_vfp_neon): Likewise
4166 (*movhf_vfp): Likewise
4167 (*movsf_vfp): Likewise
4168 (*thumb2_movsf_vfp): Likewise
4169 (*movsfcc_vfp): Likewise
4170 (*thumb2_movsfcc_vfp): Likewise
4171 * config/aarch64/aarch64-simd.md
4172 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
4173 * config/aarch64/aarch64.md
4174 (*movsi_aarch64): Replace type mov_reg with fmovs.
4175 (*movdi_aarch64): Likewise
4176 (*movsf_aarch64): Likewise
4177 (*movdf_aarch64): Likewise
4179 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
4180 * config/arm/iwmmxt.md
4181 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
4182 * config/arm/arm1020e.md: Update with new attributes.
4183 * config/arm/cortex-a15-neon.md: Update with new attributes.
4184 * config/arm/cortex-a5.md: Update with new attributes.
4185 * config/arm/cortex-a53.md: Update with new attributes.
4186 * config/arm/cortex-a7.md: Update with new attributes.
4187 * config/arm/cortex-a8-neon.md: Update with new attributes.
4188 * config/arm/cortex-a9.md: Update with new attributes.
4189 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4190 * config/arm/cortex-r4f.md: Update with new attributes.
4191 * config/arm/marvell-pj4.md: Update with new attributes.
4192 * config/arm/vfp11.md: Update with new attributes.
4194 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4196 * config/aarch64/aarch64.md
4197 (*madd<mode>): Fix type attribute.
4198 (*maddsi_uxtw): Likewise.
4199 (*msub<mode>): Likewise.
4200 (*msubsi_uxtw): Likewise.
4201 (<su_optab>maddsidi4): Likewise.
4202 (<su_optab>msubsidi4): Likewise.
4204 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4206 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
4207 * config/arm/arm.md (core_cycles): Remove fdiv.
4208 * config/arm/vfp.md:
4209 (*sqrtsf2_vfp): Update for attribute changes.
4210 (*sqrtdf2_vfp): Likewise.
4211 * config/aarch64/aarch64.md:
4212 (sqrt<mode>2): Update for attribute changes.
4213 * config/arm/arm1020e.md: Update with new attributes.
4214 * config/arm/cortex-a15-neon.md: Update with new attributes.
4215 * config/arm/cortex-a5.md: Update with new attributes.
4216 * config/arm/cortex-a53.md: Update with new attributes.
4217 * config/arm/cortex-a7.md: Update with new attributes.
4218 * config/arm/cortex-a8-neon.md: Update with new attributes.
4219 * config/arm/cortex-a9.md: Update with new attributes.
4220 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4221 * config/arm/cortex-r4f.md: Update with new attributes.
4222 * config/arm/marvell-pj4.md: Update with new attributes.
4223 * config/arm/vfp11.md: Update with new attributes.
4225 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4227 * config/arm/types.md
4228 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
4229 * config/aarch64/aarch64.md
4230 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
4232 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
4233 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
4234 (float<GPI:mode><GPF:mode>2): Likewise.
4236 (*truncsisf2_vfp): Update with new attributes.
4237 (*truncsidf2_vfp): Likewise.
4238 (fixuns_truncsfsi2): Likewise.
4239 (fixuns_truncdfsi2): Likewise.
4240 (*floatsisf2_vfp): Likewise.
4241 (*floatsidf2_vfp): Likewise.
4242 (floatunssisf2): Likewise.
4243 (floatunssidf2): Likewise.
4244 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
4245 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
4246 * config/arm/arm1020e.md: Update with new attributes.
4247 * config/arm/cortex-a15-neon.md: Update with new attributes.
4248 * config/arm/cortex-a5.md: Update with new attributes.
4249 * config/arm/cortex-a53.md: Update with new attributes.
4250 * config/arm/cortex-a7.md: Update with new attributes.
4251 * config/arm/cortex-a8-neon.md: Update with new attributes.
4252 * config/arm/cortex-a9.md: Update with new attributes.
4253 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4254 * config/arm/cortex-r4f.md: Update with new attributes.
4255 * config/arm/marvell-pj4.md: Update with new attributes.
4256 * config/arm/vfp11.md: Update with new attributes.
4258 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4260 * config/aarch64/arm_neon.h
4261 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
4262 (vqtbx<1,2,3,4><q>_s8): Likewise.
4264 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4266 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
4267 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
4268 (add<mode>3): Add type attribute.
4269 (add<mode>3): Likewise.
4270 (usadd<mode>3): Likewise.
4271 (ssadd<mode>3): Likewise.
4272 (sub<mode>3): Likewise.
4273 (sub<mode>3): Likewise.
4274 (ussub<mode>3): Likewise.
4275 (sssub<mode>3): Likewise.
4276 (ssmulsa3): Likewise.
4277 (usmulusa3): Likewise.
4278 (arm_usatsihi): Likewise.
4280 (*movdi_vfp): Add types for all instructions.
4281 (*movdi_vfp_cortexa8): Likewise.
4282 (*movhf_vfp_neon): Likewise.
4283 (*movhf_vfp): Likewise.
4284 (*movdf_vfp): Likewise.
4285 (*thumb2_movdf_vfp): Likewise.
4286 (*thumb2_movdfcc_vfp): Likewise.
4287 * config/arm/arm.md: Add type attribute to all insn patterns.
4288 (*thumb1_adddi3): Add type attribute.
4289 (*arm_adddi3): Likewise.
4290 (*adddi_sesidi_di): Likewise.
4291 (*adddi_zesidi_di): Likewise.
4292 (*thumb1_addsi3): Likewise.
4293 (addsi3_compare0): Likewise.
4294 (*addsi3_compare0_scratch): Likewise.
4295 (*compare_negsi_si): Likewise.
4296 (cmpsi2_addneg): Likewise.
4297 (*addsi3_carryin_<optab>): Likewise.
4298 (*addsi3_carryin_alt2_<optab>): Likewise.
4299 (*addsi3_carryin_clobercc_<optab>): Likewise.
4300 (*subsi3_carryin): Likewise.
4301 (*subsi3_carryin_const): Likewise.
4302 (*subsi3_carryin_compare): Likewise.
4303 (*subsi3_carryin_compare_const): Likewise.
4304 (*arm_subdi3): Likewise.
4305 (*thumb_subdi3): Likewise.
4306 (*subdi_di_zesidi): Likewise.
4307 (*subdi_di_sesidi): Likewise.
4308 (*subdi_zesidi_di): Likewise.
4309 (*subdi_sesidi_di): Likewise.
4310 (*subdi_zesidi_ze): Likewise.
4311 (thumb1_subsi3_insn): Likewise.
4312 (*arm_subsi3_insn): Likewise.
4313 (*anddi3_insn): Likewise.
4314 (*anddi_zesidi_di): Likewise.
4315 (*anddi_sesdi_di): Likewise.
4316 (*ne_zeroextracts): Likewise.
4317 (*ne_zeroextracts): Likewise.
4318 (*ite_ne_zeroextr): Likewise.
4319 (*ite_ne_zeroextr): Likewise.
4320 (*anddi_notdi_di): Likewise.
4321 (*anddi_notzesidi): Likewise.
4322 (*anddi_notsesidi): Likewise.
4323 (andsi_notsi_si): Likewise.
4324 (thumb1_bicsi3): Likewise.
4325 (*iordi3_insn): Likewise.
4326 (*iordi_zesidi_di): Likewise.
4327 (*iordi_sesidi_di): Likewise.
4328 (*thumb1_iorsi3_insn): Likewise.
4329 (*xordi3_insn): Likewise.
4330 (*xordi_zesidi_di): Likewise.
4331 (*xordi_sesidi_di): Likewise.
4332 (*arm_xorsi3): Likewise.
4333 (*andsi_iorsi3_no): Likewise.
4334 (*smax_0): Likewise.
4335 (*smax_m1): Likewise.
4336 (*arm_smax_insn): Likewise.
4337 (*smin_0): Likewise.
4338 (*arm_smin_insn): Likewise.
4339 (*arm_umaxsi3): Likewise.
4340 (*arm_uminsi3): Likewise.
4341 (*minmax_arithsi): Likewise.
4342 (*minmax_arithsi_): Likewise.
4343 (*satsi_<SAT:code>): Likewise.
4344 (arm_ashldi3_1bit): Likewise.
4345 (arm_ashrdi3_1bit): Likewise.
4346 (arm_lshrdi3_1bit): Likewise.
4347 (*arm_negdi2): Likewise.
4348 (*thumb1_negdi2): Likewise.
4349 (*arm_negsi2): Likewise.
4350 (*thumb1_negsi2): Likewise.
4351 (*negdi_extendsid): Likewise.
4352 (*negdi_zero_extend): Likewise.
4353 (*arm_abssi2): Likewise.
4354 (*thumb1_abssi2): Likewise.
4355 (*arm_neg_abssi2): Likewise.
4356 (*thumb1_neg_abss): Likewise.
4357 (one_cmpldi2): Likewise.
4358 (extend<mode>di2): Likewise.
4359 (*compareqi_eq0): Likewise.
4360 (*arm_extendhisi2addsi): Likewise.
4361 (*arm_movdi): Likewise.
4362 (*thumb1_movdi_insn): Likewise.
4363 (*arm_movt): Likewise.
4364 (*thumb1_movsi_insn): Likewise.
4365 (pic_add_dot_plus_four): Likewise.
4366 (pic_add_dot_plus_eight): Likewise.
4367 (tls_load_dot_plus_eight): Likewise.
4368 (*thumb1_movhi_insn): Likewise.
4369 (*thumb1_movsf_insn): Likewise.
4370 (*movdf_soft_insn): Likewise.
4371 (*thumb_movdf_insn): Likewise.
4372 (cbranchsi4_insn): Likewise.
4373 (cbranchsi4_scratch): Likewise.
4374 (*negated_cbranchsi4): Likewise.
4375 (*tbit_cbranch): Likewise.
4376 (*tlobits_cbranch): Likewise.
4377 (*tstsi3_cbranch): Likewise.
4378 (*cbranchne_decr1): Likewise.
4379 (*addsi3_cbranch): Likewise.
4380 (*addsi3_cbranch_scratch): Likewise.
4381 (*arm_cmpdi_insn): Likewise.
4382 (*arm_cmpdi_unsig): Likewise.
4383 (*arm_cmpdi_zero): Likewise.
4384 (*thumb_cmpdi_zero): Likewise.
4385 (*deleted_compare): Likewise.
4386 (*mov_scc): Likewise.
4387 (*mov_negscc): Likewise.
4388 (*mov_notscc): Likewise.
4389 (*cstoresi_eq0_thumb1_insn): Likewise.
4390 (cstoresi_nltu_thumb1): Likewise.
4391 (cstoresi_ltu_thu): Likewise.
4392 (thumb1_addsi3_addgeu): Likewise.
4393 (*arm_jump): Likewise.
4394 (*thumb_jump): Likewise.
4395 (*check_arch2): Likewise.
4396 (arm_casesi_internal): Likewise.
4397 (thumb1_casesi_dispatch): Likewise.
4398 (*arm_indirect_jump): Likewise.
4399 (*thumb1_indirect_jump): Likewise.
4401 (*and_scc): Likewise.
4402 (*ior_scc): Likewise.
4403 (*compare_scc): Likewise.
4404 (*cond_move): Likewise.
4405 (*cond_arith): Likewise.
4406 (*cond_sub): Likewise.
4407 (*cmp_ite0): Likewise.
4408 (*cmp_ite1): Likewise.
4409 (*cmp_and): Likewise.
4410 (*cmp_ior): Likewise.
4411 (*ior_scc_scc): Likewise.
4412 (*ior_scc_scc_cmp): Likewise.
4413 (*and_scc_scc): Likewise.
4414 (*and_scc_scc_cmp): Likewise.
4415 (*and_scc_scc_nod): Likewise.
4416 (*negscc): Likewise.
4417 (movcond_addsi): Likewise.
4418 (movcond): Likewise.
4419 (*ifcompare_plus_move): Likewise.
4420 (*if_plus_move): Likewise.
4421 (*ifcompare_move_plus): Likewise.
4422 (*if_move_plus): Likewise.
4423 (*ifcompare_arith_arith): Likewise.
4424 (*if_arith_arith): Likewise.
4425 (*ifcompare_arith_move): Likewise.
4426 (*if_arith_move): Likewise.
4427 (*ifcompare_move_arith): Likewise.
4428 (*if_move_arith): Likewise.
4429 (*ifcompare_move_not): Likewise.
4430 (*if_move_not): Likewise.
4431 (*ifcompare_not_move): Likewise.
4432 (*if_not_move): Likewise.
4433 (*ifcompare_shift_move): Likewise.
4434 (*if_shift_move): Likewise.
4435 (*ifcompare_move_shift): Likewise.
4436 (*if_move_shift): Likewise.
4437 (*ifcompare_shift_shift): Likewise.
4438 (*ifcompare_not_arith): Likewise.
4439 (*ifcompare_arith_not): Likewise.
4440 (*if_arith_not): Likewise.
4441 (*ifcompare_neg_move): Likewise.
4442 (*if_neg_move): Likewise.
4443 (*ifcompare_move_neg): Likewise.
4444 (*if_move_neg): Likewise.
4445 (prologue_thumb1_interwork): Likewise.
4446 (*cond_move_not): Likewise.
4447 (*sign_extract_onebit): Likewise.
4448 (*not_signextract_onebit): Likewise.
4449 (stack_tie): Likewise.
4450 (align_4): Likewise.
4451 (align_8): Likewise.
4452 (consttable_end): Likewise.
4453 (consttable_1): Likewise.
4454 (consttable_2): Likewise.
4455 (consttable_4): Likewise.
4456 (consttable_8): Likewise.
4457 (consttable_16): Likewise.
4458 (*thumb1_tablejump): Likewise.
4459 (prefetch): Likewise.
4460 (force_register_use): Likewise.
4461 (thumb_eh_return): Likewise.
4462 (load_tp_hard): Likewise.
4463 (load_tp_soft): Likewise.
4464 (tlscall): Likewise.
4465 (*arm_movtas_ze): Likewise.
4466 (*arm_rev): Likewise.
4467 (*arm_revsh): Likewise.
4468 (*arm_rev16): Likewise.
4469 * config/arm/thumb2.md
4470 (*thumb2_smaxsi3): Likewise.
4471 (*thumb2_sminsi3): Likewise.
4472 (*thumb32_umaxsi3): Likewise.
4473 (*thumb2_uminsi3): Likewise.
4474 (*thumb2_negdi2): Likewise.
4475 (*thumb2_abssi2): Likewise.
4476 (*thumb2_neg_abss): Likewise.
4477 (*thumb2_movsi_insn): Likewise.
4478 (tls_load_dot_plus_four): Likewise.
4479 (*thumb2_movhi_insn): Likewise.
4480 (*thumb2_mov_scc): Likewise.
4481 (*thumb2_mov_negs): Likewise.
4482 (*thumb2_mov_negs): Likewise.
4483 (*thumb2_mov_nots): Likewise.
4484 (*thumb2_mov_nots): Likewise.
4485 (*thumb2_movsicc_): Likewise.
4486 (*thumb2_movsfcc_soft_insn): Likewise.
4487 (*thumb2_indirect_jump): Likewise.
4488 (*thumb2_and_scc): Likewise.
4489 (*thumb2_ior_scc): Likewise.
4490 (*thumb2_ior_scc_strict_it): Likewise.
4491 (*thumb2_cond_move): Likewise.
4492 (*thumb2_cond_arith): Likewise.
4493 (*thumb2_cond_ari): Likewise.
4494 (*thumb2_cond_sub): Likewise.
4495 (*thumb2_negscc): Likewise.
4496 (*thumb2_movcond): Likewise.
4497 (thumb2_casesi_internal): Likewise.
4498 (thumb2_casesi_internal_pic): Likewise.
4499 (*thumb2_alusi3_short): Likewise.
4500 (*thumb2_mov<mode>_shortim): Likewise.
4501 (*thumb2_addsi_short): Likewise.
4502 (*thumb2_subsi_short): Likewise.
4503 (thumb2_addsi3_compare0): Likewise.
4504 (*thumb2_cbz): Likewise.
4505 (*thumb2_cbnz): Likewise.
4506 (*thumb2_one_cmplsi2_short): Likewise.
4507 (*thumb2_negsi2_short): Likewise.
4508 (*orsi_notsi_si): Likewise.
4509 * config/arm/arm1020e.md: Update with new attributes.
4510 * config/arm/arm1026ejs.md: Update with new attributes.
4511 * config/arm/arm1136jfs.md: Update with new attributes.
4512 * config/arm/arm926ejs.md: Update with new attributes.
4513 * config/arm/cortex-a15.md: Update with new attributes.
4514 * config/arm/cortex-a5.md: Update with new attributes.
4515 * config/arm/cortex-a53.md: Update with new attributes.
4516 * config/arm/cortex-a7.md: Update with new attributes.
4517 * config/arm/cortex-a8.md: Update with new attributes.
4518 * config/arm/cortex-a9.md: Update with new attributes.
4519 * config/arm/cortex-m4.md: Update with new attributes.
4520 * config/arm/cortex-r4.md: Update with new attributes.
4521 * config/arm/fa526.md: Update with new attributes.
4522 * config/arm/fa606te.md: Update with new attributes.
4523 * config/arm/fa626te.md: Update with new attributes.
4524 * config/arm/fa726te.md: Update with new attributes.
4526 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4528 * config/aarch64/aarch64-simd.md
4529 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
4530 <vwx> iterator to ensure correct register choice.
4531 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4532 (aarch64_sqdmull_n<mode>): Likewise.
4533 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4534 * config/aarch64/arm_neon.h
4535 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
4536 (vml<as><q>_n_<su>16): Likewise.
4537 (vml<as>l_high_lane<q>_<su>16): Likewise.
4538 (vml<as>l_high_n_<su>16): Likewise.
4539 (vml<as>l_lane<q>_<su>16): Likewise.
4540 (vml<as>l_n_<su>16): Likewise.
4541 (vmul<q>_lane<q>_<su>16): Likewise.
4542 (vmul<q>_n_<su>16): Likewise.
4543 (vmull_lane<q>_<su>16): Likewise.
4544 (vmull_n_<su>16): Likewise.
4545 (vmull_high_lane<q>_<su>16): Likewise.
4546 (vmull_high_n_<su>16): Likewise.
4547 (vqrdmulh<q>_n_s16): Likewise.
4549 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
4551 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
4552 have the correct lane parameter.
4554 2013-09-06 Richard Biener <rguenther@suse.de>
4556 * cfganal.c (control_dependences::~control_dependences):
4557 Properly free all of the vector.
4559 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
4562 * config/i386/i386.c (ix86_conditional_register_usage):
4563 Proper initialize extended SSE registers.
4565 2013-09-06 Jan Hubicka <jh@suse.cz>
4567 PR tree-optimization/58311
4568 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
4570 2013-09-06 Jan Hubicka <jh@suse.cz>
4572 * Makefile.in (tree-sra.o): Update dependencies.
4573 * tree-sra.c: Include ipa-utils.h
4574 (scan_function): Use recursive_call_p.
4575 (has_caller_p): New function.
4576 (cgraph_for_node_and_aliases): Count also callers of aliases.
4578 2013-09-06 Jan Hubicka <jh@suse.cz>
4581 * cgraph.h (symtab_semantically_equivalent_p): Declare.
4582 * tree-tailcall.c: Include ipa-utils.h.
4583 (find_tail_calls): Use it.
4584 * ipa-pure-const.c (check_call): Likewise.
4585 * ipa-utils.c (recursive_call_p): New function.
4586 * ipa-utils.h (recursive_call_p): Dclare.
4587 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
4588 (symtab_semantically_equivalent_p): New function.
4589 * Makefile.in (tree-tailcall.o): Update dependencies.
4591 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
4593 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
4596 2013-09-06 Richard Biener <rguenther@suse.de>
4598 * lto-streamer.h (lto_global_var_decls): Remove.
4599 * Makefile.in (OBJS): Remove lto-symtab.o.
4600 (lto-symtab.o): Remove.
4601 (GTFILES): Remove lto-symtab.c
4602 * lto-symtab.c: Move to lto/
4604 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4606 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
4607 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
4608 (UNSPEC_FPINT_RINT): New constant definitions.
4609 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
4610 definition with 2 attributes.
4611 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
4612 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
4615 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4617 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
4618 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
4619 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
4620 attribute to "z196_alone".
4621 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
4624 2013-09-06 Richard Biener <rguenther@suse.de>
4626 * basic-block.h (class control_dependences): New.
4627 * tree-ssa-dce.c (control_dependence_map): Remove.
4629 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
4630 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
4631 find_pdom, find_control_dependence, find_all_control_dependences):
4633 (mark_control_dependent_edges_necessary,
4634 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
4635 tree_dce_done, perform_tree_ssa_dce): Adjust.
4636 * cfganal.c (set_control_dependence_map_bit,
4637 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
4638 find_all_control_dependences): Move from tree-ssa-dce.c and
4639 implement as methods of control_dependences class.
4640 (control_dependences::control_dependences): New.
4641 (control_dependences::~control_dependences): Likewise.
4642 (control_dependences::get_edges_dependent_on): Likewise.
4643 (control_dependences::get_edge): Likewise.
4645 2013-09-04 Jan Hubicka <jh@suse.cz>
4647 * tree.c (types_same_for_odr): Drop overactive check.
4648 * ipa-devirt.c (hash_type_name): Likewise.
4650 2013-09-04 Jan Hubicka <jh@suse.cz>
4652 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
4653 (analyze_functions): ... here.
4655 2013-09-04 Jan Hubicka <jh@suse.cz>
4658 * cgraphunit.c (analyze_functions): Clear AUX fields
4659 after processing; initialize assembler name has.
4661 2013-09-05 Jeff Law <law@redhat.com>
4663 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
4664 from thread_around_empty_block. Record threading path into PATH.
4665 Recurse if threading through the initial block is successful.
4666 (thread_across_edge): Corresponding changes to slightly simplify.
4668 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4670 * config/aarch64/aarch64.md
4671 (type): Remove frecpe, frecps, frecpx.
4672 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
4673 fix to be a TARGET_SIMD instruction.
4674 (aarch64_frecps): Remove.
4675 * config/aarch64/aarch64-simd.md
4676 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
4677 (aarch64_frecps<mode>): Handle all float/vector of float modes.
4679 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4680 Sofiane Naci <sofiane.naci@arm.com>
4682 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
4683 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
4684 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
4685 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
4686 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
4687 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
4688 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
4689 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
4690 into "clz, "rbit". Rename "shift" to "shift_imm".
4691 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
4692 changes. Update for attribute changes all occurrences of arlo_* and
4694 * config/arm/arm-fixed.md: Update for attribute changes
4695 all occurrences of arlo_* types.
4696 * config/arm/thumb2.md: Update for attribute changes all occurrences
4698 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
4699 (cortexa7_older_only): Likewise.
4700 (cortexa7_younger): Likewise.
4701 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
4702 (1020alu_shift_op): Likewise.
4703 (1020alu_shift_reg_op): Likewise.
4704 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
4705 (alu_shift_op): Likewise.
4706 (alu_shift_reg_op): Likewise.
4707 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
4708 (11_alu_shift_op): Likewise.
4709 (11_alu_shift_reg_op): Likewise.
4710 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
4711 (9_alu_shift_reg_op): Likewise.
4712 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
4714 (cortex_a15_alu_shift): Likewise.
4715 (cortex_a15_alu_shift_reg): Likewise.
4716 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
4718 (cortex_a5_alu_shift): Likewise.
4719 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
4721 (cortex_a53_alu_shift): Likewise.
4722 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
4724 (cortex_a7_alu_reg): Likewise.
4725 (cortex_a7_alu_shift): Likewise.
4726 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
4728 (cortex_a8_alu_shift): Likewise.
4729 (cortex_a8_alu_shift_reg): Likewise.
4730 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
4731 (cortex_a9_dp_shift): Likewise.
4732 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
4734 * config/arm/cortex-r4.md
4735 (cortex_r4_alu): Update for attribute changes.
4736 (cortex_r4_mov): Likewise.
4737 (cortex_r4_alu_shift_reg): Likewise.
4738 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
4739 (526_alu_shift_op): Likewise.
4740 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
4741 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
4742 (626te_alu_shift_op): Likewise.
4743 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
4744 (726te_alu_shift_op): Likewise.
4745 (726te_alu_shift_reg_op): Likewise.
4746 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
4747 (mp626_alu_shift_op): Likewise.
4748 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
4749 (pj4_alu_conds): Likewise.
4750 (pj4_shift): Likewise.
4751 (pj4_shift_conds): Likewise.
4752 (pj4_alu_shift): Likewise.
4753 (pj4_alu_shift_conds): Likewise.
4754 * config/aarch64/aarch64.md: Update for attribute change
4755 all occurrences of arlo_* and shift* types.
4757 2013-09-05 Mike Stump <mikestump@comcast.net>
4759 * tree.h: Move documentation for tree_function_decl to tree-core.h
4760 with the declaration.
4762 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
4765 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
4766 looking for widest mode.
4768 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
4770 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
4772 2013-09-05 Richard Biener <rguenther@suse.de>
4774 PR tree-optimization/58137
4775 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
4776 Do not create vectors of pointers.
4777 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
4778 types for the components of the vector initializer.
4779 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
4780 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
4782 2013-09-05 Martin Jambor <mjambor@suse.cz>
4784 * ipa-prop.c (remove_described_reference): Accept missing references,
4785 return false if that hppens, otherwise return true.
4786 (cgraph_node_for_jfunc): New function.
4787 (try_decrement_rdesc_refcount): Likewise.
4788 (try_make_edge_direct_simple_call): Use them.
4789 (ipa_edge_removal_hook): Remove references from rdescs.
4790 (ipa_edge_duplication_hook): Clone rdescs and their references
4791 when the new edge has the same caller as the old one.
4792 * cgraph.c (cgraph_resolve_speculation): Remove speculative
4793 reference before removing any edges.
4795 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
4797 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
4799 * thumb2.md (thumb2_storewb_parisi): New pattern.
4801 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
4803 * config/aarch64/aarch64-option-extensions.def: Add
4804 AARCH64_OPT_EXTENSION of 'crc'.
4805 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
4806 (AARCH64_ISA_CRC): Ditto.
4807 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
4808 description of the CRC extension.
4810 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
4812 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
4813 * config/rs6000/linux.h: Ditto.
4814 * alpha/linux.h: Ditto.
4815 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
4816 no_c99_libc_has_function.
4817 * config/c6x/uclinux-elf.h: Ditto.
4818 * config/lm32/uclinux-elf.h: Ditto.
4819 * config/m68k/uclinux.h: Ditto.
4820 * config/moxie/uclinux.h: Ditto.
4821 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
4822 (crisv32-*-linux*, cris-*-linux*): Ditto.
4823 * config/bfin/bfin.c: Include "tm_p.h".
4825 2013-09-05 Richard Biener <rguenther@suse.de>
4827 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
4828 check for a definition without a basic-block.
4830 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4831 Sofiane Naci <sofiane.naci@arm.com>
4833 * config/aarch64/aarch64.md
4834 (*movti_aarch64): Rename r_2_f and f_2_r.
4835 (*movsf_aarch64): Likewise.
4836 (*movdf_aarch64): Likewise.
4837 (*movtf_aarch64): Likewise.
4838 (aarch64_movdi_<mode>low): Likewise.
4839 (aarch64_movdi_<mode>high): Likewise.
4840 (aarch64_mov<mode>high_di): Likewise.
4841 (aarch64_mov<mode>low_di): Likewise.
4842 (aarch64_movtilow_tilow): Likewise.
4843 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
4844 values to config/arm/types.md
4845 (attribute "conds"): Update for attribute change.
4846 (anddi3_insn): Likewise.
4847 (iordi3_insn): Likewise.
4848 (xordi3_insn): Likewise.
4849 (one_cmpldi2): Likewise.
4850 * config/arm/types.md (type): Add Neon types.
4851 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
4852 use "type" attribute.
4853 (movmisalign<mode>_neon_store): Likewise.
4854 (movmisalign<mode>_neon_load): Likewise.
4855 (vec_set<mode>_internal): Likewise.
4856 (vec_setv2di_internal): Likewise.
4857 (vec_extract<mode>): Likewise.
4858 (vec_extractv2di): Likewise.
4859 (add<mode>3_neon): Likewise.
4860 (adddi3_neon): Likewise.
4861 (sub<mode>3_neon): Likewise.
4862 (subdi3_neon): Likewise.
4863 (mul<mode>3_neon): Likewise.
4864 (mul<mode>3add<mode>_neon): Likewise.
4865 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
4866 (fma<VCVTF:mode>4)): Likewise.
4867 (fma<VCVTF:mode>4_intrinsic): Likewise.
4868 (fmsub<VCVTF:mode>4)): Likewise.
4869 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4870 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4871 (ior<mode>3): Likewise.
4872 (and<mode>3): Likewise.
4873 (anddi3_neon): Likewise.
4874 (orn<mode>3_neon): Likewise.
4875 (orndi3_neon): Likewise.
4876 (bic<mode>3_neon): Likewise.
4877 (bicdi3_neon): Likewise.
4878 (xor<mode>3): Likewise.
4879 (one_cmpl<mode>2): Likewise.
4880 (abs<mode>2): Likewise.
4881 (neg<mode>2): Likewise.
4882 (umin<mode>3_neon): Likewise.
4883 (umax<mode>3_neon): Likewise.
4884 (smin<mode>3_neon): Likewise.
4885 (smax<mode>3_neon): Likewise.
4886 (vashl<mode>3): Likewise.
4887 (vashr<mode>3_imm): Likewise.
4888 (vlshr<mode>3_imm): Likewise.
4889 (ashl<mode>3_signed): Likewise.
4890 (ashl<mode>3_unsigned): Likewise.
4891 (neon_load_count): Likewise.
4892 (ashldi3_neon_noclobber): Likewise.
4893 (signed_shift_di3_neon): Likewise.
4894 (unsigned_shift_di3_neon): Likewise.
4895 (ashrdi3_neon_imm_noclobber): Likewise.
4896 (lshrdi3_neon_imm_noclobber): Likewise.
4897 (widen_ssum<mode>3): Likewise.
4898 (widen_usum<mode>3): Likewise.
4899 (quad_halves_<code>v4si): Likewise.
4900 (quad_halves_<code>v4sf): Likewise.
4901 (quad_halves_<code>v8hi): Likewise.
4902 (quad_halves_<code>v16qi): Likewise.
4903 (reduc_splus_v2di): Likewise.
4904 (neon_vpadd_internal<mode>): Likewise.
4905 (neon_vpsmin<mode>): Likewise.
4906 (neon_vpsmax<mode>): Likewise.
4907 (neon_vpumin<mode>): Likewise.
4908 (neon_vpumax<mode>): Likewise.
4909 (ss_add<mode>_neon): Likewise.
4910 (us_add<mode>_neon): Likewise.
4911 (ss_sub<mode>_neon): Likewise.
4912 (us_sub<mode>_neon): Likewise.
4913 (neon_vadd<mode>_unspec): Likewise.
4914 (neon_vaddl<mode>): Likewise.
4915 (neon_vaddw<mode>): Likewise.
4916 (neon_vhadd<mode>): Likewise.
4917 (neon_vqadd<mode>): Likewise.
4918 (neon_vaddhn<mode>): Likewise.
4919 (neon_vmul<mode>): Likewise.
4920 (neon_vmla<mode>): Likewise.
4921 (neon_vmlal<mode>): Likewise.
4922 (neon_vmls<mode>): Likewise.
4923 (neon_vmlsl<mode>): Likewise.
4924 (neon_vqdmulh<mode>): Likewise.
4925 (neon_vqdmlal<mode>): Likewise.
4926 (neon_vqdmlsl<mode>): Likewise.
4927 (neon_vmull<mode>): Likewise.
4928 (neon_vqdmull<mode>): Likewise.
4929 (neon_vsub<mode>_unspec): Likewise.
4930 (neon_vsubl<mode>): Likewise.
4931 (neon_vsubw<mode>): Likewise.
4932 (neon_vqsub<mode>): Likewise.
4933 (neon_vhsub<mode>): Likewise.
4934 (neon_vsubhn<mode>): Likewise.
4935 (neon_vceq<mode>): Likewise.
4936 (neon_vcge<mode>): Likewise.
4937 (neon_vcgeu<mode>): Likewise.
4938 (neon_vcgt<mode>): Likewise.
4939 (neon_vcgtu<mode>): Likewise.
4940 (neon_vcle<mode>): Likewise.
4941 (neon_vclt<mode>): Likewise.
4942 (neon_vcage<mode>): Likewise.
4943 (neon_vcagt<mode>): Likewise.
4944 (neon_vtst<mode>): Likewise.
4945 (neon_vabd<mode>): Likewise.
4946 (neon_vabdl<mode>): Likewise.
4947 (neon_vaba<mode>): Likewise.
4948 (neon_vabal<mode>): Likewise.
4949 (neon_vmax<mode>): Likewise.
4950 (neon_vmin<mode>): Likewise.
4951 (neon_vpaddl<mode>): Likewise.
4952 (neon_vpadal<mode>): Likewise.
4953 (neon_vpmax<mode>): Likewise.
4954 (neon_vpmin<mode>): Likewise.
4955 (neon_vrecps<mode>): Likewise.
4956 (neon_vrsqrts<mode>): Likewise.
4957 (neon_vqabs<mode>): Likewise.
4958 (neon_vqneg<mode>): Likewise.
4959 (neon_vcls<mode>): Likewise.
4960 (clz<mode>2): Likewise.
4961 (popcount<mode>2): Likewise.
4962 (neon_vrecpe): Likewise.
4963 (neon_vrsqrte): Likewise.
4964 (neon_vget_lane<mode>_sext_internal): Likewise.
4965 (neon_vget_lane<mode>_zext_internal): Likewise.
4966 (neon_vdup_n<mode>): Likewise.
4967 (neon_vdup_nv2di): Likewise.
4968 (neon_vdpu_lane<mode>_internal): Likewise.
4969 (neon_vswp<mode>): Likewise.
4970 (float<mode><V_cvtto>2): Likewise.
4971 (floatuns<mode><V_cvtto>2): Likewise.
4972 (fix_trunc<mode><V_cvtto>)2): Likewise
4973 (fixuns_trunc<mode><V_cvtto)2): Likewise.
4974 (neon_vcvt<mode>): Likewise.
4975 (neon_vcvtv4sfv4hf): Likewise.
4976 (neon_vcvtv4hfv4sf): Likewise.
4977 (neon_vcvt_n<mode>): Likewise.
4978 (neon_vmovn<mode>): Likewise.
4979 (neon_vqmovn<mode>): Likewise.
4980 (neon_vqmovun<mode>): Likewise.
4981 (neon_vmovl<mode>): Likewise.
4982 (neon_vmul_lane<mode>): Likewise.
4983 (neon_vmull_lane<mode>): Likewise.
4984 (neon_vqdmull_lane<mode>): Likewise.
4985 (neon_vqdmulh_lane<mode>): Likewise.
4986 (neon_vmla_lane<mode>): Likewise.
4987 (neon_vmlal_lane<mode>): Likewise.
4988 (neon_vqdmlal_lane<mode>): Likewise.
4989 (neon_vmls_lane<mode>): Likewise.
4990 (neon_vmlsl_lane<mode>): Likewise.
4991 (neon_vqdmlsl_lane<mode>): Likewise.
4992 (neon_vext<mode>): Likewise.
4993 (neon_vrev64<mode>): Likewise.
4994 (neon_vrev32<mode>): Likewise.
4995 (neon_vrev16<mode>): Likewise.
4996 (neon_vbsl<mode>_internal): Likewise.
4997 (neon_vshl<mode>): Likewise.
4998 (neon_vqshl<mode>): Likewise.
4999 (neon_vshr_n<mode>): Likewise.
5000 (neon_vshrn_n<mode>): Likewise.
5001 (neon_vqshrn_n<mode>): Likewise.
5002 (neon_vqshrun_n<mode>): Likewise.
5003 (neon_vshl_n<mode>): Likewise.
5004 (neon_vqshl_n<mode>): Likewise.
5005 (neon_vqshlu_n<mode>): Likewise.
5006 (neon_vshll_n<mode>): Likewise.
5007 (neon_vsra_n<mode>): Likewise.
5008 (neon_vsri_n<mode>): Likewise.
5009 (neon_vsli_n<mode>): Likewise.
5010 (neon_vtbl1v8qi): Likewise.
5011 (neon_vtbl2v8qi): Likewise.
5012 (neon_vtbl3v8qi): Likewise.
5013 (neon_vtbl4v8qi): Likewise.
5014 (neon_vtbx1v8qi): Likewise.
5015 (neon_vtbx2v8qi): Likewise.
5016 (neon_vtbx3v8qi): Likewise.
5017 (neon_vtbx4v8qi): Likewise.
5018 (neon_vtrn<mode>_internal): Likewise.
5019 (neon_vzip<mode>_internal): Likewise.
5020 (neon_vuzp<mode>_internal): Likewise.
5021 (neon_vld1<mode>): Likewise.
5022 (neon_vld1_lane<mode>): Likewise.
5023 (neon_vld1_dup<mode>): Likewise.
5024 (neon_vld1_dupv2di): Likewise.
5025 (neon_vst1<mode>): Likewise.
5026 (neon_vst1_lane<mode>): Likewise.
5027 (neon_vld2<mode>): Likewise.
5028 (neon_vld2_lane<mode>): Likewise.
5029 (neon_vld2_dup<mode>): Likewise.
5030 (neon_vst2<mode>): Likewise.
5031 (neon_vst2_lane<mode>): Likewise.
5032 (neon_vld3<mode>): Likewise.
5033 (neon_vld3qa<mode>): Likewise.
5034 (neon_vld3qb<mode>): Likewise.
5035 (neon_vld3_lane<mode>): Likewise.
5036 (neon_vld3_dup<mode>): Likewise.
5037 (neon_vst3<mode>): Likewise.
5038 (neon_vst3qa<mode>): Likewise.
5039 (neon_vst3qb<mode>): Likewise.
5040 (neon_vst3_lane<mode>): Likewise.
5041 (neon_vld4<mode>): Likewise.
5042 (neon_vld4qa<mode>): Likewise.
5043 (neon_vld4qb<mode>): Likewise.
5044 (neon_vld4_lane<mode>): Likewise.
5045 (neon_vld4_dup<mode>): Likewise.
5046 (neon_vst4<mode>): Likewise.
5047 (neon_vst4qa<mode>): Likewise.
5048 (neon_vst4qb<mode>): Likewise.
5049 (neon_vst4_lane<mode>): Likewise.
5050 (neon_vec_unpack<US>_lo_<mode>): Likewise.
5051 (neon_vec_unpack<US>_hi_<mode>): Likewise.
5052 (neon_vec_<US>mult_lo_<mode>): Likewise.
5053 (neon_vec_<US>mult_hi_<mode>): Likewise.
5054 (neon_vec_<US>shiftl_<mode>): Likewise.
5055 (neon_unpack<US>_<mode>): Likewise.
5056 (neon_vec_<US>mult_<mode>): Likewise.
5057 (vec_pack_trunc_<mode>): Likewise.
5058 (neon_vec_pack_trunk_<mode>): Likewise.
5059 (neon_vabd<mode>_2): Likewise.
5060 (neon_vabd<mode>_3): Likewise.
5061 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5062 (thumb2_movsi_vfp): Likewise.
5063 (movdi_vfp): Likewise.
5064 (movdi_vfp_cortexa8): Likewise.
5065 (movhf_vfp_neon): Likewise.
5066 (movhf_vfp): Likewiwse.
5067 (movsf_vfp): Likewiwse.
5068 (thumb2_movsf_vfp): Likewiwse.
5069 (movdf_vfp): Likewise.
5070 (thumb2_movdf_vfp): Likewise.
5071 (movsfcc_vfp): Likewise.
5072 (thumb2_movsfcc_vfp): Likewise.
5073 (movdfcc_vfp): Likewise.
5074 (thumb2_movdfcc_vfp): Likewise.
5075 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
5076 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
5077 (v10_v2c): Likewise.
5078 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
5080 (cortex_a15_neon_int_2): Likewise.
5081 (cortex_a15_neon_int_3): Likewise.
5082 (cortex_a15_neon_int_4): Likewise.
5083 (cortex_a15_neon_int_5): Likewise.
5084 (cortex_a15_neon_vqneg_vqabs): Likewise.
5085 (cortex_a15_neon_vmov): Likewise.
5086 (cortex_a15_neon_vaba): Likewise.
5087 (cortex_a15_neon_vaba_qqq): Likewise.
5088 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5089 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5090 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
5091 scalar_64_32_long_scalar): Likewise.
5092 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5093 (cortex_a15_neon_mla_qqq_8_16): Likewise.
5094 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5095 lotype_qdd_64_32_long): Likewise.
5096 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5097 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5098 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
5099 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5100 (cortex_a15_neon_shift_1): Likewise.
5101 (cortex_a15_neon_shift_2): Likewise.
5102 (cortex_a15_neon_shift_3): Likewise.
5103 (cortex_a15_neon_vshl_ddd): Likewise.
5104 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5105 (cortex_a15_neon_vsra_vrsra): Likewise.
5106 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
5107 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
5108 (cortex_a15_neon_fp_vmul_ddd): Likewise.
5109 (cortex_a15_neon_fp_vmul_qqd): Likewise.
5110 (cortex_a15_neon_fp_vmla_ddd): Likewise.
5111 (cortex_a15_neon_fp_vmla_qqq): Likewise.
5112 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5113 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5114 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5115 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5116 (cortex_a15_neon_bp_simple): Likewise.
5117 (cortex_a15_neon_bp_2cycle): Likewise.
5118 (cortex_a15_neon_bp_3cycle): Likewise.
5119 (cortex_a15_neon_vld1_1_2_regs): Likewise.
5120 (cortex_a15_neon_vld1_3_4_regs): Likewise.
5121 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5122 (cortex_a15_neon_vld2_4_regs): Likewise.
5123 (cortex_a15_neon_vld3_vld4): Likewise.
5124 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5125 (cortex_a15_neon_vst1_3_4_regs): Likewise.
5126 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
5127 (cortex_a15_neon_vst3_vst4): Likewise.
5128 (cortex_a15_neon_vld1_vld2_lane): Likewise.
5129 (cortex_a15_neon_vld3_vld4_lane" 10
5130 (cortex_a15_neon_vst1_vst2_lane): Likewise.
5131 (cortex_a15_neon_vst3_vst4_lane): Likewise.
5132 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
5133 (cortex_a15_neon_ldm_2): Likewise.0
5134 (cortex_a15_neon_stm_2): Likewise.
5135 (cortex_a15_neon_mcr): Likewise.
5136 (cortex_a15_neon_mcr_2_mcrr): Likewise.
5137 (cortex_a15_neon_mrc): Likewise.
5138 (cortex_a15_neon_mrrc): Likewise.
5139 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5141 (cortex_a15_alu_shift): Likewise.
5142 (cortex_a15_alu_shift_reg): Likewise.
5143 (cortex_a15_mult32): Likewise.
5144 (cortex_a15_mult64): Likewise.
5145 (cortex_a15_block): Likewise.
5146 (cortex_a15_branch): Likewise.
5147 (cortex_a15_load1): Likewise.
5148 (cortex_a15_load3): Likewise.
5149 (cortex_a15_store1): Likewise.
5150 (cortex_a15_store3): Likewise.
5151 (cortex_a15_call): Likewise.
5152 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
5153 (cortex_a5_f2r): Likewise.
5154 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
5156 (cortex_a53_f2r): Likewise.
5157 * config/arm/cortex-a7.md
5158 (cortex_a7_branch): Update for attribute change.
5159 (cortex_a7_call): Likewise.
5160 (cortex_a7_alu_imm): Likewise.
5161 (cortex_a7_alu_reg): Likewise.
5162 (cortex_a7_alu_shift): Likewise.
5163 (cortex_a7_mul): Likewise.
5164 (cortex_a7_load1): Likewise.
5165 (cortex_a7_store1): Likewise.
5166 (cortex_a7_load2): Likewise.
5167 (cortex_a7_store2): Likewise.
5168 (cortex_a7_load3): Likewise.
5169 (cortex_a7_store3): Likewise.
5170 (cortex_a7_load4): Likewise.
5171 (cortex_a7_store4): Likewise.
5172 (cortex_a7_fpalu): Likewise.
5173 (cortex_a7_fconst): Likewise.
5174 (cortex_a7_fpmuls): Likewise.
5175 (cortex_a7_neon_mul): Likewise.
5176 (cortex_a7_fpmacs): Likewise.
5177 (cortex_a7_neon_mla: Likewise.
5178 (cortex_a7_fpmuld: Likewise.
5179 (cortex_a7_fpmacd: Likewise.
5180 (cortex_a7_fpfmad: Likewise.
5181 (cortex_a7_fdivs: Likewise.
5182 (cortex_a7_fdivd: Likewise.
5183 (cortex_a7_r2f: Likewise.
5184 (cortex_a7_f2r: Likewise.
5185 (cortex_a7_f_flags: Likewise.
5186 (cortex_a7_f_loads: Likewise.
5187 (cortex_a7_f_loadd: Likewise.
5188 (cortex_a7_f_stores: Likewise.
5189 (cortex_a7_f_stored: Likewise.
5190 (cortex_a7_neon): Likewise.
5191 * config/arm/cortex-a8-neon.md
5192 (cortex_a8_neon_mrc): Update for attribute change.
5193 (cortex_a8_neon_mrrc): Likewise.
5194 (cortex_a8_neon_int_1): Likewise.
5195 (cortex_a8_neon_int_2): Likewise.
5196 (cortex_a8_neon_int_3): Likewise.
5197 (cortex_a8_neon_int_4): Likewise.
5198 (cortex_a8_neon_int_5): Likewise.
5199 (cortex_a8_neon_vqneg_vqabs): Likewise.
5200 (cortex_a8_neon_vmov): Likewise.
5201 (cortex_a8_neon_vaba): Likewise.
5202 (cortex_a8_neon_vaba_qqq): Likewise.
5203 (cortex_a8_neon_vsma): Likewise.
5204 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5205 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5206 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
5207 long_scalar): Likewise.
5208 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5209 (cortex_a8_neon_mla_qqq_8_16): Likewise.
5210 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5211 long_scalar_qdd_64_32_long): Likewise.
5212 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5213 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5214 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
5215 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5216 (cortex_a8_neon_shift_1): Likewise.
5217 (cortex_a8_neon_shift_2): Likewise.
5218 (cortex_a8_neon_shift_3): Likewise.
5219 (cortex_a8_neon_vshl_ddd): Likewise.
5220 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5221 (cortex_a8_neon_vsra_vrsra): Likewise.
5222 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
5223 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
5224 (cortex_a8_neon_fp_vsum): Likewise.
5225 (cortex_a8_neon_fp_vmul_ddd): Likewise.
5226 (cortex_a8_neon_fp_vmul_qqd): Likewise.
5227 (cortex_a8_neon_fp_vmla_ddd): Likewise.
5228 (cortex_a8_neon_fp_vmla_qqq): Likewise.
5229 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
5230 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
5231 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5232 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5233 (cortex_a8_neon_bp_simple): Likewise.
5234 (cortex_a8_neon_bp_2cycle): Likewise.
5235 (cortex_a8_neon_bp_3cycle): Likewise.
5236 (cortex_a8_neon_ldr): Likewise.
5237 (cortex_a8_neon_str): Likewise.
5238 (cortex_a8_neon_vld1_1_2_regs): Likewise.
5239 (cortex_a8_neon_vld1_3_4_regs): Likewise.
5240 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5241 (cortex_a8_neon_vld2_4_regs): Likewise.
5242 (cortex_a8_neon_vld3_vld4): Likewise.
5243 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5244 (cortex_a8_neon_vst1_3_4_regs): Likewise.
5245 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
5246 (cortex_a8_neon_vst3_vst4): Likewise.
5247 (cortex_a8_neon_vld1_vld2_lane): Likewise.
5248 (cortex_a8_neon_vld3_vld4_lane): Likewise.
5249 (cortex_a8_neon_vst1_vst2_lane): Likewise.
5250 (cortex_a8_neon_vst3_vst4_lane): Likewise.
5251 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
5252 (cortex_a8_neon_mcr): Likewise.
5253 (cortex_a8_neon_mcr_2_mcrr): Likewise.
5254 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5255 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
5257 (ca9_neon_mrrc): Likewise.
5258 (cortex_a9_neon_int_1): Likewise.
5259 (cortex_a9_neon_int_2): Likewise.
5260 (cortex_a9_neon_int_3): Likewise.
5261 (cortex_a9_neon_int_4): Likewise.
5262 (cortex_a9_neon_int_5): Likewise.
5263 (cortex_a9_neon_vqneg_vqabs): Likewise.
5264 (cortex_a9_neon_vmov): Likewise.
5265 (cortex_a9_neon_vaba): Likewise.
5266 (cortex_a9_neon_vaba_qqq): Likewise.
5267 (cortex_a9_neon_vsma): Likewise.
5268 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5269 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5270 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
5271 long_scalar): Likewise.
5272 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5273 (cortex_a9_neon_mla_qqq_8_16): Likewise.
5274 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5275 long_scalar_qdd_64_32_long): Likewise.
5276 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5277 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5278 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
5279 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5280 (cortex_a9_neon_shift_1): Likewise.
5281 (cortex_a9_neon_shift_2): Likewise.
5282 (cortex_a9_neon_shift_3): Likewise.
5283 (cortex_a9_neon_vshl_ddd): Likewise.
5284 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5285 (cortex_a9_neon_vsra_vrsra): Likewise.
5286 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
5287 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
5288 (cortex_a9_neon_fp_vsum): Likewise.
5289 (cortex_a9_neon_fp_vmul_ddd): Likewise.
5290 (cortex_a9_neon_fp_vmul_qqd): Likewise.
5291 (cortex_a9_neon_fp_vmla_ddd): Likewise.
5292 (cortex_a9_neon_fp_vmla_qqq): Likewise.
5293 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
5294 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
5295 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5296 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5297 (cortex_a9_neon_bp_simple): Likewise.
5298 (cortex_a9_neon_bp_2cycle): Likewise.
5299 (cortex_a9_neon_bp_3cycle): Likewise.
5300 (cortex_a9_neon_ldr): Likewise.
5301 (cortex_a9_neon_str): Likewise.
5302 (cortex_a9_neon_vld1_1_2_regs): Likewise.
5303 (cortex_a9_neon_vld1_3_4_regs): Likewise.
5304 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5305 (cortex_a9_neon_vld2_4_regs): Likewise.
5306 (cortex_a9_neon_vld3_vld4): Likewise.
5307 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5308 (cortex_a9_neon_vst1_3_4_regs): Likewise.
5309 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
5310 (cortex_a9_neon_vst3_vst4): Likewise.
5311 (cortex_a9_neon_vld1_vld2_lane): Likewise.
5312 (cortex_a9_neon_vld3_vld4_lane): Likewise.
5313 (cortex_a9_neon_vst1_vst2_lane): Likewise.
5314 (cortex_a9_neon_vst3_vst4_lane): Likewise.
5315 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5316 (cortex_a9_neon_mcr): Likewise.
5317 (cortex_a9_neon_mcr_2_mcrr): Likewise.
5318 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
5319 (cortex_a9_fps): Likewise.
5320 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
5322 (cortex_m4_fmuls): Likewise.
5323 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
5325 (cortex_r4_mrc): Likewise.
5326 * config/arm/iterators.md: Update comment referring to neon_type.
5327 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
5328 (iwmmxt_movsi_insn): Likewise.
5329 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
5331 (pj4_core_to_vfp): Likewise.
5332 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
5334 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
5335 (vfp_fstore): Likewise.
5336 * doc/md.texi: Change references to neon_type to refer to type.
5338 2013-09-04 Dodji Seketeli <dodji@redhat.com>
5340 * tree.h (DECL_BUILT_IN): Fix typo in comment.
5342 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
5344 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
5347 2013-09-04 Easwaran Raman <eraman@google.com>
5350 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
5351 (build_and_add_sum): Use it.
5352 (appears_later_in_bb): Simplify code.
5354 2013-09-04 Teresa Johnson <tejohnson@google.com>
5356 * dumpfile.c (dump_finish): Don't close stderr/stdout.
5358 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
5360 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
5362 2013-09-04 Jan Hubicka <jh@suse.cz>
5364 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
5365 * ipa-devirt.c: Include diganostic.h
5366 (odr_type_d): Add types and types_set.
5367 (hash_type_name): Work for types with vtables during LTO.
5368 (odr_hasher::remove): Fix comment; destroy types_set.
5369 (add_type_duplicate): New function,
5370 (get_odr_type): Use it.
5371 (dump_type_inheritance_graph): Dump type duplicates.
5372 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
5374 * tree.c (types_same_for_odr): Give exact answers on types with
5377 2013-09-04 Dodji Seketeli <dodji@redhat.com>
5379 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
5380 explaining their differences.
5382 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
5384 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
5386 2013-09-03 Jeff Law <law@redhat.com>
5388 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
5389 when not threading through a joiner block. Pass joiner/no joiner
5390 state to register_jump_thread.
5391 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
5392 state from argument rather than implying on path length.
5393 Dump the entire jump thread path into debugging dump.
5394 * tree-flow.h (register_jump_thread): Update prototype.
5396 2013-08-29 Xinliang David Li <davidxl@google.com>
5398 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
5399 Remove a trivial gcc_assert.
5401 2013-08-29 Xinliang David Li <davidxl@google.com>
5403 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
5404 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
5405 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
5406 Delay base decl alignment adjustment.
5407 * tree-vectorizer.c (vect_destroy_datarefs): New function.
5408 * tree-vectorizer.h: New data structure.
5409 (set_dr_misalignment): New function.
5410 (dr_misalignment): Ditto.
5411 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
5412 (vectorizable_load): Ditto.
5413 (ensure_base_align): New function.
5414 (vectorize_loops): Add dbg_cnt support.
5415 (execute_vect_slp): Ditto.
5416 * dbgcnt.def: New debug counter.
5417 * Makefile: New dependency.
5419 2013-09-03 Meador Inge <meadori@codesourcery.com>
5423 2013-08-30 Meador Inge <meadori@codesourcery.com>
5425 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
5427 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
5429 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
5430 function descriptor.
5432 2013-09-03 Richard Biener <rguenther@suse.de>
5434 * tree-affine.c (add_elt_to_tree): Fix association issue,
5435 avoid useless converts and make sure to always return a
5436 properly typed result.
5438 2013-09-03 Richard Biener <rguenther@suse.de>
5441 * fold-const.c (negate_expr_p): Fix division case.
5442 (negate_expr): Likewise.
5444 2013-09-03 Richard Biener <rguenther@suse.de>
5447 * tree-streamer-out.c: Include tm.h.
5448 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
5450 2013-09-03 Jan Hubicka <jh@suse.cz>
5452 * tree-profile.c (tree_profiling): Cleanup CFG when done.
5454 2013-09-03 Alan Modra <amodra@gmail.com>
5456 * config.gcc (powerpc*-*-linux*): Add support for little-endian
5457 multilibs to big-endian target and vice versa.
5458 * config/rs6000/t-linux64: Use := assignment on all vars.
5459 (MULTILIB_EXTRA_OPTS): Remove fPIC.
5460 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
5461 * config/rs6000/t-linux64le: New file.
5462 * config/rs6000/t-linux64bele: New file.
5463 * config/rs6000/t-linux64lebe: New file.
5465 2013-09-02 Jan Hubicka <jh@suse.cz>
5467 * ipa-inline-transform.c (inline_transform): Do not
5468 optimize_inline_calls when not optimizing.
5470 2013-09-02 Jan Hubicka <jh@suse.cz>
5472 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
5473 duplicated nodes for assembler names.
5474 * symtab.c (symtab_unregister_node): Do not attempt to unlink
5475 hard registers from assembler name hash.
5477 2013-09-02 Jan Hubicka <jh@suse.cz>
5479 * ipa-split.c (execute_split_functions): Split externally visible
5480 functions called once.
5482 2013-09-02 Martin Jambor <mjambor@suse.cz>
5485 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
5486 linked list. When finding the correct duplicate, also consider also
5487 the caller in additon to its inlined_to node.
5489 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
5491 * config/aarch64/aarch64-simd-builtins.def
5492 (dup_lane_scalar): Remove.
5493 * config/aarch64/aarch64-simd.md
5494 (aarch64_simd_dup): Add 'w->w' alternative.
5495 (aarch64_dup_lane<mode>): Allow for VALL.
5496 (aarch64_dup_lane_scalar<mode>): Remove.
5497 (aarch64_dup_lane_<vswap_width_name><mode>): New.
5498 (aarch64_get_lane_signed<mode>): Add w->w altenative.
5499 (aarch64_get_lane_unsigned<mode>): Likewise.
5500 (aarch64_get_lane<mode>): Likewise.
5501 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
5502 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
5503 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
5504 (VCON): Change container of V2SF.
5505 (vswap_width_name): Likewise.
5506 * config/aarch64/arm_neon.h
5507 (__aarch64_vdup_lane_any): New.
5508 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
5509 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
5510 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
5512 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
5515 * expr.c (emit_move_complex): Do not move complex FP values as parts if
5516 the source or the destination is a single hard register.
5518 2013-09-02 Richard Biener <rguenther@suse.de>
5521 * tree-scalar-evolution.c (instantiate_scev_name): Allow
5524 2013-09-02 Richard Biener <rguenther@suse.de>
5526 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
5527 arithmetic to sizetype.
5529 2013-09-02 Bin Cheng <bin.cheng@arm.com>
5531 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
5532 Find auto-increment use both before and after candidate.
5534 2013-09-02 Marek Polacek <polacek@redhat.com>
5536 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
5538 2013-09-01 Jan Hubicka <jh@suse.cz>
5540 * Makefile.in: Add ipa-profile.o
5541 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
5542 * cgraph.c (struct cgraph_propagate_frequency_data,
5543 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
5544 ipa-profile.c; replace cgraph_ by ipa_ prefix.
5545 * cgraph.h (cgraph_propagate_frequency): Remove.
5546 * ipa-inline-analysis.c: Include ipa-utils.h;
5547 drop duplicated cfgloop.h.
5548 (inline_update_callee_summaries): Update.
5549 * ipa-profile.c: New file.
5550 * ipa-utils.h (ipa_propagate_frequency): Declare.
5551 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
5552 data-streamer.h, value-prof.h.
5553 (symtab_remove_unreachable_nodes): Update profile.
5554 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
5555 account_time_size, cmp_counts, dump_histogram,
5556 ipa_profile_generate_summary, ipa_profile_write_summary,
5557 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
5558 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
5559 Move to ipa-profile.c.
5561 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
5563 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
5565 2013-09-01 Jan Hubicka <jh@suse.cz>
5567 * common.opt (fdevirtualize-speculatively): New function.
5568 * invoke.texi (fdevirtualize-speculatively): Document.
5569 * ipa-devirt.c: Include ipa-inline.h
5570 (likely_target_p): New function.
5571 (ipa_devirt): New function.
5572 (gate_ipa_devirt): New function.
5573 (pass_data_ipa_devirt): New static var.
5574 (pass_ipa_devirt): Likewise.
5575 (make_pass_ipa_devirt): New function.
5576 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
5577 (common_handle_option): Disable devirtualization when
5578 value range profiling is available.
5579 * passes.def (pass_ipa_devirt): Add.
5580 * timever.def (TV_IPA_DEVIRT): New timevar.
5581 * tree-pass.h (make_pass_ipa_devirt):
5583 2013-09-01 Iain Sandoe <iain@codesourcery.com>
5585 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
5586 include sanitize(undefined).
5588 2013-08-31 Diego Novillo <dnovillo@google.com>
5590 * Makefile.in (TREE_CORE_H): Define.
5592 (GTFILES): Add tree-core.h.
5593 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
5595 * tree-core.h: New file.
5596 Move all data structures, enum, typedefs, global
5597 declarations and constants from ...
5600 2013-08-31 Jan Hubicka <jh@suse.cz>
5602 * bulitins.c (expand_builtin): Do not early exit for gcov
5603 instrumented functions.
5605 2013-08-31 Marek Polacek <polacek@redhat.com>
5607 * ubsan.c: Include tm_p.h.
5609 2013-08-31 Jan Hubicka <jh@suse.cz>
5611 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
5614 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
5615 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5616 * tree-cfg.c (verify_gimple_label): ... here.
5617 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
5618 (ipa_merge_profiles): New function.
5619 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
5620 (lto_input_function_body): Likewise.
5621 * ipa-utils.h (ipa_merge_profiles): Declare.
5622 * lto-streamer.h (lto_input_function_body): Update prototype.
5623 (emit_label_in_global_context_p): Remove.
5624 * lto-symtab.c: Include ipa-utils.h
5625 (lto_cgraph_replace_node): Use ipa_merge_profiles.
5627 2013-08-31 Jan Hubicka <jh@suse.cz>
5629 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
5631 2013-08-31 Jan Hubicka <jh@suse.cz>
5633 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
5635 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
5637 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
5638 "cmp" RTX before signed_comparison_operator check to account
5641 2013-08-30 Jan Hubicka <jh@suse.cz>
5643 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
5644 (detect_type_change_1): Rename to ...
5645 (detect_type_change): ... this one; early return on non-polymorphic
5647 (detect_type_change_ssa): Add comp_type parameter; update
5648 use of detect_type_change.
5649 (compute_complex_assign_jump_func): Add param_type parameter;
5650 update use of detect_type_change_ssa.
5651 (compute_complex_ancestor_jump_func): Likewise.
5652 (ipa_get_callee_param_type): New function.
5653 (ipa_compute_jump_functions_for_edge): Compute parameter type;
5654 update calls to the jump function computation functions.
5656 2013-08-30 Teresa Johnson <tejohnson@google.com>
5657 Steven Bosscher <steven@gcc.gnu.org>
5659 * cfgrtl.c (fixup_new_cold_bb): New routine.
5660 (commit_edge_insertions): Invoke fixup_partitions.
5661 (find_partition_fixes): New routine.
5662 (fixup_partitions): Ditto.
5663 (verify_hot_cold_block_grouping): Update comments.
5664 (rtl_verify_edges): Invoke find_partition_fixes.
5665 (rtl_verify_bb_pointers): Update comments.
5666 (rtl_verify_bb_layout): Ditto.
5667 * basic-block.h (probably_never_executed_edge_p): Declare.
5668 (fixup_partitions): Ditto.
5669 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
5670 * bb-reorder.c (sanitize_hot_paths): New function.
5671 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
5673 * predict.c (probably_never_executed_edge_p): New routine.
5674 * cfg.c (check_bb_profile): Add partition insanity warnings.
5676 2013-08-30 Meador Inge <meadori@codesourcery.com>
5678 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
5680 2013-08-30 Marek Polacek <polacek@redhat.com>
5682 * Makefile.in (ubsan.o): Add.
5683 (c-family/c-ubsan.o): Add.
5684 (builtins.o): Add ubsan.h dependency.
5685 * ubsan.h: New file.
5686 * ubsan.c: New file.
5687 * common.opt: Add -fsanitize=undefined option.
5688 (flag_sanitize): Add variable.
5689 (fsanitize=): Add option. Add Driver.
5690 (fsanitize=thread): Remove option.
5691 (fsanitize=address): Likewise.
5692 (static-libubsan): New option.
5693 * doc/invoke.texi: Document the new flag and -static-libubsan.
5694 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
5695 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
5696 * builtin-attrs.def (ATTR_COLD): Define.
5697 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
5698 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
5699 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
5700 * flag-types.h (sanitize_code): New enum.
5701 * opts.c (common_handle_option): Parse command line arguments
5702 of -fsanitize=. Add -fsanitize=unreachable option.
5703 * varasm.c (get_variable_section): Adjust.
5704 (assemble_noswitch_variable): Likewise.
5705 (assemble_variable): Likewise.
5706 (output_constant_def_contents): Likewise.
5707 (categorize_decl_for_section): Likewise.
5708 (place_block_symbol): Likewise.
5709 (output_object_block): Likewise.
5710 * builtins.def: Likewise.
5711 * toplev.c (compile_file): Likewise.
5712 (process_options): Likewise.
5713 * cppbuiltin.c: Likewise.
5714 * tsan.c (tsan_pass): Likewise.
5715 (tsan_gate): Likewise.
5716 (tsan_gate_O0): Likewise.
5717 * cfgexpand.c (partition_stack_vars): Likewise.
5718 (expand_stack_vars): Likewise.
5719 (defer_stack_allocation): Likewise.
5720 (expand_used_vars): Likewise.
5721 * cfgcleanup.c (old_insns_match_p): Likewise.
5722 * asan.c (asan_finish_file): Likewise.
5723 (asan_instrument): Likewise.
5724 (gate_asan): Likewise.
5725 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
5726 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
5727 (asan_global_struct): Use pointer_sized_int_node instead
5728 calling build_nonstandard_integer_type.
5729 (initialize_sanitizer_builtins): Likewise.
5730 (asan_finish_file): Likewise.
5731 * gcc.c: Document %{%:function(args):X}.
5732 (static_spec_functions): Add sanitize.
5733 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
5734 store funcval != NULL there.
5735 (do_spec_1): Adjust handle_spec_function caller.
5736 (handle_braces): Allow %:function(args) as condition.
5737 (sanitize_spec_function): New function.
5738 (ADD_STATIC_LIBUBSAN_LIBS): Define.
5739 (LIBUBSAN_SPEC): Likewise.
5740 (LIBUBSAN_EARLY_SPEC): Likewise.
5741 (SANITIZER_SPEC): Handle libubsan.
5742 (SANITIZER_EARLY_SPEC): Likewise.
5743 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
5744 instead of fsanitize=address.
5745 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
5746 instead of fsanitize=address*.
5747 * builtins.c: Include ubsan.h.
5748 (fold_builtin_0): Instrument __builtin_unreachable.
5749 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
5750 instead of flag_asan.
5751 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
5752 (pointer_sized_int_node): Define.
5753 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
5755 2013-08-30 Mike Stump <mikestump@comcast.net>
5757 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
5760 2013-08-29 Jan Hubicka <jh@suse.cz>
5762 * cgraph.c (cgraph_function_body_availability): Handle weakref
5764 * passes.def: Remove pass_fixup_cfg.
5765 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
5766 track when we need to remove functions.
5767 (gate_ipa_inline): Execute inlining always; add comment why.
5768 (pass_data_ipa_inline): Remove TODO_remove_functions.
5769 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
5770 do not produce summaries.
5771 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
5772 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
5773 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
5774 constant pool and vtable.
5776 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
5778 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
5779 New arm_neon.h's internal macros to specify 64-bit constants.
5780 Avoid using stdint.h's macros.
5782 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
5784 * recog.c (verify_changes): Verify that changes[i].old is non-zero
5785 before applying REG_P.
5787 2013-08-30 Jakub Jelinek <jakub@redhat.com>
5789 PR tree-optimization/58277
5790 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
5791 after seeing too many stmts with vdef in between dombb and current
5792 bb, invalidate everything.
5794 2013-08-30 Richard Biener <rguenther@suse.de>
5796 * fold-const.c (fold_single_bit_test): Fix overflow test.
5798 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
5800 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
5801 and which can live in a register, always retrieve the value on entry.
5802 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
5803 passed by invisible reference specially.
5804 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
5805 (vt_add_function_parameter): Correctly deal with a parameter passed by
5806 invisible reference.
5808 2013-08-30 Jan Hubicka <jh@suse.cz>
5810 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
5812 2013-08-30 Richard Biener <rguenther@suse.de>
5814 PR tree-optimization/58228
5815 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
5816 allow invariant loads in nested loop vectorization.
5818 2013-08-30 Richard Biener <rguenther@suse.de>
5820 PR tree-optimization/58223
5821 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
5822 (has_anti_or_output_dependence): ... this and adjust to also
5823 look for output dependences.
5824 (mark_nodes_having_upstream_mem_writes): Adjust.
5825 (rdg_flag_uses): Likewise.
5827 2013-08-30 Richard Biener <rguenther@suse.de>
5829 PR tree-optimization/58010
5830 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
5831 assert that we have a loop-closed PHI.
5833 2013-08-29 Jan Hubicka <jh@suse.cz>
5835 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
5836 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
5837 * lto-section-in.c (lto_free_function_in_decl_state): New function.
5838 (lto_free_function_in_decl_state_for_node): New function.
5840 2013-08-29 Xinliang David Li <davidxl@google.com>
5842 * loop-unroll.c (report_unroll_peel): Minor message change.
5843 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
5844 Emit alignment peeling message with default -fopt-info.
5845 (vect_loop_versioning): Emit loop version info message.
5846 * tree-vectorizer.c (vectorize_loops): Minor message change.
5847 (execute_vect_slp): Ditto.
5849 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
5851 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
5852 of the clone from the DECL_NAME of the original function.
5854 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
5856 * passes.c (register_pass): Add overload.
5857 * tree-pass.h (register_pass): Forward declare it. Add comment.
5859 2013-08-29 Jan Hubicka <jh@suse.cz>
5861 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
5862 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
5863 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
5864 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
5865 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
5866 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
5867 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
5868 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
5870 2013-08-29 Teresa Johnson <tejohnson@google.com>
5872 * dumpfile.c (dump_loc): Output column number.
5873 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
5874 * doc/invoke.texi: Document optall -fopt-info flag.
5875 * profile.c (read_profile_edge_counts): Use new dump framework.
5876 (compute_branch_probabilities): Ditto.
5877 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
5878 when pass not in any opt group.
5879 * pass_manager.h (pass_manager::get_pass_profile): New method.
5880 * value-prof.c (check_counter): Use new dump framework.
5881 (check_ic_target): Ditto.
5882 * coverage.c (get_coverage_counts): Ditto.
5883 (coverage_init): Setup new dump framework.
5885 2013-08-29 Richard Biener <rguenther@suse.de>
5887 PR tree-optimization/58246
5888 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
5889 handle the dominance check inside a basic-block.
5891 2013-08-29 Richard Biener <rguenther@suse.de>
5894 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
5895 of default defs that appear in abnormal PHI nodes.
5897 2013-08-29 Richard Biener <rguenther@suse.de>
5899 PR tree-optimization/57685
5900 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
5901 single-use operands to avoid exponential complexity.
5903 2013-08-28 Dehao Chen <dehao@google.com>
5905 * ipa-inline.c (edge_badness): Fix integer underflow.
5907 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
5909 * gtm-builtins.def (_ITM_free): Declare leaf.
5911 2013-08-28 Jakub Jelinek <jakub@redhat.com>
5914 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
5915 (*tls_local_dynamic_base_64_largepic): Likewise.
5916 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
5917 Remove predicate from call operand.
5918 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
5919 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
5921 2013-08-28 Jeff Law <law@redhat.com>
5923 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
5924 checks for the number of predecessors and successors allowed.
5925 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
5926 which require copying a joiner block if there is a request which
5927 is a subpath that requires no joiner block copying.
5929 2013-08-28 Jan Hubicka <jh@suse.cz>
5931 * lto-streamer-out.c (DFS_write_tree_body): Drop
5932 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
5933 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
5934 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
5935 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
5936 Do not read DECL_ERROR_ISSUED.
5937 (unpack_ts_decl_with_vis_value_fields): Do not read
5939 (lto_input_ts_binfo_tree_pointers): Do not read
5940 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
5941 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
5942 write DECL_ERROR_ISSUED..
5943 (pack_ts_decl_with_vis_value_fields): Do not write
5945 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
5946 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
5947 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
5948 * tree.h (tree_decl_common): Update comment.
5949 (DECL_ERROR_ISSUED): Remove.
5951 2013-08-28 Jakub Jelinek <jakub@redhat.com>
5954 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
5956 2013-08-28 Jan Hubicka <jh@suse.cz>
5958 * builtins.def (free): Declare leaf.
5960 2013-08-27 David Malcolm <dmalcolm@redhat.com>
5963 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
5964 * configure: Regenerate.
5966 2013-08-27 Martin Jambor <mjambor@suse.cz>
5968 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
5969 (ipa_ancestor_jf_data): Likewise.
5970 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
5971 (ipa_get_jf_pass_through_type_preserved): New function.
5972 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
5973 (ipa_get_jf_ancestor_type_preserved): New function.
5974 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
5975 (ipa_get_jf_ancestor_result): Likewise.
5976 (propagate_vals_accross_pass_through): Use
5977 ipa_get_jf_pass_through_result to do all the value mappings.
5978 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
5979 type_preserved flag.
5980 (ipa_set_jf_cst_copy): New function.
5981 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
5982 (ipa_set_jf_arith_pass_through): Likewise.
5983 (ipa_set_ancestor_jf): Likewise.
5984 (compute_complex_assign_jump_func): Set type_preserved instead of
5986 (ipa_compute_jump_functions_for_edge): Likewise.
5987 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
5988 (update_jump_functions_after_inlining): Update type_preserved.
5989 Explicitely create jump functions when combining one with pass_through.
5990 (ipa_write_jump_function): Stream the type_preserved flags.
5991 (ipa_read_jump_function): Likewise.
5993 2013-08-27 Jakub Jelinek <jakub@redhat.com>
5994 Aldy Hernandez <aldyh@redhat.com>
5996 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
5997 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
5998 * function.h (struct function): Add has_force_vect_loops and
6000 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
6001 * gimple.c (gimple_build_omp_critical): Add KIND argument and
6003 * gimple.def: Update CLAUSES comments.
6004 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
6005 (gimple_build_omp_for): Add argument to prototype.
6006 (gimple_omp_for_kind): New.
6007 (gimple_omp_for_set_kind): New.
6008 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
6009 GOVD_DATA_SHARE_CLASS.
6010 (enum omp_region_type): Add ORT_SIMD.
6011 (gimple_add_tmp_var): Handle ORT_SIMD.
6012 (gimplify_var_or_parm_decl): Same.
6013 (is_gimple_stmt): Same.
6014 (omp_firstprivatize_variable): Same.
6015 (omp_add_variable): Only use splay_tree_insert if lookup failed.
6016 (omp_notice_variable): Handle ORT_SIMD.
6017 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
6018 (omp_check_private): Handle ORT_SIMD.
6019 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
6021 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
6022 Handle OMP_CLAUSE_LASTPRIVATE.
6023 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
6025 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
6026 (gimplify_expr): Handle OMP_SIMD.
6027 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
6028 (expand_GOMP_SIMD_VF): New.
6029 (expand_GOMP_SIMD_LAST_LANE): New.
6030 * internal-fn.def (GOMP_SIMD_LANE): New.
6031 (GOMP_SIMD_VF): New.
6032 (GOMP_SIMD_LAST_LANE): New.
6033 * omp-low.c: Include target.h.
6034 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
6036 (check_omp_nesting_restrictions): Same.
6038 (lower_rec_simd_input_clauses): New.
6039 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
6041 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
6042 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
6043 (expand_omp_build_assign): New.
6044 (expand_omp_for_init_counts): New.
6045 (expand_omp_for_init_vars): New.
6046 (extract_omp_for_update_vars): New.
6047 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
6048 and rewrite accordingly.
6049 (expand_omp_simd): New.
6050 (expand_omp_for): Use expand_omp_simd.
6051 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
6052 (lower_omp_for): Do not lower the body.
6053 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
6055 * tree-flow.h (find_omp_clause): Remove prototype.
6056 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
6057 forcing vectorization of the loop, or if flag_tree_vectorize.
6058 (gate_tree_if_conversion): Similarly.
6059 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
6060 gimple_build_omp_for.
6061 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
6062 * tree-parloops (create_parallel_loop): Pass kind argument to
6063 gimple_build_omp_for.
6064 * tree-pretty-print.c (dump_omp_clause): Add cases for
6065 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
6066 OMP_CLAUSE__SIMDUID_.
6067 (dump_generic_node): Handle OMP_SIMD.
6068 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
6069 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
6070 unroll OMP_SIMD loops here.
6071 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
6072 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
6074 (vect_analyze_data_refs): Handle simd loops.
6075 * tree-vect-loop.c (vectorizable_live_operation): Handle
6077 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
6078 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
6079 (vectorizable_load): Same.
6080 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
6081 (struct simduid_to_vf): New.
6082 (simduid_to_vf::hash): New.
6083 (simduid_to-vf::equal): New.
6084 (struct simd_array_to_simduid): New.
6085 (simd_array_to_simduid::hash): New.
6086 (simd_array_to_simduid::equal): New.
6087 (adjust_simduid_builtins): New.
6088 (struct note_simd_array_uses_struct): New.
6089 (note_simd_array_uses_cb): New.
6090 (note_simd_array_uses): New.
6091 (vectorize_loops): Handle simd hints and adjust simd builtins
6093 * tree-vectorizer.h (struct _stmt_vec_info): Add
6094 simd_lane_access_p field.
6095 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
6096 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
6097 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
6098 (omp_clause_code_name): Same.
6099 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
6100 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
6101 * tree.def (OMP_SIMD): New entry.
6102 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
6103 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
6104 (OMP_CLAUSE_DECL): Adjust range for new clauses.
6105 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
6106 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
6107 (OMP_CLAUSE_LINEAR_STEP): New.
6108 (OMP_CLAUSE_SAFELEN_EXPR): New.
6109 (OMP_CLAUSE__SIMDUID__DECL): New.
6110 (find_omp_clause): New prototype.
6112 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
6114 * config/i386/driver-i386.c (host_detect_local_cpu): Update
6115 Haswell processor detection.
6117 2013-08-27 Christian Widmer <shadow@umbrox.de>
6120 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
6121 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
6122 AVX2 capable processors.
6124 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
6126 * config/aarch64/arm_neon.h: Replace all inline asm implementations
6127 of vget_low_* with implementations in terms of other intrinsics.
6129 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
6132 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
6133 values to -1, 0 and 1.
6135 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
6137 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
6138 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
6139 (<optab><mode>3_insn): Remove.
6140 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
6141 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
6142 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
6143 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
6144 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
6145 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
6146 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
6147 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
6148 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
6149 (ror<mode>3_insn): Likewise.
6150 * config/aarch64/predicates.md (aarch64_simd_register): New.
6152 2013-08-27 Richard Biener <rguenther@suse.de>
6154 PR tree-optimization/57521
6155 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
6156 one edge is non-critical.
6157 (find_phi_replacement_condition): Make sure to use a non-critical
6158 edge. Cleanup and remove old bug workarounds.
6159 (bb_postdominates_preds): Remove.
6160 (if_convertible_loop_p_1): Do not compute post-dominators.
6161 (combine_blocks): Do not free post-dominators.
6162 (main_tree_if_conversion): Likewise.
6163 (pass_data_if_conversion): Add TODO_verify_ssa.
6165 2013-08-27 DJ Delorie <dj@redhat.com>
6167 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
6169 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
6171 * function.c (assign_parm_find_data_types): Set passed_mode and
6172 nominal_mode to the TYPE_MODE of nominal_type for the built
6173 pointer type in case of the struct-pass-by-reference.
6175 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6177 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
6178 (UINT16_TYPE): Change default to "unsigned int".
6180 * config/avr/avr.opt (mfract-convert-truncate): New option.
6181 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
6182 is set, round negative fractional integers according to n1169
6183 when converting to integer types.
6185 2013-08-26 Jan Hubicka <jh@suse.cz>
6187 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
6188 methods can not be called indirectly when their address is not taken.
6190 2013-08-26 Jan Hubicka <jh@suse.cz>
6192 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
6195 2013-08-26 Jan Hubicka <jh@suse.cz>
6197 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
6200 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6202 * reload.c (find_valid_class): Allow classes that do not include
6203 FIRST_PSEUDO_REGISTER - 1.
6205 2013-08-26 Jan Hubicka <jh@suse.cz>
6207 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
6208 fix edge count/frequency when speculation failed; fix type check
6209 for the direct call.
6211 2013-08-26 Jan Hubicka <jh@suse.cz>
6213 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
6215 2013-08-26 Jan Hubicka <jh@suse.cz>
6217 * ipa-inline-transform.c (inline_transform): Be ready for basic block
6218 to be changed by edge redirection.
6220 2013-08-26 Jan Hubicka <jh@suse.cz>
6222 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
6223 formating; add sanity check.
6224 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
6225 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
6226 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
6227 (resolve_noninline_speculation): Update callee keys, too.
6229 2013-08-26 Jan Hubicka <jh@suse.cz>
6231 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
6232 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
6234 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6236 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
6239 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
6241 * config/i386/i386.c (ix86_debug_options): Remove prototype.
6242 (x86_64_elf_select_section): Ditto.
6243 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
6245 (ix86_pass_by_reference): Ditto.
6246 (output_set_got): Ditto.
6247 (ix86_unary_operator_ok): Ditto.
6248 (ix86_expand_builtin): Ditto.
6250 2013-08-23 Jan Hubicka <jh@suse.cz>
6252 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
6254 2013-08-23 Jan Hubicka <jh@suse.cz>
6256 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
6257 (tree_decl_with_vis): Add FINAL field.
6259 2013-08-23 Jeff Law <law@redhat.com>
6261 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
6262 the debugging dump when the expression is fully redundant.
6264 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
6266 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
6267 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
6268 * pretty-print.c (pp_formatted_text_data): Likewise.
6269 (pp_write_text_to_stream): Likewise.
6270 (pp_write_text_as_dot_label_to_stream): Likewise.
6271 (pp_append_r): Likewise.
6272 (pp_format): Likewise.
6273 (pp_flush): Likewise.
6274 (pp_clear_output_area): Likewise.
6275 (pp_append_text): Likewise.
6276 (pp_formatted_text): Likewise.
6277 (pp_remaining_character_count_for_line): Likewise.
6278 (pp_newline): Likewise.
6279 (pp_character): Likewise.
6280 (output_buffer::~output_buffer): Define.
6281 (pretty_printer::~pretty_printer): Destruct output buffer.
6282 * pretty-print.h (output_buffer::~output_buffer): Declare.
6283 (pretty_printer::~pretty_printer): Declare virtual.
6285 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
6288 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
6289 HOST_WIDE_INT_M1U): New macros.
6290 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
6291 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
6292 unsigned -1 for lshift.
6293 * cse.c (cse_insn): Likewise.
6294 * double-int.c (rshift_double, lshift_double): Likewise.
6295 * builtins.c (fold_builtin_bitop): Likewise.
6296 * combine.c (force_to_mode): Likewise.
6297 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
6298 * simplify-rtx.c (simplify_const_unary_operation,
6299 simplify_const_binary_operation): Likewise.
6300 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
6301 check_va_list_escapes): Likewise.
6302 * rtlanal.c (nonzero_bits1): Likewise.
6303 * expmed.c (expand_smod_pow2): Likewise.
6304 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
6306 2013-08-23 Jan Hubicka <jh@suse.cz>
6308 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
6309 as having address taken.
6311 2013-08-23 Jan Hubicka <jh@suse.cz>
6313 * ipa-utils.h (method_class_type): Declare.
6314 * ipa-devirt.c (method_class_type): Export.
6316 * cgraphunit.c (analyze_functions): Do basic devirtualization;
6317 do not walk base classes of anonymous types.
6319 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
6321 PR rtl-optimization/58220
6323 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
6324 handle SEQUENCE insns properly.
6326 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
6328 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
6330 (pp_newline_and_indent): Likewise.
6331 (pp_separate_with): Likewise.
6332 * pretty-print.c (pp_newline_and_flush): Define.
6333 (pp_newline_and_indent): Likewise.
6334 (pp_separate_with): Likewise.
6336 2013-08-23 Jakub Jelinek <jakub@redhat.com>
6339 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
6340 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
6342 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
6344 * config/i386/predicates.md (ext_sse_reg_operand): New.
6345 * config/i386/i386.md (*movti_internal): Use
6346 predicate to determine if EVEX is needed.
6347 (*movsi_internal): Ditto.
6348 (*movdf_internal): Ditto.
6349 (*movsf_internal): Ditto.
6350 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
6352 2013-08-23 Jakub Jelinek <jakub@redhat.com>
6354 PR tree-optimization/58209
6355 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
6356 (find_tail_calls): Give up for pointer result types if m is non-NULL.
6357 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
6358 emit POINTER_PLUS_EXPR.
6359 (create_tailcall_accumulator): For pointer result type accumulate in
6362 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
6364 * configure.ac: Add backslashes missing from the last change.
6365 * configure: Regenerate.
6367 2013-08-22 Jan Hubicka <jh@susue.cz>
6369 * ipa.c (function_and_variable_visibility): First remember function
6370 was global and then make it local.
6372 2013-08-22 Julian Brown <julian@codesourcery.com>
6374 * configure.ac: Add aarch64 to list of arches which use "nop" in
6376 * configure: Regenerate.
6378 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6380 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
6381 gnu_libc_has_function.
6382 * config/s390/tpf.h: Likewise.
6384 2013-08-22 Jan Hubicka <jh@susue.cz>
6386 * timevar.c (validate_phases): Add cast.
6388 2013-08-22 Jan Hubicka <jh@susue.cz>
6390 * timevar.c (validate_phases): Use size_t for memory.
6391 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
6393 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
6395 * pretty-print.h (output_buffer::output_buffer): Declare.
6396 (pretty_printer::pretty_printer): Likewise.
6397 (pp_construct): Remove.
6398 * pretty-print.c (output_buffer::output_buffer): Define.
6399 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
6400 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
6401 (print_gimple_expr): Likewise.
6402 (print_gimple_seq): Likewise.
6403 (gimple_dump_bb): Likewise.
6404 * sched-vis.c (dump_value_slim): Likewise.
6405 (dump_insn_slim): Likewise.
6406 (dump_rtl_slim): Likewise.
6407 (str_pattern_slim): Likewise.
6408 * tree-mudflap.c (mf_varname_tree): Likewise.
6409 * graph.c (print_graph_cfg): Likewise.
6410 (start_graph_dump): Likewise.
6411 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
6413 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
6414 pretty printer initialization.
6415 * coretypes.h (diagnostic_context): Remove superflous type alias
6417 (pretty_printer): Likewise. Declare directly as a class.
6418 (pretty_print_info): Remove declaration as class.
6419 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
6420 and pp_clear_output_area.
6421 (asan_add_global): Likewise.
6423 2013-08-22 Jan Hubicka <jh@suse.cz>
6425 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
6426 * ipa-utils.h (update_type_inheritance_graph): Declare.
6427 (possible_polymorphic_call_target_p): Declare.
6428 (possible_polymorphic_call_target_p): New.
6429 * ipa-devirt.c: Update toplevel comments.
6430 (cached_polymorphic_call_targets): Move up.
6431 (odr_type_d): Move ID down.
6432 (polymorphic_type_binfo_p): Update comment.
6433 (odr_hasher::remove): Likewise;
6434 (get_odr_type): Set anonymous_namespace.
6435 (dump_odr_type): Dump it.
6436 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
6437 (maybe_record_node): Record node in cached_polymorphic_call_targets.
6438 (record_binfo): Add comment.
6439 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
6441 (devirt_node_removal_hook): Do not iCE when cache is freed.
6442 (possible_polymorphic_call_target_p): New predicate.
6443 (update_type_inheritance_graph): New function.
6445 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
6446 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6447 Sergey Lega <sergey.s.lega@intel.com>
6448 Anna Tikhonova <anna.tikhonova@intel.com>
6449 Ilya Tocar <ilya.tocar@intel.com>
6450 Andrey Turetskiy <andrey.turetskiy@intel.com>
6451 Ilya Verbin <ilya.verbin@intel.com>
6452 Kirill Yukhin <kirill.yukhin@intel.com>
6453 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6455 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
6456 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
6457 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
6458 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
6459 (OPTION_MASK_ISA_AVX2_UNSET): Update.
6460 (OPTION_MASK_ISA_AVX512F_UNSET): New.
6461 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
6462 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
6463 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
6464 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
6465 OPT_mavx512pf, OPT_mavx512er cases.
6466 * config/i386/constraints.md (v): New constraint.
6467 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
6468 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
6469 (bit_AVX512CD): New.
6470 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6471 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
6472 * config/i386/i386-c.c (ix86_target_macros_internal):
6473 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
6475 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
6476 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
6477 * config/i386/i386.c (regclass_map, dbx_register_map)
6478 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
6479 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
6480 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
6482 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
6483 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
6484 -mavx512pf options. Fix formatting.
6485 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
6486 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
6487 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
6488 -mavx512cd, -mavx512pf options.
6489 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
6490 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
6491 (ix86_preferred_output_reload_class): Replace SSE_REGS with
6493 (ix86_hard_regno_mode_ok): Support 512-bit registers.
6494 (ix86_set_reg_reg_cost): Ditto.
6495 (x86_order_regs_for_local_alloc): Ditto.
6496 (MAX_VECT_LEN): Extend to 64-byte.
6497 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
6498 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
6499 (TARGET_AVX512ER, TARGET_AVX512CD): New.
6500 (BIGGEST_ALIGNMENT): Extend to 512-bits.
6501 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
6502 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
6503 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
6504 (SSE_REG_MODE_P): Support new modes.
6505 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
6506 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
6507 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
6508 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
6509 (REG_CLASS_CONTENTS): Add new registers.
6510 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
6511 (EXT_REX_SSE_REGNO_P): New.
6512 (HI_REGISTER_NAMES): Add new registers.
6513 * config/i386/i386.md: Define constants for new registers.
6514 (mode): Add new 512-bit modes.
6515 (prefix): Support evex prefix.
6516 (isa): Support avx512f, noavx512f, fma_avx512f.
6517 (ssemodesuffix): Add new 512-bit modes.
6519 (*movxi_internal_avx512f): Ditto.
6520 (*movdi_internal): Replace constraint "x" with the new constraint "v".
6522 (*movsi_internal): Likewise.
6523 (*movdf_internal): Likewise.
6524 (*movsf_internal): Likewise.
6525 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
6526 (<code><mode>3): Likewise.
6527 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
6529 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
6530 with the new constraint "v".
6531 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
6533 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
6534 with the new constraint "v".
6535 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
6536 (<sse2>_storedqu<avxsizesuffix>): Likewise.
6537 (*<plusminus_insn><mode>3): Likewise.
6538 (<sse>_vm<plusminus_insn><mode>3): Likewise.
6539 (*mul<mode>3): Likewise.
6540 (<sse>_vmmul<mode>3): Likewise.
6541 (<sse>_div<mode>3): Likewise.
6542 (<sse>_vmdiv<mode>3): Likewise.
6543 (<sse>_sqrt<mode>2): Likewise.
6544 (<sse>_vmsqrt<mode>2): Likewise.
6545 (*<code><mode>3_finite): Likewise.
6546 (*<code><mode>3) <smaxmin>: Likewise.
6547 (<sse>_vm<code><mode>3): Likewise.
6548 (*<code><mode>3) <any_logic>: Likewise.
6549 (*fma_fmadd_<mode>): Likewise.
6550 (*fma_fmsub_<mode>): Likewise.
6551 (*fma_fnmadd_<mode>): Likewise.
6552 (*fma_fnmsub_<mode>): Likewise.
6553 (*fma_fmaddsub_<mode>): Likewise.
6554 (*fma_fmsubadd_<mode>): Likewise.
6555 (*fmai_fmadd_<mode>): Likewise.
6556 (*fmai_fmsub_<mode>): Likewise.
6557 (*fmai_fnmadd_<mode>): Likewise.
6558 (*fmai_fnmsub_<mode>): Likewise.
6559 (sse_cvtsi2ss): Likewise.
6560 (sse_cvtsi2ssq): Likewise.
6561 (sse_cvtss2si): Likewise.
6562 (sse_cvtss2si_2): Likewise.
6563 (sse_cvtss2siq): Likewise.
6564 (sse_cvtss2siq_2): Likewise.
6565 (sse_cvttss2si): Likewise.
6566 (sse_cvtss2siq_2): Likewise.
6567 (float<sseintvecmodelower><mode>2): Likewise.
6568 (sse2_cvtsd2si_2): Likewise.
6569 (sse2_cvtsd2siq_2): Likewise.
6570 (*<plusminus_insn><mode>3): Likewise.
6571 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
6572 (*<sse4_1_avx2>_mul<mode>3): Likewise.
6573 (ashr<mode>3): Likewise.
6574 (<shift_insn><mode>3): Likewise.
6575 (avx2_<code><mode>3): Likewise.
6576 (*avx2_<code><mode>3): Likewise.
6577 (*andnot<mode>3): Likewise.
6578 (*<code><mode>3) <any_logic>: Likewise.
6579 (abs<mode>2): Likewise.
6580 (avx2_permvar<mode>): Likewise.
6581 (avx2_perm<mode>_1): Likewise.
6582 (*avx_vpermilp<mode>): Likewise.
6583 (avx_vpermilvar<mode>3): Likewise.
6584 (avx2_ashrv<mode>): Likewise.
6585 (avx2_<shift_insn>v<mode>): Likewise.
6586 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
6588 * doc/rtl.texi: Document XImode.
6590 2013-08-21 Jeff Law <law@redhat.com>
6592 * tree-flow.h (register_jump_thread): Pass vector of edges
6593 instead of each important edge.
6594 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
6595 thread path into a vector and pass that to register_jump_thread.
6596 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
6597 passed in edge vector to the current 3-edge form.
6600 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
6602 * dce.c (fini_dce): Call df_analyze again just in case
6603 delete_unmarked_insns removed anything.
6605 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
6607 * reload.h (struct reg_equivs): Rename to ..
6608 (struct reg_equivs_s): .. this.
6610 2013-08-20 Martin Liska <marxin.liska@gmail.com>
6612 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
6614 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6616 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
6618 2013-08-21 Jeff Law <law@redhat.com>
6620 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
6621 simplify assignments too. If the RHS collapses to a singleton
6622 range, then return the value for the range.
6624 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
6626 * config/i386/sse.md (V16): Rename to...
6628 (mov<mode>): Update iterator name.
6629 (*mov<mode>_internal): Ditto.
6630 (push<mode>1): Ditto.
6631 (movmisalign<mode>): Ditto.
6633 2013-08-20 Jan Hubicka <jh@suse.cz>
6636 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
6637 entry for direct edges.
6638 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
6640 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6642 Revert my last two changes, r201865 and r201864:
6645 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6647 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
6648 instances can own GC refs.
6650 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
6651 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
6652 (gcc::context::gt_pch_nx): Likewise.
6653 (gcc::context::gt_pch_nx): Likewise.
6654 * ggc.h (gt_ggc_mx <T>): New.
6655 (gt_pch_nx_with_op <T>): New.
6656 (gt_pch_nx <T>): New.
6657 * passes.c (opt_pass::gt_ggc_mx): New.
6658 (opt_pass::gt_pch_nx): New.
6659 (opt_pass::gt_pch_nx_with_op): New.
6660 (pass_manager::gt_ggc_mx): New.
6661 (pass_manager::gt_pch_nx): New.
6662 (pass_manager::gt_pch_nx_with_op): New.
6663 (pass_manager::operator new): Use
6664 ggc_internal_cleared_alloc_stat rather than xcalloc.
6665 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
6666 (pass_manager::gt_ggc_mx): New.
6667 (pass_manager::gt_pch_nx): New.
6668 (pass_manager::gt_pch_nx_with_op): New.
6669 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
6670 (opt_pass::operator new): New.
6671 (opt_pass::gt_ggc_mx): New.
6672 (opt_pass::gt_pch_nx): New.
6673 (opt_pass::gt_pch_nx_with_op): New.
6676 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6678 * Makefile.in (GTFILES): Add context.h.
6679 * context.c (gcc::context::operator new): New.
6680 (gcc::context::gt_ggc_mx): New.
6681 (gcc::context::gt_pch_nx): New.
6682 (gcc::context::gt_pch_nx): New.
6683 * context.h (gcc::context): Add GTY((user)) marking.
6684 (gcc::context::operator new): New.
6685 (gcc::context::gt_ggc_mx): New.
6686 (gcc::context::gt_pch_nx): New.
6687 (gcc::context::gt_pch_nx): New.
6688 (g): Add GTY marking.
6689 (gt_ggc_mx (gcc::context *)): New.
6690 (gt_pch_nx (gcc::context *)): New.
6691 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
6692 void *cookie)): New.
6693 * gengtype.c (open_base_files) <ifiles>: Add context.h.
6695 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
6697 * dce.c (fini_dce): Call df_analyze again just in case
6698 delete_unmarked_insns removed anything.
6700 2013-08-20 Teresa Johnson <tejohnson@google.com>
6702 PR rtl-optimizations/57451
6703 * final.c (reemit_insn_block_notes): Prevent lexical blocks
6704 from crossing split section boundaries.
6706 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
6708 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
6709 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
6710 with MULTLIB_DEFAULTS.
6712 2013-08-20 Nick Clifton <nickc@redhat.com>
6714 * target.def (narrow_volatile_bitfield): Note that the default
6715 value is false, not !TARGET_STRICT_ALIGN.
6716 * doc/tm.texi: Regenerate.
6718 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
6720 Fix LIB_SPEC for systems without libpthread.
6722 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
6723 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
6725 * config/i386/linux-common.h: Likewise.
6726 * config/mips/linux-common.h: Likewise.
6728 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6730 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
6733 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6735 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
6736 instances can own GC refs.
6738 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
6739 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
6740 (gcc::context::gt_pch_nx): Likewise.
6741 (gcc::context::gt_pch_nx): Likewise.
6742 * ggc.h (gt_ggc_mx <T>): New.
6743 (gt_pch_nx_with_op <T>): New.
6744 (gt_pch_nx <T>): New.
6745 * passes.c (opt_pass::gt_ggc_mx): New.
6746 (opt_pass::gt_pch_nx): New.
6747 (opt_pass::gt_pch_nx_with_op): New.
6748 (pass_manager::gt_ggc_mx): New.
6749 (pass_manager::gt_pch_nx): New.
6750 (pass_manager::gt_pch_nx_with_op): New.
6751 (pass_manager::operator new): Use
6752 ggc_internal_cleared_alloc_stat rather than xcalloc.
6753 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
6754 (pass_manager::gt_ggc_mx): New.
6755 (pass_manager::gt_pch_nx): New.
6756 (pass_manager::gt_pch_nx_with_op): New.
6757 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
6758 (opt_pass::operator new): New.
6759 (opt_pass::gt_ggc_mx): New.
6760 (opt_pass::gt_pch_nx): New.
6761 (opt_pass::gt_pch_nx_with_op): New.
6763 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6765 * Makefile.in (GTFILES): Add context.h.
6766 * context.c (gcc::context::operator new): New.
6767 (gcc::context::gt_ggc_mx): New.
6768 (gcc::context::gt_pch_nx): New.
6769 (gcc::context::gt_pch_nx): New.
6770 * context.h (gcc::context): Add GTY((user)) marking.
6771 (gcc::context::operator new): New.
6772 (gcc::context::gt_ggc_mx): New.
6773 (gcc::context::gt_pch_nx): New.
6774 (gcc::context::gt_pch_nx): New.
6775 (g): Add GTY marking.
6776 (gt_ggc_mx (gcc::context *)): New.
6777 (gt_pch_nx (gcc::context *)): New.
6778 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
6779 void *cookie)): New.
6780 * gengtype.c (open_base_files) <ifiles>: Add context.h.
6782 2013-08-20 Alan Modra <amodra@gmail.com>
6785 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
6786 (rs6000_emit_epilogue): Likewise.
6788 2013-08-19 Dehao Chen <dehao@google.com>
6790 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
6792 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
6793 Jakub Jelinek <jakub@redhat.com>
6795 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
6796 (BUILT_IN_FABSD64): Likewise.
6797 (BUILT_IN_FABSD128): Likewise.
6798 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
6799 (fold_builtin_1): Likewise.
6800 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
6801 destination and source operands.
6802 (*abstd2_fpr): Likewise.
6803 (*nabstd2_fpr): Likewise.
6805 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
6807 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
6810 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
6812 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
6815 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
6817 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
6818 * builtins.c (default_libc_has_function): New.
6819 (gnu_libc_has_function): Ditto.
6820 (no_c99_libc_has_function): Ditto.
6821 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
6822 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
6823 (fold_builtin_sincos): Likewise.
6824 (fold_builtin_cexp): Likewise.
6825 * builtins.def (DEF_C94_BUILTIN): Likewise.
6826 (DEF_C99_BUILTIN): Likewise.
6827 (DEF_C99_C90RES_BUILTIN): Likewise.
6828 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
6829 definitions to using this define.
6830 * config/darwin-protos.h (darwin_libc_has_function): New.
6831 * config/darwin.c (darwin_libc_has_function): Ditto.
6832 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
6833 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
6834 * config/darwin.h: Ditto.
6835 * config/elfos.h: Ditto.
6836 * config/freebsd.h: Ditto.
6837 * config/i386/cygming.h: Ditto.
6838 * config/i386/djgpp.h: Ditto.
6839 * config/i386/i386-interix.h: Ditto.
6840 * config/microblaze/microblaze.h: Ditto.
6841 * config/mmix/mmix.h: Ditto.
6842 * config/gnu-user.h: Ditto.
6843 * config/ia64/hpux.h: Ditto.
6844 * config/pa/pa-hpux.h: Ditto.
6845 * config/pdp11/pdp11.h: Ditto.
6846 * config/picochip/picochip.h: Ditto.
6847 * config/linux.h: Ditto.
6848 * config/netbsd.h: Ditto.
6849 * config/openbsd.h: Ditto.
6850 * config/rs6000/aix43.h: Ditto.
6851 * config/rs6000/aix51.h: Ditto.
6852 * config/rs6000/aix52.h: Ditto.
6853 * config/rs6000/aix53.h: Ditto.
6854 * config/rs6000/aix61.h: Ditto.
6855 * config/rs6000/darwin.h: Ditto.
6856 * config/rs6000/linux.h: Ditto.
6857 * config/rs6000/linux64.h: Ditto.
6858 * config/s390/tpf.h: Ditto.
6859 * config/sol2-10.h: Ditto.
6860 * config/sol2.h: Ditto.
6861 * config/vms/vms.h: Ditto.
6862 * config/vxworks.h: Ditto.
6863 * config/linux-android.c (linux_android_libc_has_function):
6864 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
6865 * config/linux-protos.h (linux_android_libc_has_function):
6867 * config/i386/i386.c (ix86_libc_has_function): New.
6868 * config/i386/i386-protos.h
6869 (ix86_libc_has_function): New declaration.
6870 * config/i386/i386.md
6871 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
6872 ("isinf<mode>2): Likewise.
6873 * convert.c (convert_to_integer): Using new target hook
6874 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
6875 TARGET_C99_FUNCTIONS.
6876 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
6877 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
6878 * coretypes.h (function_class): New enum for different
6879 classes of functions.
6880 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
6881 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
6882 (TARGET_HAS_SINCOS): Likewise.
6883 (TARGET_LIBC_HAS_FUNCTION): New.
6884 * doc/tm.texi: Regenerated.
6885 * targhooks.h (default_libc_has_function): New declaration.
6886 (no_c99_libc_has_function): Ditto.
6887 (gnu_libc_has_function): Ditto.
6888 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
6889 and TARGET_HAS_SINCOS.
6891 2013-08-18 Jan Hubicka <jh@suse.cz>
6893 * Makeifle-in (ipa-devirt.o): New.
6894 (GTFILES): Add ipa-utils.h and ipa-devirt.c
6895 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
6896 (analyze_functions): Look into possible targets of polymorphic call.
6897 * dumpfile.c (dump_files): Add type-inheritance dump.
6898 * dumpfile.h (TDI_inheritance): New.
6899 * ipa-devirt.c: New file.
6900 * ipa-utils.h (odr_type_d): Forward declare.
6901 (odr_type): New type.
6902 (build_type_inheritance_graph): Declare.
6903 (possible_polymorphic_call_targets): Declare and introduce inline
6904 variant when only edge is pased.
6905 (dump_possible_polymorphic_call_targets): Likewise.
6906 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
6907 * tree.c (type_in_anonymous_namespace_p): Break out from ...
6908 (types_same_for_odr): ... here.
6909 * tree.h (type_in_anonymous_namespace_p): Declare.
6911 2013-08-18 Jakub Jelinek <jakub@redhat.com>
6913 PR tree-optimization/58006
6914 * tree-parloops.c (take_address_of): Don't ICE if get_name
6916 (eliminate_local_variables_stmt): Remove clobber stmts.
6918 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
6920 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
6921 error message is issued for an alias to undefined symbol.
6923 2013-08-18 Jan Hubicka <jh@suse.cz>
6925 * cgraph.c (cgraph_create_indirect_edge): Discover
6926 polymorphic calls and record basic info into indirect_info.
6927 * gimple-fold.c (gimple_fold_call): When doing BINFO based
6928 devirtualization, ignore objc function calls.
6929 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
6930 call with no parm index info.
6931 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
6932 * tree.c (virtual_method_call_p): New function.
6933 * tree.h (virtual_method_call_p): Declare.
6935 2013-08-16 Jan Hubicka <jh@suse.cz>
6938 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
6940 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
6942 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
6945 2013-08-16 David Malcolm <dmalcolm@redhat.com>
6947 * gengtype.c (type_for_name): Add special-case support for
6948 locating types within the "gcc::" namespace.
6949 (open_base_files): Emit a "using namespace gcc" directive.
6951 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
6954 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
6955 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
6957 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
6958 array instead of each individual operand as a separate argument.
6959 (emit_fusion_gpr_load): Likewise.
6960 (expand_fusion_gpr_load): Add new function declaration.
6962 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
6963 signature to have the operands passed as an array, instead of as
6964 separate arguments. Allow ZERO_EXTEND to be in the memory
6965 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
6966 depend on the register live/dead flags when peepholes are run.
6967 (expand_fusion_gpr_load): New function to be called from the
6968 peephole2 pass, to change the register that addis sets to be the
6970 (emit_fusion_gpr_load): Change the calling signature to have the
6971 operands passed as an array, instead of as separate arguments.
6972 Allow ZERO_EXTEND to be in the memory address, and also
6973 SIGN_EXTEND if -mpower8-fusion-sign.
6975 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
6977 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
6978 adjust the register addis loads up in the peephole2 pass. Do not
6979 depend on the register live/dead state when the peephole pass is done.
6981 2013-08-16 David Malcolm <dmalcolm@redhat.com>
6983 * gengtype.c (create_user_defined_type): Ensure that the kind
6984 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
6985 declaration is seen before the GTY((user)) marking.
6987 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
6990 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
6992 2013-08-16 Jan Hubicka <jh@suse.cz>
6994 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
6995 arugment expected_type.
6996 (gimple_fold_call): Use it.
6997 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
6998 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
6999 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
7000 (try_make_edge_direct_virtual_call): Likewise.
7001 * tree.c (obj_type_ref_class): New.
7002 * tree.h (obj_type_ref_class): Use it.
7004 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
7006 * sched-vis.c (rtl_slim_pp_initialized): Remove.
7007 (rtl_slim_pp): Likewise.
7008 (init_rtl_slim_pretty_print): Likewise.
7009 (dump_value_slim): Don't call it. Use local pretty printer.
7010 (dump_insn_slim): Likewise.
7011 (dump_rtl_slim): Likewise.
7012 (str_pattern_slim): Likewise.
7013 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
7016 2013-08-16 Jakub Jelinek <jakub@redhat.com>
7018 PR tree-optimization/58164
7019 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
7020 walk gimple_goto_dest of GIMPLE_GOTO.
7022 PR tree-optimization/58165
7023 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
7024 bi_call must be the last stmt in a bb, don't split_block, instead
7025 use fallthru edge from it and give up if there is none.
7026 Release conds vector when returning early.
7028 2013-08-14 Xinliang David Li <davidxl@google.com>
7030 * config/i386/i386.c (ix86_option_override_internal):
7031 Remove unused variable and field.
7033 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7036 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
7037 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
7038 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
7039 and Linux, correct BLKmode alignment when 128-bit alignment is
7040 required and compatibility flag is not set.
7041 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
7042 for zero-size arguments when compatibility flag is not set.
7044 2013-08-14 Jakub Jelinek <jakub@redhat.com>
7046 PR tree-optimization/58145
7047 * tree-sra.c (build_ref_for_offset): If prev_base has
7048 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
7050 2013-08-14 Xinliang David Li <davidxl@google.com>
7052 * config/i386/i386.c (ix86_option_override_internal):
7053 Fix uninitialized variable error.
7055 2013-08-14 Xinliang David Li <davidxl@google.com>
7057 * config/i386/i386.opt: Define two new options.
7058 * config/i386/x86-tune.def: Add arch selector field in macros.
7059 * config/i386/i386.h: Adjust macro definition.
7060 * config/i386/i386.c (ix86_option_override_internal):
7062 (parse_mtune_ctrl_str): New function.
7063 (set_ix86_tune_features): New function.
7064 (ix86_function_specific_restore): Call the new helper function.
7066 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
7068 PR rtl-optimization/57662
7069 * sel-sched.c (code_motion_process_successors): When the current insn
7070 is removed after the recursive traversal, break from the loop.
7071 Add comments and debug printouts.
7073 2013-08-14 Jakub Jelinek <jakub@redhat.com>
7074 Alexandre Oliva <aoliva@redhat.com>
7077 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
7078 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
7079 there also UNSPEC_PLTOFF.
7081 2013-08-14 Marek Polacek <polacek@redhat.com>
7083 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
7084 NUM_CONDITIONS bit positions.
7086 2013-08-13 Cary Coutant <ccoutant@google.com>
7088 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
7089 (attr_checksum): Hash vector contents instead of pointer.
7090 (attr_checksum_ordered): Likewise.
7092 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
7094 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
7095 when Pmode != word_mode. Add length_address attribute.
7096 (sse3_monitor_<mode>): Merge from sse3_monitor and
7097 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
7098 Pmode != word_mode. Update insn length attribute.
7099 * config/i386/i386.c (ix86_option_override_internal): Update
7100 ix86_gen_monitor selection for merged sse3_monitor insn.
7102 2013-08-13 Julian Brown <julian@codesourcery.com>
7104 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
7105 perform invalid legitimization on greater-than-word-size modes for
7108 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
7110 * ira.c (setup_class_translate_array): Use aclass instead of cl
7111 for classes not fully covered by allocno classes.
7113 2013-08-13 Jakub Jelinek <jakub@redhat.com>
7115 PR tree-optimization/57661
7116 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
7117 * tree-inline.c (tree_function_versioning): Initialize it.
7118 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
7119 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
7120 that is not being copied.
7123 * asan.c (instrument_strlen_call): Fix typo in comment.
7124 Use char * type even for the lhs of POINTER_PLUS_EXPR.
7126 2013-08-13 Steve Ellcey <sellcey@mips.com>
7128 * config/mips/mips.md (prefetch): Use lw instead of ld on
7129 loongson in 32bit mode.
7131 2013-08-13 Nick Clifton <nickc@redhat.com>
7133 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
7135 2013-08-13 Jan Hubicka <jh@suse.cz>
7137 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
7138 introduced edge; fix typo in sanity check.
7139 (cgraph_resolve_speculation): Export; improve diagnostic.
7140 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
7141 speculation at type mismatch.
7142 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
7143 (cgraph_resolve_speculation): Declare.
7144 (symtab_can_be_discarded): New function.
7145 * value-prof.c (gimple_ic_transform): Remove actual transform code.
7146 * ipa-inline-transform.c (speculation_removed): New global var.
7147 (clone_inlined_nodes): See if speculation can be removed.
7148 (inline_call): If speculations was removed, we growths may not match.
7149 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
7150 (speculation_useful_p): New function.
7151 (resolve_noninline_speculation): New function.
7152 (inline_small_functions): Resolve useless speculations.
7153 * ipa-inline.h (speculation_useful_p): Declare
7154 * ipa.c (can_replace_by_local_alias): Simplify.
7155 (ipa_profile): Produce speculative calls in non-lto, too;
7156 add simple cost model; produce local aliases.
7158 2013-08-13 David Malcolm <dmalcolm@redhat.com>
7160 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
7163 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
7165 * config/i386/i386.c (ix86_function_versions): Use error + inform.
7167 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
7169 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
7170 iterator instead of X87MODEF.
7172 2013-08-12 Perez Read <netfirewall@gmail.com>
7175 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
7176 operand 0 for intel asm alternative.
7177 (*movabs<mode>_2): Ditto for operand 1.
7179 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
7181 * config/aarch64/arm_none.h
7182 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
7184 2013-08-12 Nick Clifton <nickc@redhat.com>
7186 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
7188 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
7190 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
7191 expand for QI/HImode operand to produce more effictive code for
7192 unsigned char(short) --> float(double) conversion.
7194 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
7196 * doc/invoke.texi: Mention that -ftls-model does not force the final
7199 2013-08-12 Marek Polacek <polacek@redhat.com>
7200 Marc Glisse <marc.glisse@inria.fr>
7202 PR tree-optimization/57980
7203 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
7204 when creating -1 constant.
7206 2013-08-10 Jan Hubicka <jh@suse.cz>
7208 Workaround binutils PR14342.
7209 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
7210 (gimple_init_edge_profiler): Likewise.
7211 (gimple_gen_ic_func_profiler): Likewise.
7213 2013-08-09 Jan Hubicka <jh@suse.cz>
7215 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
7217 2013-08-09 Xinliang David Li <davidxl@google.com>
7219 * config/i386/stringop.def: New file.
7220 * config/i386/stringop.opt: New file.
7221 * config/i386/i386-opts.h: Include stringopt.def.
7222 * config/i386/i386.opt: Include stringopt.opt.
7223 * config/i386/i386.c (ix86_option_override_internal):
7224 Override default size based stringop inline strategies with options.
7225 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
7228 2013-08-09 Jan Hubicka <jh@suse.cz>
7230 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
7232 2013-08-09 Jan Hubicka <jh@suse.cz>
7234 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
7236 (dump_cgraph_node): Dump profile-id.
7237 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
7238 and common_target_probability.
7239 * lto-cgraph.c (lto_output_edge): Stream common targets.
7240 (lto_output_node): Stream profile ids.
7241 (input_node): Stream profile ids.
7242 (input_edge): Stream common targets.
7243 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
7244 * ipa.c: Include value-prof.h
7245 (ipa_profile_generate_summary): Turn indirect call statement histograms
7246 into common targets.
7247 (ipa_profile): Turn common targets into speculative edges.
7249 2013-08-09 Jan Hubicka <jh@suse.cz>
7251 * cgraph.h (cgraph_node): Add profile_id.
7252 * value-prof.c (cgraph_node_map): Turn into pointer_map.
7253 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
7254 (del_node_map): Update.
7255 (find_func_by_funcdef_no): Replace by ...
7256 (find_func_by_profile_id): ... this one.
7257 (gimple_ic_transform): Do not remove useful histograms when
7258 speculation is not done; dump info when indirect call removal
7260 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
7261 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
7262 (__gcov_indirect_call_profiler_v2): .. this one.
7263 * profile.h (init_node_map): Update.
7264 * coverage.c (coverage_compute_profile_id): New function.
7265 * coverage.h (coverage_compute_profile_id): Declare.
7266 * tree-profile.c (init_ic_make_global_vars): Make
7267 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
7268 (gimple_init_edge_profiler): Update prototype of
7269 __gcov_indirect_call_profiler.
7270 (gimple_gen_ic_func_profiler): Simplify.
7271 (tree_profiling): Use init_node_map
7273 2013-08-09 Jan Hubicka <jh@suse.cz>
7275 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
7276 non-speculative refs.
7277 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
7278 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
7279 (cgraph_set_call_stmt): Likewise.
7280 (cgraph_create_edge_1): Fix release checking compilatoin;
7282 (cgraph_free_edge): Free indirect info.
7283 (cgraph_turn_edge_to_speculative): New function.
7284 (cgraph_speculative_call_info): New function.
7285 (cgraph_make_edge_direct): Return direct edge; handle speculation.
7286 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
7287 (dump_cgraph_node): Dump speculation.
7288 (verify_edge_count_and_frequency): Accept speculative edges.
7289 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
7290 (verify_cgraph_node): Handle speculation.
7291 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
7292 (cgraph_set_call_stmt): Update prototype.
7293 (cgraph_make_edge_direct): Update prototype.
7294 (cgraph_speculative_call_info): Declare.
7295 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
7296 to change; update call of ipa_find_references.
7297 * ipa-ref.c (ipa_record_reference): Fix return value; clear
7298 lto_stmt_uid and speculative flags.
7299 (ipa_dump_references): Dump speculation.
7300 (ipa_clone_references): Clone speculative flag.
7301 (ipa_clone_referring): Likewise.
7302 (ipa_clone_ref): New function.
7303 (ipa_find_reference): Look into lto_stmt_uids
7304 (ipa_clear_stmts_in_references): Do not clear speculative calls.
7305 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
7306 (ipa_find_reference): Update declaration.
7307 (ipa_clone_ref): Declare.
7308 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
7309 stream speculative flag.
7310 (lto_output_ref): Stream statements uids and speculation.
7311 (input_ref): Likewise.
7312 (input_edge): Stream speuclation.
7313 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
7314 (cgraph_set_call_stmt_including_clones): Handle speculation.
7315 * ipa-inline.c (heap_edge_removal_hook): New function.
7316 (inline_small_functions): Register it.
7317 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
7318 also initialize refs.
7319 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
7321 (try_make_edge_direct_simple_call): Likewise.
7322 (try_make_edge_direct_simple_call): Likewise.
7323 (update_indirect_edges_after_inlining): Likewise.
7324 (remove_described_reference): Look proper lto_stmt_uid.
7325 (propagate_controlled_uses): Likewise.
7326 (propagate_controlled_uses): Liekwise.
7327 * tree-inline.c (copy_bb): Copy speculative edges.
7328 (redirect_all_calls): New function.
7329 (copy_cfg_body): Do redirection after loop info is updated.
7330 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
7332 2013-08-09 Jan Hubicka <jh@suse.cz>
7334 * lto-streamer-out.c (output_function): Renumber PHIs.
7335 * lto-streamer-in.c (input_function): Likewise.
7337 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
7339 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
7340 (get_lane_unsigned): Likewise.
7341 (dup_lane_scalar): Likewise.
7342 (get_lane): enable for VALL.
7343 * config/aarch64/aarch64-simd.md
7344 (aarch64_dup_lane_scalar<mode>): Remove.
7345 (aarch64_get_lane_signed<mode>): Likewise.
7346 (aarch64_get_lane_unsigned<mode>): Likewise.
7347 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
7348 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
7349 (aarch64_get_lane<mode>): Enable for all vector modes.
7350 (aarch64_get_lanedi): Remove misleading constraints.
7351 * config/aarch64/arm_neon.h
7352 (__aarch64_vget_lane_any): Define.
7353 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
7354 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
7355 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
7356 * config/aarch64/iterators.md (VDQQH): New.
7360 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
7362 * configure.ac: Add GAS check for LEON instructions on SPARC.
7363 * configure: Regenerate.
7364 * config.in: Likewise.
7365 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
7367 * config/sparc/sparc.opt (LEON, LEON3): New masks.
7368 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
7370 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
7371 (AS_LEON_FLAG): New macro.
7372 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
7373 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
7374 Deal with LEON and LEON3 for the memory model.
7375 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
7376 (atomic_compare_and_swap<mode>_1): Likewise.
7377 (*atomic_compare_and_swap<mode>_1): Likewise.
7379 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7381 * config/arm/neon.md (vcond): Fix floating-point vector
7382 comparisons against 0.
7384 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
7386 * lra-constraints.c (emit_spill_move): Remove assert.
7387 (process_alt_operands): Add more debugging
7388 output. Increase reject for spilling into memory. Decrease
7389 reject for reloading scratch.
7390 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
7392 2013-08-08 Steve Ellcey <sellcey@mips.com>
7394 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
7395 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
7396 micromips incompatible. Add nan2008.
7397 (MULTILIB_DIRNAMES): Add nan2008.
7398 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
7399 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
7400 and micromips incompatible. Add nan2008.
7401 (MULTILIB_DIRNAMES): Add nan2008.
7402 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
7404 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
7406 PR rtl-optimization/58079
7407 * combine.c (combine_simplify_rtx): Avoid using SUBST if
7408 simplify_comparison has widened a comparison with an integer.
7410 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7412 * config/arm/neon.md (movmisalign<mode>): Disable when we
7413 don't allow unaligned accesses.
7414 (*movmisalign<mode>_neon_store): Likewise.
7415 (*movmisalign<mode>_neon_load): Likewise.
7416 (*movmisalign<mode>_neon_store): Likewise.
7417 (*movmisalign<mode>_neon_load): Likewise.
7419 2013-08-08 Jan Hubicka <jh@suse.cz>
7421 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
7422 (make_pass_rebuild_cgraph_edges): Also clear references.
7423 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
7424 * ipa-inline-transform.c (inline_transform): Remove all references
7426 * cgraphunit.c (expand_function): Remove all references after
7428 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
7429 (ipa_find_reference): Rewrite to iterator.
7430 (remove_stmt_references): Likewise.
7431 (ipa_clear_stmts_in_references): New function.
7432 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
7433 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
7435 * ipa-split.c (split_function): Remove references in split function.
7437 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7440 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
7441 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
7443 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7446 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
7447 suggested mode for the assignment isn't compatible with the
7450 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
7453 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
7455 2013-08-07 Xinliang David Li <davidxl@google.com>
7457 * config/i386/i386.opt: New option -mtune-ctrl=.
7458 * config/i386/x86-tune.def: New file.
7459 * config/i386/i386.h: include x86-tune.def.
7460 * config/i386/i386.c (ix86_option_override_internal):
7461 Parsing -mtune-ctrl= option and set tune features.
7463 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
7466 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
7467 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
7470 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
7472 * rtl.h (update_alignments): Declare.
7473 * final.c (grow_label_align): New function extracted from...
7474 (shorten_branches): ...here. Call it.
7475 (update_alignments): New function.
7476 * reorg.c (sibling_labels): New variable.
7477 (get_label_before): Add SIBLING parameter. If it is non-zero, push
7478 the new label along with it onto the sibling_labels vector.
7479 (fill_simple_delay_slots): Adjust call to get_label_before.
7480 (fill_slots_from_thread): Likewise.
7481 (relax_delay_slots): Likewise.
7482 (make_return_insns): Likewise.
7483 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
7485 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
7487 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
7488 document its semantics.
7489 (diagnostic_report_diagnostic): Adjust accordingly.
7491 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7493 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
7494 (sparc_option_override): ...and port to new C++ pass API.
7495 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
7497 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
7499 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
7501 2013-08-06 Caroline Tice <cmtice@google.com>
7503 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
7504 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
7505 * tree-pass.h: Add pass_vtable_verify.
7506 * varasm.c (assemble_variable): Add code to properly set the comdat
7507 section and name for the .vtable_map_vars section.
7508 (assemble_vtyv_preinit_initializer): New function.
7509 (default_sectin_type_flags): Make sure .vtable_map_vars section has
7511 * output.h: Add function decl for assemble_vtv_preinit_initializer.
7512 * vtable-verify.c: New file.
7513 * vtable-verify.h: New file.
7514 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
7515 initialiation levels.
7516 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
7517 * passes.def: Insert pass_vtable_verify.
7518 * aclocal.m4: Reorder includes.
7519 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
7520 -fvtv-counts options.
7521 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
7522 as appropriate, if -fvtable-verify=... is used.
7523 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
7524 -fvtable-verify=... is used.
7525 * Makefile.in (OBJS): Add vtable-verify.o to list.
7526 (vtable-verify.o): Add new build rule.
7527 (GTFILES): Add vtable-verify.c to list.
7528 * common.opt (fvtable-verify=): New flag.
7529 (vtv_priority): Values for fvtable-verify= flag.
7530 (fvtv-counts): New flag.
7531 (fvtv-debug): New flag.
7532 * tree.h (save_vtable_map_decl): New extern function decl.
7534 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7536 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
7537 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
7538 (pass_data_rl78_devirt): ...new pass_data instance and...
7539 (make_pass_rl78_devirt): ...new function.
7540 (rl78_asm_file_start): Port pass registration to new C++ API.
7542 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7544 * coretypes.h (rtl_opt_pass): Add.
7545 (gcc::context): Add.
7546 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
7547 (epiphany_init): Port to new C++ pass API.
7548 (epiphany_optimize_mode_switching): Likewise.
7549 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
7550 (pass_manager::get_pass_mode_switching): New.
7551 (pass_manager::get_pass_peephole2): New.
7552 * mode-switching.c (pass_mode_switching): Add clone method.
7553 * recog.c (pass_peephole2): Add clone method.
7554 (pass_split_all_insns): Add clone method.
7556 2013-08-06 David Malcolm <dmalcolm@redhat.com>
7558 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
7559 (mips_option_override): ...here, porting to new C++ API for passes.
7561 2013-08-06 Jan Hubicka <jh@suse.cz>
7563 * cgraph.c (cgraph_get_body): New function based on lto.c
7565 * cgraph.h (cgraph_get_body): Declare.
7566 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
7568 * cgraphunit.c (expand_function): Get body prior expanding.
7569 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
7570 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
7572 * passes.c (do_per_function_toporder): Get body.
7573 * tree-inline.c (expand_call_inline): Get body prior inlining it.
7574 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
7576 2013-08-06 Martin Jambor <mjambor@suse.cz>
7579 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
7580 is not re-finalized. Rename second parameter to no_collect.
7582 2013-08-06 Martin Jambor <mjambor@suse.cz>
7585 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
7586 MEM_REF has proper alignment information.
7588 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
7591 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
7593 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
7594 rtx (*) (rtx, ...) with insn_gen_fn.
7595 * genoutput.c (output_insn_data): Cast gen_? function pointers to
7596 insn_gen_fn::stored_funcptr. Add initializer braces.
7598 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7600 Rewrite how instances of passes are cloned to remove assumptions
7601 about their sizes (thus allowing pass subclasses to have
7602 additional data fields, albeit non-GC-managed ones at this point).
7604 * passes.c (make_pass_instance): Now that passes have clone
7605 methods, rewrite this function to eliminate XNEW and memcpy
7606 calls that used hardcoded sizes. Since this function no longer
7607 creates pass instances, rename it to...
7608 (add_pass_instance): ...this. Document the old way that passes were
7609 numbered and flagged, and rework this function to continue using it.
7610 (next_pass_1): Add an initial_pass argument for use by
7612 (position_pass): When adding multiple instances of a pass, use
7613 the pass's clone method, rather than relying on the XNEW/memcpy
7614 within the former make_pass_instance (now add_pass_instance).
7615 (pass_manager::pass_manager): When invoking next_pass_1, also supply
7616 the initial instance of the current pass within the pass manager.
7618 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7620 This is the automated part of the conversion of passes from C
7621 structs to C++ classes.
7623 Patch autogenerated by refactor_passes.py from
7624 https://github.com/davidmalcolm/gcc-refactoring-scripts
7625 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
7627 * asan.c (pass_asan): Convert from a global struct to a subclass of
7628 gimple_opt_pass along with...
7629 (pass_data_asan): ...new pass_data instance and...
7630 (make_pass_asan): ...new function.
7631 (pass_asan_O0): Convert from a global struct to a subclass of
7632 gimple_opt_pass along with...
7633 (pass_data_asan_O0): ...new pass_data instance and...
7634 (make_pass_asan_O0): ...new function.
7635 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
7636 subclass of rtl_opt_pass along with...
7637 (pass_data_inc_dec): ...new pass_data instance and...
7638 (make_pass_inc_dec): ...new function.
7639 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
7640 a subclass of rtl_opt_pass along with...
7641 (pass_data_reorder_blocks): ...new pass_data instance and...
7642 (make_pass_reorder_blocks): ...new function.
7643 (pass_duplicate_computed_gotos): Convert from a global struct to a
7644 subclass of rtl_opt_pass along with...
7645 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
7646 (make_pass_duplicate_computed_gotos): ...new function.
7647 (pass_partition_blocks): Convert from a global struct to a subclass of
7648 rtl_opt_pass along with...
7649 (pass_data_partition_blocks): ...new pass_data instance and...
7650 (make_pass_partition_blocks): ...new function.
7651 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
7652 struct to a subclass of rtl_opt_pass along with...
7653 (pass_data_branch_target_load_optimize1): ...new pass_data instance
7655 (make_pass_branch_target_load_optimize1): ...new function.
7656 (pass_branch_target_load_optimize2): Convert from a global struct to a
7657 subclass of rtl_opt_pass along with...
7658 (pass_data_branch_target_load_optimize2): ...new pass_data instance
7660 (make_pass_branch_target_load_optimize2): ...new function.
7661 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
7662 of rtl_opt_pass along with...
7663 (pass_data_jump): ...new pass_data instance and...
7664 (make_pass_jump): ...new function.
7665 (pass_jump2): Convert from a global struct to a subclass of
7666 rtl_opt_pass along with...
7667 (pass_data_jump2): ...new pass_data instance and...
7668 (make_pass_jump2): ...new function.
7669 * cfgexpand.c (pass_expand): Convert from a global struct to a
7670 subclass of rtl_opt_pass along with...
7671 (pass_data_expand): ...new pass_data instance and...
7672 (make_pass_expand): ...new function.
7673 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
7674 of rtl_opt_pass along with...
7675 (pass_data_free_cfg): ...new pass_data instance and...
7676 (make_pass_free_cfg): ...new function.
7677 (pass_into_cfg_layout_mode): Convert from a global struct to a
7678 subclass of rtl_opt_pass along with...
7679 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
7680 (make_pass_into_cfg_layout_mode): ...new function.
7681 (pass_outof_cfg_layout_mode): Convert from a global struct to a
7682 subclass of rtl_opt_pass along with...
7683 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
7684 (make_pass_outof_cfg_layout_mode): ...new function.
7685 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
7686 struct to a subclass of gimple_opt_pass along with...
7687 (pass_data_build_cgraph_edges): ...new pass_data instance and...
7688 (make_pass_build_cgraph_edges): ...new function.
7689 (pass_rebuild_cgraph_edges): Convert from a global struct to a
7690 subclass of gimple_opt_pass along with...
7691 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
7692 (make_pass_rebuild_cgraph_edges): ...new function.
7693 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
7694 subclass of gimple_opt_pass along with...
7695 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
7697 (make_pass_remove_cgraph_callee_edges): ...new function.
7698 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
7699 struct to a subclass of rtl_opt_pass along with...
7700 (pass_data_stack_adjustments): ...new pass_data instance and...
7701 (make_pass_stack_adjustments): ...new function.
7702 * combine.c (pass_combine): Convert from a global struct to a subclass
7703 of rtl_opt_pass along with...
7704 (pass_data_combine): ...new pass_data instance and...
7705 (make_pass_combine): ...new function.
7706 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
7707 global struct to a subclass of rtl_opt_pass along with...
7708 (pass_data_compare_elim_after_reload): ...new pass_data instance
7710 (make_pass_compare_elim_after_reload): ...new function.
7711 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
7712 of rtl_opt_pass along with...
7713 (pass_data_rtl_cprop): ...new pass_data instance and...
7714 (make_pass_rtl_cprop): ...new function.
7715 * cse.c (pass_cse): Convert from a global struct to a subclass of
7716 rtl_opt_pass along with...
7717 (pass_data_cse): ...new pass_data instance and...
7718 (make_pass_cse): ...new function.
7719 (pass_cse2): Convert from a global struct to a subclass of
7720 rtl_opt_pass along with...
7721 (pass_data_cse2): ...new pass_data instance and...
7722 (make_pass_cse2): ...new function.
7723 (pass_cse_after_global_opts): Convert from a global struct to a
7724 subclass of rtl_opt_pass along with...
7725 (pass_data_cse_after_global_opts): ...new pass_data instance and...
7726 (make_pass_cse_after_global_opts): ...new function.
7727 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
7728 of rtl_opt_pass along with...
7729 (pass_data_ud_rtl_dce): ...new pass_data instance and...
7730 (make_pass_ud_rtl_dce): ...new function.
7731 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
7732 rtl_opt_pass along with...
7733 (pass_data_fast_rtl_dce): ...new pass_data instance and...
7734 (make_pass_fast_rtl_dce): ...new function.
7735 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
7736 a subclass of rtl_opt_pass along with...
7737 (pass_data_df_initialize_opt): ...new pass_data instance and...
7738 (make_pass_df_initialize_opt): ...new function.
7739 (pass_df_initialize_no_opt): Convert from a global struct to a
7740 subclass of rtl_opt_pass along with...
7741 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
7742 (make_pass_df_initialize_no_opt): ...new function.
7743 (pass_df_finish): Convert from a global struct to a subclass of
7744 rtl_opt_pass along with...
7745 (pass_data_df_finish): ...new pass_data instance and...
7746 (make_pass_df_finish): ...new function.
7747 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
7748 rtl_opt_pass along with...
7749 (pass_data_rtl_dse1): ...new pass_data instance and...
7750 (make_pass_rtl_dse1): ...new function.
7751 (pass_rtl_dse2): Convert from a global struct to a subclass of
7752 rtl_opt_pass along with...
7753 (pass_data_rtl_dse2): ...new pass_data instance and...
7754 (make_pass_rtl_dse2): ...new function.
7755 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
7756 subclass of rtl_opt_pass along with...
7757 (pass_data_dwarf2_frame): ...new pass_data instance and...
7758 (make_pass_dwarf2_frame): ...new function.
7759 * except.c (pass_set_nothrow_function_flags): Convert from a global
7760 struct to a subclass of rtl_opt_pass along with...
7761 (pass_data_set_nothrow_function_flags): ...new pass_data instance
7763 (make_pass_set_nothrow_function_flags): ...new function.
7764 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
7765 subclass of rtl_opt_pass along with...
7766 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
7768 (make_pass_convert_to_eh_region_ranges): ...new function.
7769 * final.c (pass_compute_alignments): Convert from a global struct to a
7770 subclass of rtl_opt_pass along with...
7771 (pass_data_compute_alignments): ...new pass_data instance and...
7772 (make_pass_compute_alignments): ...new function.
7773 (pass_final): Convert from a global struct to a subclass of
7774 rtl_opt_pass along with...
7775 (pass_data_final): ...new pass_data instance and...
7776 (make_pass_final): ...new function.
7777 (pass_shorten_branches): Convert from a global struct to a subclass of
7778 rtl_opt_pass along with...
7779 (pass_data_shorten_branches): ...new pass_data instance and...
7780 (make_pass_shorten_branches): ...new function.
7781 (pass_clean_state): Convert from a global struct to a subclass of
7782 rtl_opt_pass along with...
7783 (pass_data_clean_state): ...new pass_data instance and...
7784 (make_pass_clean_state): ...new function.
7785 * function.c (pass_instantiate_virtual_regs): Convert from a global
7786 struct to a subclass of rtl_opt_pass along with...
7787 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
7788 (make_pass_instantiate_virtual_regs): ...new function.
7789 (pass_leaf_regs): Convert from a global struct to a subclass of
7790 rtl_opt_pass along with...
7791 (pass_data_leaf_regs): ...new pass_data instance and...
7792 (make_pass_leaf_regs): ...new function.
7793 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
7794 subclass of rtl_opt_pass along with...
7795 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
7797 (make_pass_thread_prologue_and_epilogue): ...new function.
7798 (pass_match_asm_constraints): Convert from a global struct to a
7799 subclass of rtl_opt_pass along with...
7800 (pass_data_match_asm_constraints): ...new pass_data instance and...
7801 (make_pass_match_asm_constraints): ...new function.
7802 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
7803 subclass of rtl_opt_pass along with...
7804 (pass_data_rtl_fwprop): ...new pass_data instance and...
7805 (make_pass_rtl_fwprop): ...new function.
7806 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
7807 rtl_opt_pass along with...
7808 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
7809 (make_pass_rtl_fwprop_addr): ...new function.
7810 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
7811 rtl_opt_pass along with...
7812 (pass_data_rtl_pre): ...new pass_data instance and...
7813 (make_pass_rtl_pre): ...new function.
7814 (pass_rtl_hoist): Convert from a global struct to a subclass of
7815 rtl_opt_pass along with...
7816 (pass_data_rtl_hoist): ...new pass_data instance and...
7817 (make_pass_rtl_hoist): ...new function.
7818 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
7819 subclass of gimple_opt_pass along with...
7820 (pass_data_lower_cf): ...new pass_data instance and...
7821 (make_pass_lower_cf): ...new function.
7822 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
7823 from a global struct to a subclass of gimple_opt_pass along with...
7824 (pass_data_strength_reduction): ...new pass_data instance and...
7825 (make_pass_strength_reduction): ...new function.
7826 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
7827 of rtl_opt_pass along with...
7828 (pass_data_rtl_ifcvt): ...new pass_data instance and...
7829 (make_pass_rtl_ifcvt): ...new function.
7830 (pass_if_after_combine): Convert from a global struct to a subclass of
7831 rtl_opt_pass along with...
7832 (pass_data_if_after_combine): ...new pass_data instance and...
7833 (make_pass_if_after_combine): ...new function.
7834 (pass_if_after_reload): Convert from a global struct to a subclass of
7835 rtl_opt_pass along with...
7836 (pass_data_if_after_reload): ...new pass_data instance and...
7837 (make_pass_if_after_reload): ...new function.
7838 * init-regs.c (pass_initialize_regs): Convert from a global struct to
7839 a subclass of rtl_opt_pass along with...
7840 (pass_data_initialize_regs): ...new pass_data instance and...
7841 (make_pass_initialize_regs): ...new function.
7842 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
7843 of ipa_opt_pass_d along with...
7844 (pass_data_ipa_cp): ...new pass_data instance and...
7845 (make_pass_ipa_cp): ...new function.
7846 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
7847 global struct to a subclass of gimple_opt_pass along with...
7848 (pass_data_inline_parameters): ...new pass_data instance and...
7849 (make_pass_inline_parameters): ...new function.
7850 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
7851 subclass of gimple_opt_pass along with...
7852 (pass_data_early_inline): ...new pass_data instance and...
7853 (make_pass_early_inline): ...new function.
7854 (pass_ipa_inline): Convert from a global struct to a subclass of
7855 ipa_opt_pass_d along with...
7856 (pass_data_ipa_inline): ...new pass_data instance and...
7857 (make_pass_ipa_inline): ...new function.
7858 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
7859 struct to a subclass of gimple_opt_pass along with...
7860 (pass_data_local_pure_const): ...new pass_data instance and...
7861 (make_pass_local_pure_const): ...new function.
7862 (pass_ipa_pure_const): Convert from a global struct to a subclass of
7863 ipa_opt_pass_d along with...
7864 (pass_data_ipa_pure_const): ...new pass_data instance and...
7865 (make_pass_ipa_pure_const): ...new function.
7866 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
7867 to a subclass of ipa_opt_pass_d along with...
7868 (pass_data_ipa_reference): ...new pass_data instance and...
7869 (make_pass_ipa_reference): ...new function.
7870 * ipa-split.c (pass_split_functions): Convert from a global struct to
7871 a subclass of gimple_opt_pass along with...
7872 (pass_data_split_functions): ...new pass_data instance and...
7873 (make_pass_split_functions): ...new function.
7874 (pass_feedback_split_functions): Convert from a global struct to a
7875 subclass of gimple_opt_pass along with...
7876 (pass_data_feedback_split_functions): ...new pass_data instance and...
7877 (make_pass_feedback_split_functions): ...new function.
7878 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
7879 global struct to a subclass of simple_ipa_opt_pass along with...
7880 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
7882 (make_pass_ipa_function_and_variable_visibility): ...new function.
7883 (pass_ipa_free_inline_summary): Convert from a global struct to a
7884 subclass of simple_ipa_opt_pass along with...
7885 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
7886 (make_pass_ipa_free_inline_summary): ...new function.
7887 (pass_ipa_whole_program_visibility): Convert from a global struct to a
7888 subclass of ipa_opt_pass_d along with...
7889 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
7891 (make_pass_ipa_whole_program_visibility): ...new function.
7892 (pass_ipa_profile): Convert from a global struct to a subclass of
7893 ipa_opt_pass_d along with...
7894 (pass_data_ipa_profile): ...new pass_data instance and...
7895 (make_pass_ipa_profile): ...new function.
7896 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
7897 ipa_opt_pass_d along with...
7898 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
7899 (make_pass_ipa_cdtor_merge): ...new function.
7900 * ira.c (pass_ira): Convert from a global struct to a subclass of
7901 rtl_opt_pass along with...
7902 (pass_data_ira): ...new pass_data instance and...
7903 (make_pass_ira): ...new function.
7904 (pass_reload): Convert from a global struct to a subclass of
7905 rtl_opt_pass along with...
7906 (pass_data_reload): ...new pass_data instance and...
7907 (make_pass_reload): ...new function.
7908 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
7909 subclass of rtl_opt_pass along with...
7910 (pass_data_cleanup_barriers): ...new pass_data instance and...
7911 (make_pass_cleanup_barriers): ...new function.
7912 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
7913 of rtl_opt_pass along with...
7914 (pass_data_loop2): ...new pass_data instance and...
7915 (make_pass_loop2): ...new function.
7916 (pass_rtl_loop_init): Convert from a global struct to a subclass of
7917 rtl_opt_pass along with...
7918 (pass_data_rtl_loop_init): ...new pass_data instance and...
7919 (make_pass_rtl_loop_init): ...new function.
7920 (pass_rtl_loop_done): Convert from a global struct to a subclass of
7921 rtl_opt_pass along with...
7922 (pass_data_rtl_loop_done): ...new pass_data instance and...
7923 (make_pass_rtl_loop_done): ...new function.
7924 (pass_rtl_move_loop_invariants): Convert from a global struct to a
7925 subclass of rtl_opt_pass along with...
7926 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
7927 (make_pass_rtl_move_loop_invariants): ...new function.
7928 (pass_rtl_unswitch): Convert from a global struct to a subclass of
7929 rtl_opt_pass along with...
7930 (pass_data_rtl_unswitch): ...new pass_data instance and...
7931 (make_pass_rtl_unswitch): ...new function.
7932 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
7933 subclass of rtl_opt_pass along with...
7934 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
7936 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7937 (pass_rtl_doloop): Convert from a global struct to a subclass of
7938 rtl_opt_pass along with...
7939 (pass_data_rtl_doloop): ...new pass_data instance and...
7940 (make_pass_rtl_doloop): ...new function.
7941 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
7942 a subclass of rtl_opt_pass along with...
7943 (pass_data_lower_subreg): ...new pass_data instance and...
7944 (make_pass_lower_subreg): ...new function.
7945 (pass_lower_subreg2): Convert from a global struct to a subclass of
7946 rtl_opt_pass along with...
7947 (pass_data_lower_subreg2): ...new pass_data instance and...
7948 (make_pass_lower_subreg2): ...new function.
7949 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
7950 struct to a subclass of ipa_opt_pass_d along with...
7951 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
7952 (make_pass_ipa_lto_gimple_out): ...new function.
7953 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
7954 of ipa_opt_pass_d along with...
7955 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
7956 (make_pass_ipa_lto_finish_out): ...new function.
7957 * mode-switching.c (pass_mode_switching): Convert from a global struct
7958 to a subclass of rtl_opt_pass along with...
7959 (pass_data_mode_switching): ...new pass_data instance and...
7960 (make_pass_mode_switching): ...new function.
7961 * modulo-sched.c (pass_sms): Convert from a global struct to a
7962 subclass of rtl_opt_pass along with...
7963 (pass_data_sms): ...new pass_data instance and...
7964 (make_pass_sms): ...new function.
7965 * omp-low.c (pass_expand_omp): Convert from a global struct to a
7966 subclass of gimple_opt_pass along with...
7967 (pass_data_expand_omp): ...new pass_data instance and...
7968 (make_pass_expand_omp): ...new function.
7969 (pass_lower_omp): Convert from a global struct to a subclass of
7970 gimple_opt_pass along with...
7971 (pass_data_lower_omp): ...new pass_data instance and...
7972 (make_pass_lower_omp): ...new function.
7973 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
7974 of gimple_opt_pass along with...
7975 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
7976 (make_pass_diagnose_omp_blocks): ...new function.
7977 * passes.c (pass_early_local_passes): Convert from a global struct to
7978 a subclass of simple_ipa_opt_pass along with...
7979 (pass_data_early_local_passes): ...new pass_data instance and...
7980 (make_pass_early_local_passes): ...new function.
7981 (pass_all_early_optimizations): Convert from a global struct to a
7982 subclass of gimple_opt_pass along with...
7983 (pass_data_all_early_optimizations): ...new pass_data instance and...
7984 (make_pass_all_early_optimizations): ...new function.
7985 (pass_all_optimizations): Convert from a global struct to a subclass
7986 of gimple_opt_pass along with...
7987 (pass_data_all_optimizations): ...new pass_data instance and...
7988 (make_pass_all_optimizations): ...new function.
7989 (pass_all_optimizations_g): Convert from a global struct to a subclass
7990 of gimple_opt_pass along with...
7991 (pass_data_all_optimizations_g): ...new pass_data instance and...
7992 (make_pass_all_optimizations_g): ...new function.
7993 (pass_rest_of_compilation): Convert from a global struct to a subclass
7994 of rtl_opt_pass along with...
7995 (pass_data_rest_of_compilation): ...new pass_data instance and...
7996 (make_pass_rest_of_compilation): ...new function.
7997 (pass_postreload): Convert from a global struct to a subclass of
7998 rtl_opt_pass along with...
7999 (pass_data_postreload): ...new pass_data instance and...
8000 (make_pass_postreload): ...new function.
8001 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
8002 subclass of rtl_opt_pass along with...
8003 (pass_data_gcse2): ...new pass_data instance and...
8004 (make_pass_gcse2): ...new function.
8005 * postreload.c (pass_postreload_cse): Convert from a global struct to
8006 a subclass of rtl_opt_pass along with...
8007 (pass_data_postreload_cse): ...new pass_data instance and...
8008 (make_pass_postreload_cse): ...new function.
8009 * predict.c (pass_profile): Convert from a global struct to a subclass
8010 of gimple_opt_pass along with...
8011 (pass_data_profile): ...new pass_data instance and...
8012 (make_pass_profile): ...new function.
8013 (pass_strip_predict_hints): Convert from a global struct to a subclass
8014 of gimple_opt_pass along with...
8015 (pass_data_strip_predict_hints): ...new pass_data instance and...
8016 (make_pass_strip_predict_hints): ...new function.
8017 * recog.c (pass_peephole2): Convert from a global struct to a subclass
8018 of rtl_opt_pass along with...
8019 (pass_data_peephole2): ...new pass_data instance and...
8020 (make_pass_peephole2): ...new function.
8021 (pass_split_all_insns): Convert from a global struct to a subclass of
8022 rtl_opt_pass along with...
8023 (pass_data_split_all_insns): ...new pass_data instance and...
8024 (make_pass_split_all_insns): ...new function.
8025 (pass_split_after_reload): Convert from a global struct to a subclass
8026 of rtl_opt_pass along with...
8027 (pass_data_split_after_reload): ...new pass_data instance and...
8028 (make_pass_split_after_reload): ...new function.
8029 (pass_split_before_regstack): Convert from a global struct to a
8030 subclass of rtl_opt_pass along with...
8031 (pass_data_split_before_regstack): ...new pass_data instance and...
8032 (make_pass_split_before_regstack): ...new function.
8033 (pass_split_before_sched2): Convert from a global struct to a subclass
8034 of rtl_opt_pass along with...
8035 (pass_data_split_before_sched2): ...new pass_data instance and...
8036 (make_pass_split_before_sched2): ...new function.
8037 (pass_split_for_shorten_branches): Convert from a global struct to a
8038 subclass of rtl_opt_pass along with...
8039 (pass_data_split_for_shorten_branches): ...new pass_data instance
8041 (make_pass_split_for_shorten_branches): ...new function.
8042 * ree.c (pass_ree): Convert from a global struct to a subclass of
8043 rtl_opt_pass along with...
8044 (pass_data_ree): ...new pass_data instance and...
8045 (make_pass_ree): ...new function.
8046 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
8047 subclass of rtl_opt_pass along with...
8048 (pass_data_stack_regs): ...new pass_data instance and...
8049 (make_pass_stack_regs): ...new function.
8050 (pass_stack_regs_run): Convert from a global struct to a subclass of
8051 rtl_opt_pass along with...
8052 (pass_data_stack_regs_run): ...new pass_data instance and...
8053 (make_pass_stack_regs_run): ...new function.
8054 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
8055 subclass of rtl_opt_pass along with...
8056 (pass_data_cprop_hardreg): ...new pass_data instance and...
8057 (make_pass_cprop_hardreg): ...new function.
8058 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
8059 subclass of rtl_opt_pass along with...
8060 (pass_data_reginfo_init): ...new pass_data instance and...
8061 (make_pass_reginfo_init): ...new function.
8062 * regmove.c (pass_regmove): Convert from a global struct to a subclass
8063 of rtl_opt_pass along with...
8064 (pass_data_regmove): ...new pass_data instance and...
8065 (make_pass_regmove): ...new function.
8066 * regrename.c (pass_regrename): Convert from a global struct to a
8067 subclass of rtl_opt_pass along with...
8068 (pass_data_regrename): ...new pass_data instance and...
8069 (make_pass_regrename): ...new function.
8070 * reorg.c (pass_delay_slots): Convert from a global struct to a
8071 subclass of rtl_opt_pass along with...
8072 (pass_data_delay_slots): ...new pass_data instance and...
8073 (make_pass_delay_slots): ...new function.
8074 (pass_machine_reorg): Convert from a global struct to a subclass of
8075 rtl_opt_pass along with...
8076 (pass_data_machine_reorg): ...new pass_data instance and...
8077 (make_pass_machine_reorg): ...new function.
8078 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
8079 of rtl_opt_pass along with...
8080 (pass_data_sched): ...new pass_data instance and...
8081 (make_pass_sched): ...new function.
8082 (pass_sched2): Convert from a global struct to a subclass of
8083 rtl_opt_pass along with...
8084 (pass_data_sched2): ...new pass_data instance and...
8085 (make_pass_sched2): ...new function.
8086 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
8087 to a subclass of rtl_opt_pass along with...
8088 (pass_data_stack_ptr_mod): ...new pass_data instance and...
8089 (make_pass_stack_ptr_mod): ...new function.
8090 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
8091 to a subclass of rtl_opt_pass along with...
8092 (pass_data_rtl_store_motion): ...new pass_data instance and...
8093 (make_pass_rtl_store_motion): ...new function.
8094 * tracer.c (pass_tracer): Convert from a global struct to a subclass
8095 of gimple_opt_pass along with...
8096 (pass_data_tracer): ...new pass_data instance and...
8097 (make_pass_tracer): ...new function.
8098 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
8099 to a subclass of gimple_opt_pass along with...
8100 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
8101 (make_pass_diagnose_tm_blocks): ...new function.
8102 (pass_lower_tm): Convert from a global struct to a subclass of
8103 gimple_opt_pass along with...
8104 (pass_data_lower_tm): ...new pass_data instance and...
8105 (make_pass_lower_tm): ...new function.
8106 (pass_tm_init): Convert from a global struct to a subclass of
8107 gimple_opt_pass along with...
8108 (pass_data_tm_init): ...new pass_data instance and...
8109 (make_pass_tm_init): ...new function.
8110 (pass_tm_mark): Convert from a global struct to a subclass of
8111 gimple_opt_pass along with...
8112 (pass_data_tm_mark): ...new pass_data instance and...
8113 (make_pass_tm_mark): ...new function.
8114 (pass_tm_edges): Convert from a global struct to a subclass of
8115 gimple_opt_pass along with...
8116 (pass_data_tm_edges): ...new pass_data instance and...
8117 (make_pass_tm_edges): ...new function.
8118 (pass_tm_memopt): Convert from a global struct to a subclass of
8119 gimple_opt_pass along with...
8120 (pass_data_tm_memopt): ...new pass_data instance and...
8121 (make_pass_tm_memopt): ...new function.
8122 (pass_ipa_tm): Convert from a global struct to a subclass of
8123 simple_ipa_opt_pass along with...
8124 (pass_data_ipa_tm): ...new pass_data instance and...
8125 (make_pass_ipa_tm): ...new function.
8126 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
8127 subclass of gimple_opt_pass along with...
8128 (pass_data_call_cdce): ...new pass_data instance and...
8129 (make_pass_call_cdce): ...new function.
8130 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
8131 subclass of gimple_opt_pass along with...
8132 (pass_data_build_cfg): ...new pass_data instance and...
8133 (make_pass_build_cfg): ...new function.
8134 (pass_split_crit_edges): Convert from a global struct to a subclass of
8135 gimple_opt_pass along with...
8136 (pass_data_split_crit_edges): ...new pass_data instance and...
8137 (make_pass_split_crit_edges): ...new function.
8138 (pass_warn_function_return): Convert from a global struct to a
8139 subclass of gimple_opt_pass along with...
8140 (pass_data_warn_function_return): ...new pass_data instance and...
8141 (make_pass_warn_function_return): ...new function.
8142 (pass_warn_function_noreturn): Convert from a global struct to a
8143 subclass of gimple_opt_pass along with...
8144 (pass_data_warn_function_noreturn): ...new pass_data instance and...
8145 (make_pass_warn_function_noreturn): ...new function.
8146 (pass_warn_unused_result): Convert from a global struct to a subclass
8147 of gimple_opt_pass along with...
8148 (pass_data_warn_unused_result): ...new pass_data instance and...
8149 (make_pass_warn_unused_result): ...new function.
8150 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
8151 a subclass of gimple_opt_pass along with...
8152 (pass_data_merge_phi): ...new pass_data instance and...
8153 (make_pass_merge_phi): ...new function.
8154 * tree-complex.c (pass_lower_complex): Convert from a global struct to
8155 a subclass of gimple_opt_pass along with...
8156 (pass_data_lower_complex): ...new pass_data instance and...
8157 (make_pass_lower_complex): ...new function.
8158 (pass_lower_complex_O0): Convert from a global struct to a subclass of
8159 gimple_opt_pass along with...
8160 (pass_data_lower_complex_O0): ...new pass_data instance and...
8161 (make_pass_lower_complex_O0): ...new function.
8162 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
8163 subclass of gimple_opt_pass along with...
8164 (pass_data_lower_eh): ...new pass_data instance and...
8165 (make_pass_lower_eh): ...new function.
8166 (pass_refactor_eh): Convert from a global struct to a subclass of
8167 gimple_opt_pass along with...
8168 (pass_data_refactor_eh): ...new pass_data instance and...
8169 (make_pass_refactor_eh): ...new function.
8170 (pass_lower_resx): Convert from a global struct to a subclass of
8171 gimple_opt_pass along with...
8172 (pass_data_lower_resx): ...new pass_data instance and...
8173 (make_pass_lower_resx): ...new function.
8174 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
8175 of gimple_opt_pass along with...
8176 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
8177 (make_pass_lower_eh_dispatch): ...new function.
8178 (pass_cleanup_eh): Convert from a global struct to a subclass of
8179 gimple_opt_pass along with...
8180 (pass_data_cleanup_eh): ...new pass_data instance and...
8181 (make_pass_cleanup_eh): ...new function.
8182 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
8183 to a subclass of simple_ipa_opt_pass along with...
8184 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
8185 (make_pass_ipa_lower_emutls): ...new function.
8186 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
8187 a subclass of gimple_opt_pass along with...
8188 (pass_data_if_conversion): ...new pass_data instance and...
8189 (make_pass_if_conversion): ...new function.
8190 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
8191 subclass of gimple_opt_pass along with...
8192 (pass_data_build_ssa): ...new pass_data instance and...
8193 (make_pass_build_ssa): ...new function.
8194 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
8195 global struct to a subclass of gimple_opt_pass along with...
8196 (pass_data_loop_distribution): ...new pass_data instance and...
8197 (make_pass_loop_distribution): ...new function.
8198 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
8199 subclass of gimple_opt_pass along with...
8200 (pass_data_mudflap_1): ...new pass_data instance and...
8201 (make_pass_mudflap_1): ...new function.
8202 (pass_mudflap_2): Convert from a global struct to a subclass of
8203 gimple_opt_pass along with...
8204 (pass_data_mudflap_2): ...new pass_data instance and...
8205 (make_pass_mudflap_2): ...new function.
8206 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
8207 subclass of gimple_opt_pass along with...
8208 (pass_data_mudflap_1): ...new pass_data instance and...
8209 (make_pass_mudflap_1): ...new function.
8210 (pass_mudflap_2): Convert from a global struct to a subclass of
8211 gimple_opt_pass along with...
8212 (pass_data_mudflap_2): ...new pass_data instance and...
8213 (make_pass_mudflap_2): ...new function.
8214 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
8215 gimple_opt_pass along with...
8216 (pass_data_nrv): ...new pass_data instance and...
8217 (make_pass_nrv): ...new function.
8218 (pass_return_slot): Convert from a global struct to a subclass of
8219 gimple_opt_pass along with...
8220 (pass_data_return_slot): ...new pass_data instance and...
8221 (make_pass_return_slot): ...new function.
8222 * tree-object-size.c (pass_object_sizes): Convert from a global struct
8223 to a subclass of gimple_opt_pass along with...
8224 (pass_data_object_sizes): ...new pass_data instance and...
8225 (make_pass_object_sizes): ...new function.
8226 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
8227 global struct to a subclass of gimple_opt_pass along with...
8228 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
8230 (make_pass_cleanup_cfg_post_optimizing): ...new function.
8231 (pass_fixup_cfg): Convert from a global struct to a subclass of
8232 gimple_opt_pass along with...
8233 (pass_data_fixup_cfg): ...new pass_data instance and...
8234 (make_pass_fixup_cfg): ...new function.
8235 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
8236 (make_pass_mudflap_1): ...new function.
8237 (pass_mudflap_2): Replace declaration with that of...
8238 (make_pass_mudflap_2): ...new function.
8239 (pass_asan): Replace declaration with that of...
8240 (make_pass_asan): ...new function.
8241 (pass_asan_O0): Replace declaration with that of...
8242 (make_pass_asan_O0): ...new function.
8243 (pass_tsan): Replace declaration with that of...
8244 (make_pass_tsan): ...new function.
8245 (pass_tsan_O0): Replace declaration with that of...
8246 (make_pass_tsan_O0): ...new function.
8247 (pass_lower_cf): Replace declaration with that of...
8248 (make_pass_lower_cf): ...new function.
8249 (pass_refactor_eh): Replace declaration with that of...
8250 (make_pass_refactor_eh): ...new function.
8251 (pass_lower_eh): Replace declaration with that of...
8252 (make_pass_lower_eh): ...new function.
8253 (pass_lower_eh_dispatch): Replace declaration with that of...
8254 (make_pass_lower_eh_dispatch): ...new function.
8255 (pass_lower_resx): Replace declaration with that of...
8256 (make_pass_lower_resx): ...new function.
8257 (pass_build_cfg): Replace declaration with that of...
8258 (make_pass_build_cfg): ...new function.
8259 (pass_early_tree_profile): Replace declaration with that of...
8260 (make_pass_early_tree_profile): ...new function.
8261 (pass_cleanup_eh): Replace declaration with that of...
8262 (make_pass_cleanup_eh): ...new function.
8263 (pass_sra): Replace declaration with that of...
8264 (make_pass_sra): ...new function.
8265 (pass_sra_early): Replace declaration with that of...
8266 (make_pass_sra_early): ...new function.
8267 (pass_early_ipa_sra): Replace declaration with that of...
8268 (make_pass_early_ipa_sra): ...new function.
8269 (pass_tail_recursion): Replace declaration with that of...
8270 (make_pass_tail_recursion): ...new function.
8271 (pass_tail_calls): Replace declaration with that of...
8272 (make_pass_tail_calls): ...new function.
8273 (pass_tree_loop): Replace declaration with that of...
8274 (make_pass_tree_loop): ...new function.
8275 (pass_tree_loop_init): Replace declaration with that of...
8276 (make_pass_tree_loop_init): ...new function.
8277 (pass_lim): Replace declaration with that of...
8278 (make_pass_lim): ...new function.
8279 (pass_tree_unswitch): Replace declaration with that of...
8280 (make_pass_tree_unswitch): ...new function.
8281 (pass_predcom): Replace declaration with that of...
8282 (make_pass_predcom): ...new function.
8283 (pass_iv_canon): Replace declaration with that of...
8284 (make_pass_iv_canon): ...new function.
8285 (pass_scev_cprop): Replace declaration with that of...
8286 (make_pass_scev_cprop): ...new function.
8287 (pass_empty_loop): Replace declaration with that of...
8288 (make_pass_empty_loop): ...new function.
8289 (pass_record_bounds): Replace declaration with that of...
8290 (make_pass_record_bounds): ...new function.
8291 (pass_graphite): Replace declaration with that of...
8292 (make_pass_graphite): ...new function.
8293 (pass_graphite_transforms): Replace declaration with that of...
8294 (make_pass_graphite_transforms): ...new function.
8295 (pass_if_conversion): Replace declaration with that of...
8296 (make_pass_if_conversion): ...new function.
8297 (pass_loop_distribution): Replace declaration with that of...
8298 (make_pass_loop_distribution): ...new function.
8299 (pass_vectorize): Replace declaration with that of...
8300 (make_pass_vectorize): ...new function.
8301 (pass_slp_vectorize): Replace declaration with that of...
8302 (make_pass_slp_vectorize): ...new function.
8303 (pass_complete_unroll): Replace declaration with that of...
8304 (make_pass_complete_unroll): ...new function.
8305 (pass_complete_unrolli): Replace declaration with that of...
8306 (make_pass_complete_unrolli): ...new function.
8307 (pass_parallelize_loops): Replace declaration with that of...
8308 (make_pass_parallelize_loops): ...new function.
8309 (pass_loop_prefetch): Replace declaration with that of...
8310 (make_pass_loop_prefetch): ...new function.
8311 (pass_iv_optimize): Replace declaration with that of...
8312 (make_pass_iv_optimize): ...new function.
8313 (pass_tree_loop_done): Replace declaration with that of...
8314 (make_pass_tree_loop_done): ...new function.
8315 (pass_ch): Replace declaration with that of...
8316 (make_pass_ch): ...new function.
8317 (pass_ccp): Replace declaration with that of...
8318 (make_pass_ccp): ...new function.
8319 (pass_phi_only_cprop): Replace declaration with that of...
8320 (make_pass_phi_only_cprop): ...new function.
8321 (pass_build_ssa): Replace declaration with that of...
8322 (make_pass_build_ssa): ...new function.
8323 (pass_build_alias): Replace declaration with that of...
8324 (make_pass_build_alias): ...new function.
8325 (pass_build_ealias): Replace declaration with that of...
8326 (make_pass_build_ealias): ...new function.
8327 (pass_dominator): Replace declaration with that of...
8328 (make_pass_dominator): ...new function.
8329 (pass_dce): Replace declaration with that of...
8330 (make_pass_dce): ...new function.
8331 (pass_dce_loop): Replace declaration with that of...
8332 (make_pass_dce_loop): ...new function.
8333 (pass_cd_dce): Replace declaration with that of...
8334 (make_pass_cd_dce): ...new function.
8335 (pass_call_cdce): Replace declaration with that of...
8336 (make_pass_call_cdce): ...new function.
8337 (pass_merge_phi): Replace declaration with that of...
8338 (make_pass_merge_phi): ...new function.
8339 (pass_split_crit_edges): Replace declaration with that of...
8340 (make_pass_split_crit_edges): ...new function.
8341 (pass_pre): Replace declaration with that of...
8342 (make_pass_pre): ...new function.
8343 (pass_profile): Replace declaration with that of...
8344 (make_pass_profile): ...new function.
8345 (pass_strip_predict_hints): Replace declaration with that of...
8346 (make_pass_strip_predict_hints): ...new function.
8347 (pass_lower_complex_O0): Replace declaration with that of...
8348 (make_pass_lower_complex_O0): ...new function.
8349 (pass_lower_complex): Replace declaration with that of...
8350 (make_pass_lower_complex): ...new function.
8351 (pass_lower_vector): Replace declaration with that of...
8352 (make_pass_lower_vector): ...new function.
8353 (pass_lower_vector_ssa): Replace declaration with that of...
8354 (make_pass_lower_vector_ssa): ...new function.
8355 (pass_lower_omp): Replace declaration with that of...
8356 (make_pass_lower_omp): ...new function.
8357 (pass_diagnose_omp_blocks): Replace declaration with that of...
8358 (make_pass_diagnose_omp_blocks): ...new function.
8359 (pass_expand_omp): Replace declaration with that of...
8360 (make_pass_expand_omp): ...new function.
8361 (pass_expand_omp_ssa): Replace declaration with that of...
8362 (make_pass_expand_omp_ssa): ...new function.
8363 (pass_object_sizes): Replace declaration with that of...
8364 (make_pass_object_sizes): ...new function.
8365 (pass_strlen): Replace declaration with that of...
8366 (make_pass_strlen): ...new function.
8367 (pass_fold_builtins): Replace declaration with that of...
8368 (make_pass_fold_builtins): ...new function.
8369 (pass_stdarg): Replace declaration with that of...
8370 (make_pass_stdarg): ...new function.
8371 (pass_early_warn_uninitialized): Replace declaration with that of...
8372 (make_pass_early_warn_uninitialized): ...new function.
8373 (pass_late_warn_uninitialized): Replace declaration with that of...
8374 (make_pass_late_warn_uninitialized): ...new function.
8375 (pass_cse_reciprocals): Replace declaration with that of...
8376 (make_pass_cse_reciprocals): ...new function.
8377 (pass_cse_sincos): Replace declaration with that of...
8378 (make_pass_cse_sincos): ...new function.
8379 (pass_optimize_bswap): Replace declaration with that of...
8380 (make_pass_optimize_bswap): ...new function.
8381 (pass_optimize_widening_mul): Replace declaration with that of...
8382 (make_pass_optimize_widening_mul): ...new function.
8383 (pass_warn_function_return): Replace declaration with that of...
8384 (make_pass_warn_function_return): ...new function.
8385 (pass_warn_function_noreturn): Replace declaration with that of...
8386 (make_pass_warn_function_noreturn): ...new function.
8387 (pass_cselim): Replace declaration with that of...
8388 (make_pass_cselim): ...new function.
8389 (pass_phiopt): Replace declaration with that of...
8390 (make_pass_phiopt): ...new function.
8391 (pass_forwprop): Replace declaration with that of...
8392 (make_pass_forwprop): ...new function.
8393 (pass_phiprop): Replace declaration with that of...
8394 (make_pass_phiprop): ...new function.
8395 (pass_tree_ifcombine): Replace declaration with that of...
8396 (make_pass_tree_ifcombine): ...new function.
8397 (pass_dse): Replace declaration with that of...
8398 (make_pass_dse): ...new function.
8399 (pass_nrv): Replace declaration with that of...
8400 (make_pass_nrv): ...new function.
8401 (pass_rename_ssa_copies): Replace declaration with that of...
8402 (make_pass_rename_ssa_copies): ...new function.
8403 (pass_sink_code): Replace declaration with that of...
8404 (make_pass_sink_code): ...new function.
8405 (pass_fre): Replace declaration with that of...
8406 (make_pass_fre): ...new function.
8407 (pass_check_data_deps): Replace declaration with that of...
8408 (make_pass_check_data_deps): ...new function.
8409 (pass_copy_prop): Replace declaration with that of...
8410 (make_pass_copy_prop): ...new function.
8411 (pass_vrp): Replace declaration with that of...
8412 (make_pass_vrp): ...new function.
8413 (pass_uncprop): Replace declaration with that of...
8414 (make_pass_uncprop): ...new function.
8415 (pass_return_slot): Replace declaration with that of...
8416 (make_pass_return_slot): ...new function.
8417 (pass_reassoc): Replace declaration with that of...
8418 (make_pass_reassoc): ...new function.
8419 (pass_rebuild_cgraph_edges): Replace declaration with that of...
8420 (make_pass_rebuild_cgraph_edges): ...new function.
8421 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
8422 (make_pass_remove_cgraph_callee_edges): ...new function.
8423 (pass_build_cgraph_edges): Replace declaration with that of...
8424 (make_pass_build_cgraph_edges): ...new function.
8425 (pass_local_pure_const): Replace declaration with that of...
8426 (make_pass_local_pure_const): ...new function.
8427 (pass_tracer): Replace declaration with that of...
8428 (make_pass_tracer): ...new function.
8429 (pass_warn_unused_result): Replace declaration with that of...
8430 (make_pass_warn_unused_result): ...new function.
8431 (pass_diagnose_tm_blocks): Replace declaration with that of...
8432 (make_pass_diagnose_tm_blocks): ...new function.
8433 (pass_lower_tm): Replace declaration with that of...
8434 (make_pass_lower_tm): ...new function.
8435 (pass_tm_init): Replace declaration with that of...
8436 (make_pass_tm_init): ...new function.
8437 (pass_tm_mark): Replace declaration with that of...
8438 (make_pass_tm_mark): ...new function.
8439 (pass_tm_memopt): Replace declaration with that of...
8440 (make_pass_tm_memopt): ...new function.
8441 (pass_tm_edges): Replace declaration with that of...
8442 (make_pass_tm_edges): ...new function.
8443 (pass_split_functions): Replace declaration with that of...
8444 (make_pass_split_functions): ...new function.
8445 (pass_feedback_split_functions): Replace declaration with that of...
8446 (make_pass_feedback_split_functions): ...new function.
8447 (pass_strength_reduction): Replace declaration with that of...
8448 (make_pass_strength_reduction): ...new function.
8449 (pass_ipa_lower_emutls): Replace declaration with that of...
8450 (make_pass_ipa_lower_emutls): ...new function.
8451 (pass_ipa_function_and_variable_visibility): Replace declaration with
8453 (make_pass_ipa_function_and_variable_visibility): ...new function.
8454 (pass_ipa_tree_profile): Replace declaration with that of...
8455 (make_pass_ipa_tree_profile): ...new function.
8456 (pass_early_local_passes): Replace declaration with that of...
8457 (make_pass_early_local_passes): ...new function.
8458 (pass_ipa_whole_program_visibility): Replace declaration with that
8460 (make_pass_ipa_whole_program_visibility): ...new function.
8461 (pass_ipa_lto_gimple_out): Replace declaration with that of...
8462 (make_pass_ipa_lto_gimple_out): ...new function.
8463 (pass_ipa_increase_alignment): Replace declaration with that of...
8464 (make_pass_ipa_increase_alignment): ...new function.
8465 (pass_ipa_inline): Replace declaration with that of...
8466 (make_pass_ipa_inline): ...new function.
8467 (pass_ipa_free_lang_data): Replace declaration with that of...
8468 (make_pass_ipa_free_lang_data): ...new function.
8469 (pass_ipa_free_inline_summary): Replace declaration with that of...
8470 (make_pass_ipa_free_inline_summary): ...new function.
8471 (pass_ipa_cp): Replace declaration with that of...
8472 (make_pass_ipa_cp): ...new function.
8473 (pass_ipa_reference): Replace declaration with that of...
8474 (make_pass_ipa_reference): ...new function.
8475 (pass_ipa_pure_const): Replace declaration with that of...
8476 (make_pass_ipa_pure_const): ...new function.
8477 (pass_ipa_pta): Replace declaration with that of...
8478 (make_pass_ipa_pta): ...new function.
8479 (pass_ipa_lto_finish_out): Replace declaration with that of...
8480 (make_pass_ipa_lto_finish_out): ...new function.
8481 (pass_ipa_tm): Replace declaration with that of...
8482 (make_pass_ipa_tm): ...new function.
8483 (pass_ipa_profile): Replace declaration with that of...
8484 (make_pass_ipa_profile): ...new function.
8485 (pass_ipa_cdtor_merge): Replace declaration with that of...
8486 (make_pass_ipa_cdtor_merge): ...new function.
8487 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
8489 (make_pass_cleanup_cfg_post_optimizing): ...new function.
8490 (pass_init_datastructures): Replace declaration with that of...
8491 (make_pass_init_datastructures): ...new function.
8492 (pass_fixup_cfg): Replace declaration with that of...
8493 (make_pass_fixup_cfg): ...new function.
8494 (pass_expand): Replace declaration with that of...
8495 (make_pass_expand): ...new function.
8496 (pass_instantiate_virtual_regs): Replace declaration with that of...
8497 (make_pass_instantiate_virtual_regs): ...new function.
8498 (pass_rtl_fwprop): Replace declaration with that of...
8499 (make_pass_rtl_fwprop): ...new function.
8500 (pass_rtl_fwprop_addr): Replace declaration with that of...
8501 (make_pass_rtl_fwprop_addr): ...new function.
8502 (pass_jump): Replace declaration with that of...
8503 (make_pass_jump): ...new function.
8504 (pass_jump2): Replace declaration with that of...
8505 (make_pass_jump2): ...new function.
8506 (pass_lower_subreg): Replace declaration with that of...
8507 (make_pass_lower_subreg): ...new function.
8508 (pass_cse): Replace declaration with that of...
8509 (make_pass_cse): ...new function.
8510 (pass_fast_rtl_dce): Replace declaration with that of...
8511 (make_pass_fast_rtl_dce): ...new function.
8512 (pass_ud_rtl_dce): Replace declaration with that of...
8513 (make_pass_ud_rtl_dce): ...new function.
8514 (pass_rtl_dce): Replace declaration with that of...
8515 (make_pass_rtl_dce): ...new function.
8516 (pass_rtl_dse1): Replace declaration with that of...
8517 (make_pass_rtl_dse1): ...new function.
8518 (pass_rtl_dse2): Replace declaration with that of...
8519 (make_pass_rtl_dse2): ...new function.
8520 (pass_rtl_dse3): Replace declaration with that of...
8521 (make_pass_rtl_dse3): ...new function.
8522 (pass_rtl_cprop): Replace declaration with that of...
8523 (make_pass_rtl_cprop): ...new function.
8524 (pass_rtl_pre): Replace declaration with that of...
8525 (make_pass_rtl_pre): ...new function.
8526 (pass_rtl_hoist): Replace declaration with that of...
8527 (make_pass_rtl_hoist): ...new function.
8528 (pass_rtl_store_motion): Replace declaration with that of...
8529 (make_pass_rtl_store_motion): ...new function.
8530 (pass_cse_after_global_opts): Replace declaration with that of...
8531 (make_pass_cse_after_global_opts): ...new function.
8532 (pass_rtl_ifcvt): Replace declaration with that of...
8533 (make_pass_rtl_ifcvt): ...new function.
8534 (pass_into_cfg_layout_mode): Replace declaration with that of...
8535 (make_pass_into_cfg_layout_mode): ...new function.
8536 (pass_outof_cfg_layout_mode): Replace declaration with that of...
8537 (make_pass_outof_cfg_layout_mode): ...new function.
8538 (pass_loop2): Replace declaration with that of...
8539 (make_pass_loop2): ...new function.
8540 (pass_rtl_loop_init): Replace declaration with that of...
8541 (make_pass_rtl_loop_init): ...new function.
8542 (pass_rtl_move_loop_invariants): Replace declaration with that of...
8543 (make_pass_rtl_move_loop_invariants): ...new function.
8544 (pass_rtl_unswitch): Replace declaration with that of...
8545 (make_pass_rtl_unswitch): ...new function.
8546 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
8547 (make_pass_rtl_unroll_and_peel_loops): ...new function.
8548 (pass_rtl_doloop): Replace declaration with that of...
8549 (make_pass_rtl_doloop): ...new function.
8550 (pass_rtl_loop_done): Replace declaration with that of...
8551 (make_pass_rtl_loop_done): ...new function.
8552 (pass_web): Replace declaration with that of...
8553 (make_pass_web): ...new function.
8554 (pass_cse2): Replace declaration with that of...
8555 (make_pass_cse2): ...new function.
8556 (pass_df_initialize_opt): Replace declaration with that of...
8557 (make_pass_df_initialize_opt): ...new function.
8558 (pass_df_initialize_no_opt): Replace declaration with that of...
8559 (make_pass_df_initialize_no_opt): ...new function.
8560 (pass_reginfo_init): Replace declaration with that of...
8561 (make_pass_reginfo_init): ...new function.
8562 (pass_inc_dec): Replace declaration with that of...
8563 (make_pass_inc_dec): ...new function.
8564 (pass_stack_ptr_mod): Replace declaration with that of...
8565 (make_pass_stack_ptr_mod): ...new function.
8566 (pass_initialize_regs): Replace declaration with that of...
8567 (make_pass_initialize_regs): ...new function.
8568 (pass_combine): Replace declaration with that of...
8569 (make_pass_combine): ...new function.
8570 (pass_if_after_combine): Replace declaration with that of...
8571 (make_pass_if_after_combine): ...new function.
8572 (pass_ree): Replace declaration with that of...
8573 (make_pass_ree): ...new function.
8574 (pass_partition_blocks): Replace declaration with that of...
8575 (make_pass_partition_blocks): ...new function.
8576 (pass_match_asm_constraints): Replace declaration with that of...
8577 (make_pass_match_asm_constraints): ...new function.
8578 (pass_regmove): Replace declaration with that of...
8579 (make_pass_regmove): ...new function.
8580 (pass_split_all_insns): Replace declaration with that of...
8581 (make_pass_split_all_insns): ...new function.
8582 (pass_fast_rtl_byte_dce): Replace declaration with that of...
8583 (make_pass_fast_rtl_byte_dce): ...new function.
8584 (pass_lower_subreg2): Replace declaration with that of...
8585 (make_pass_lower_subreg2): ...new function.
8586 (pass_mode_switching): Replace declaration with that of...
8587 (make_pass_mode_switching): ...new function.
8588 (pass_sms): Replace declaration with that of...
8589 (make_pass_sms): ...new function.
8590 (pass_sched): Replace declaration with that of...
8591 (make_pass_sched): ...new function.
8592 (pass_ira): Replace declaration with that of...
8593 (make_pass_ira): ...new function.
8594 (pass_reload): Replace declaration with that of...
8595 (make_pass_reload): ...new function.
8596 (pass_clean_state): Replace declaration with that of...
8597 (make_pass_clean_state): ...new function.
8598 (pass_branch_prob): Replace declaration with that of...
8599 (make_pass_branch_prob): ...new function.
8600 (pass_value_profile_transformations): Replace declaration with that
8602 (make_pass_value_profile_transformations): ...new function.
8603 (pass_postreload_cse): Replace declaration with that of...
8604 (make_pass_postreload_cse): ...new function.
8605 (pass_gcse2): Replace declaration with that of...
8606 (make_pass_gcse2): ...new function.
8607 (pass_split_after_reload): Replace declaration with that of...
8608 (make_pass_split_after_reload): ...new function.
8609 (pass_branch_target_load_optimize1): Replace declaration with that
8611 (make_pass_branch_target_load_optimize1): ...new function.
8612 (pass_thread_prologue_and_epilogue): Replace declaration with that
8614 (make_pass_thread_prologue_and_epilogue): ...new function.
8615 (pass_stack_adjustments): Replace declaration with that of...
8616 (make_pass_stack_adjustments): ...new function.
8617 (pass_peephole2): Replace declaration with that of...
8618 (make_pass_peephole2): ...new function.
8619 (pass_if_after_reload): Replace declaration with that of...
8620 (make_pass_if_after_reload): ...new function.
8621 (pass_regrename): Replace declaration with that of...
8622 (make_pass_regrename): ...new function.
8623 (pass_cprop_hardreg): Replace declaration with that of...
8624 (make_pass_cprop_hardreg): ...new function.
8625 (pass_reorder_blocks): Replace declaration with that of...
8626 (make_pass_reorder_blocks): ...new function.
8627 (pass_branch_target_load_optimize2): Replace declaration with that
8629 (make_pass_branch_target_load_optimize2): ...new function.
8630 (pass_leaf_regs): Replace declaration with that of...
8631 (make_pass_leaf_regs): ...new function.
8632 (pass_split_before_sched2): Replace declaration with that of...
8633 (make_pass_split_before_sched2): ...new function.
8634 (pass_compare_elim_after_reload): Replace declaration with that of...
8635 (make_pass_compare_elim_after_reload): ...new function.
8636 (pass_sched2): Replace declaration with that of...
8637 (make_pass_sched2): ...new function.
8638 (pass_stack_regs): Replace declaration with that of...
8639 (make_pass_stack_regs): ...new function.
8640 (pass_stack_regs_run): Replace declaration with that of...
8641 (make_pass_stack_regs_run): ...new function.
8642 (pass_df_finish): Replace declaration with that of...
8643 (make_pass_df_finish): ...new function.
8644 (pass_compute_alignments): Replace declaration with that of...
8645 (make_pass_compute_alignments): ...new function.
8646 (pass_duplicate_computed_gotos): Replace declaration with that of...
8647 (make_pass_duplicate_computed_gotos): ...new function.
8648 (pass_variable_tracking): Replace declaration with that of...
8649 (make_pass_variable_tracking): ...new function.
8650 (pass_free_cfg): Replace declaration with that of...
8651 (make_pass_free_cfg): ...new function.
8652 (pass_machine_reorg): Replace declaration with that of...
8653 (make_pass_machine_reorg): ...new function.
8654 (pass_cleanup_barriers): Replace declaration with that of...
8655 (make_pass_cleanup_barriers): ...new function.
8656 (pass_delay_slots): Replace declaration with that of...
8657 (make_pass_delay_slots): ...new function.
8658 (pass_split_for_shorten_branches): Replace declaration with that of...
8659 (make_pass_split_for_shorten_branches): ...new function.
8660 (pass_split_before_regstack): Replace declaration with that of...
8661 (make_pass_split_before_regstack): ...new function.
8662 (pass_convert_to_eh_region_ranges): Replace declaration with that
8664 (make_pass_convert_to_eh_region_ranges): ...new function.
8665 (pass_shorten_branches): Replace declaration with that of...
8666 (make_pass_shorten_branches): ...new function.
8667 (pass_set_nothrow_function_flags): Replace declaration with that of...
8668 (make_pass_set_nothrow_function_flags): ...new function.
8669 (pass_dwarf2_frame): Replace declaration with that of...
8670 (make_pass_dwarf2_frame): ...new function.
8671 (pass_final): Replace declaration with that of...
8672 (make_pass_final): ...new function.
8673 (pass_rtl_seqabstr): Replace declaration with that of...
8674 (make_pass_rtl_seqabstr): ...new function.
8675 (pass_release_ssa_names): Replace declaration with that of...
8676 (make_pass_release_ssa_names): ...new function.
8677 (pass_early_inline): Replace declaration with that of...
8678 (make_pass_early_inline): ...new function.
8679 (pass_inline_parameters): Replace declaration with that of...
8680 (make_pass_inline_parameters): ...new function.
8681 (pass_update_address_taken): Replace declaration with that of...
8682 (make_pass_update_address_taken): ...new function.
8683 (pass_convert_switch): Replace declaration with that of...
8684 (make_pass_convert_switch): ...new function.
8685 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
8686 to a subclass of simple_ipa_opt_pass along with...
8687 (pass_data_ipa_tree_profile): ...new pass_data instance and...
8688 (make_pass_ipa_tree_profile): ...new function.
8689 * tree-sra.c (pass_sra_early): Convert from a global struct to a
8690 subclass of gimple_opt_pass along with...
8691 (pass_data_sra_early): ...new pass_data instance and...
8692 (make_pass_sra_early): ...new function.
8693 (pass_sra): Convert from a global struct to a subclass of
8694 gimple_opt_pass along with...
8695 (pass_data_sra): ...new pass_data instance and...
8696 (make_pass_sra): ...new function.
8697 (pass_early_ipa_sra): Convert from a global struct to a subclass of
8698 gimple_opt_pass along with...
8699 (pass_data_early_ipa_sra): ...new pass_data instance and...
8700 (make_pass_early_ipa_sra): ...new function.
8701 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
8702 subclass of gimple_opt_pass along with...
8703 (pass_data_ccp): ...new pass_data instance and...
8704 (make_pass_ccp): ...new function.
8705 (pass_fold_builtins): Convert from a global struct to a subclass of
8706 gimple_opt_pass along with...
8707 (pass_data_fold_builtins): ...new pass_data instance and...
8708 (make_pass_fold_builtins): ...new function.
8709 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
8710 subclass of gimple_opt_pass along with...
8711 (pass_data_copy_prop): ...new pass_data instance and...
8712 (make_pass_copy_prop): ...new function.
8713 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
8714 global struct to a subclass of gimple_opt_pass along with...
8715 (pass_data_rename_ssa_copies): ...new pass_data instance and...
8716 (make_pass_rename_ssa_copies): ...new function.
8717 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
8718 subclass of gimple_opt_pass along with...
8719 (pass_data_dce): ...new pass_data instance and...
8720 (make_pass_dce): ...new function.
8721 (pass_dce_loop): Convert from a global struct to a subclass of
8722 gimple_opt_pass along with...
8723 (pass_data_dce_loop): ...new pass_data instance and...
8724 (make_pass_dce_loop): ...new function.
8725 (pass_cd_dce): Convert from a global struct to a subclass of
8726 gimple_opt_pass along with...
8727 (pass_data_cd_dce): ...new pass_data instance and...
8728 (make_pass_cd_dce): ...new function.
8729 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
8730 subclass of gimple_opt_pass along with...
8731 (pass_data_dominator): ...new pass_data instance and...
8732 (make_pass_dominator): ...new function.
8733 (pass_phi_only_cprop): Convert from a global struct to a subclass of
8734 gimple_opt_pass along with...
8735 (pass_data_phi_only_cprop): ...new pass_data instance and...
8736 (make_pass_phi_only_cprop): ...new function.
8737 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
8738 subclass of gimple_opt_pass along with...
8739 (pass_data_dse): ...new pass_data instance and...
8740 (make_pass_dse): ...new function.
8741 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
8742 a subclass of gimple_opt_pass along with...
8743 (pass_data_forwprop): ...new pass_data instance and...
8744 (make_pass_forwprop): ...new function.
8745 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
8746 struct to a subclass of gimple_opt_pass along with...
8747 (pass_data_tree_ifcombine): ...new pass_data instance and...
8748 (make_pass_tree_ifcombine): ...new function.
8749 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
8750 subclass of gimple_opt_pass along with...
8751 (pass_data_ch): ...new pass_data instance and...
8752 (make_pass_ch): ...new function.
8753 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
8754 subclass of gimple_opt_pass along with...
8755 (pass_data_tree_loop): ...new pass_data instance and...
8756 (make_pass_tree_loop): ...new function.
8757 (pass_tree_loop_init): Convert from a global struct to a subclass of
8758 gimple_opt_pass along with...
8759 (pass_data_tree_loop_init): ...new pass_data instance and...
8760 (make_pass_tree_loop_init): ...new function.
8761 (pass_lim): Convert from a global struct to a subclass of
8762 gimple_opt_pass along with...
8763 (pass_data_lim): ...new pass_data instance and...
8764 (make_pass_lim): ...new function.
8765 (pass_tree_unswitch): Convert from a global struct to a subclass of
8766 gimple_opt_pass along with...
8767 (pass_data_tree_unswitch): ...new pass_data instance and...
8768 (make_pass_tree_unswitch): ...new function.
8769 (pass_predcom): Convert from a global struct to a subclass of
8770 gimple_opt_pass along with...
8771 (pass_data_predcom): ...new pass_data instance and...
8772 (make_pass_predcom): ...new function.
8773 (pass_vectorize): Convert from a global struct to a subclass of
8774 gimple_opt_pass along with...
8775 (pass_data_vectorize): ...new pass_data instance and...
8776 (make_pass_vectorize): ...new function.
8777 (pass_graphite): Convert from a global struct to a subclass of
8778 gimple_opt_pass along with...
8779 (pass_data_graphite): ...new pass_data instance and...
8780 (make_pass_graphite): ...new function.
8781 (pass_graphite_transforms): Convert from a global struct to a subclass
8782 of gimple_opt_pass along with...
8783 (pass_data_graphite_transforms): ...new pass_data instance and...
8784 (make_pass_graphite_transforms): ...new function.
8785 (pass_check_data_deps): Convert from a global struct to a subclass of
8786 gimple_opt_pass along with...
8787 (pass_data_check_data_deps): ...new pass_data instance and...
8788 (make_pass_check_data_deps): ...new function.
8789 (pass_iv_canon): Convert from a global struct to a subclass of
8790 gimple_opt_pass along with...
8791 (pass_data_iv_canon): ...new pass_data instance and...
8792 (make_pass_iv_canon): ...new function.
8793 (pass_scev_cprop): Convert from a global struct to a subclass of
8794 gimple_opt_pass along with...
8795 (pass_data_scev_cprop): ...new pass_data instance and...
8796 (make_pass_scev_cprop): ...new function.
8797 (pass_record_bounds): Convert from a global struct to a subclass of
8798 gimple_opt_pass along with...
8799 (pass_data_record_bounds): ...new pass_data instance and...
8800 (make_pass_record_bounds): ...new function.
8801 (pass_complete_unroll): Convert from a global struct to a subclass of
8802 gimple_opt_pass along with...
8803 (pass_data_complete_unroll): ...new pass_data instance and...
8804 (make_pass_complete_unroll): ...new function.
8805 (pass_complete_unrolli): Convert from a global struct to a subclass of
8806 gimple_opt_pass along with...
8807 (pass_data_complete_unrolli): ...new pass_data instance and...
8808 (make_pass_complete_unrolli): ...new function.
8809 (pass_parallelize_loops): Convert from a global struct to a subclass
8810 of gimple_opt_pass along with...
8811 (pass_data_parallelize_loops): ...new pass_data instance and...
8812 (make_pass_parallelize_loops): ...new function.
8813 (pass_loop_prefetch): Convert from a global struct to a subclass of
8814 gimple_opt_pass along with...
8815 (pass_data_loop_prefetch): ...new pass_data instance and...
8816 (make_pass_loop_prefetch): ...new function.
8817 (pass_iv_optimize): Convert from a global struct to a subclass of
8818 gimple_opt_pass along with...
8819 (pass_data_iv_optimize): ...new pass_data instance and...
8820 (make_pass_iv_optimize): ...new function.
8821 (pass_tree_loop_done): Convert from a global struct to a subclass of
8822 gimple_opt_pass along with...
8823 (pass_data_tree_loop_done): ...new pass_data instance and...
8824 (make_pass_tree_loop_done): ...new function.
8825 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
8826 struct to a subclass of gimple_opt_pass along with...
8827 (pass_data_cse_reciprocals): ...new pass_data instance and...
8828 (make_pass_cse_reciprocals): ...new function.
8829 (pass_cse_sincos): Convert from a global struct to a subclass of
8830 gimple_opt_pass along with...
8831 (pass_data_cse_sincos): ...new pass_data instance and...
8832 (make_pass_cse_sincos): ...new function.
8833 (pass_optimize_bswap): Convert from a global struct to a subclass of
8834 gimple_opt_pass along with...
8835 (pass_data_optimize_bswap): ...new pass_data instance and...
8836 (make_pass_optimize_bswap): ...new function.
8837 (pass_optimize_widening_mul): Convert from a global struct to a
8838 subclass of gimple_opt_pass along with...
8839 (pass_data_optimize_widening_mul): ...new pass_data instance and...
8840 (make_pass_optimize_widening_mul): ...new function.
8841 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
8842 subclass of gimple_opt_pass along with...
8843 (pass_data_phiopt): ...new pass_data instance and...
8844 (make_pass_phiopt): ...new function.
8845 (pass_cselim): Convert from a global struct to a subclass of
8846 gimple_opt_pass along with...
8847 (pass_data_cselim): ...new pass_data instance and...
8848 (make_pass_cselim): ...new function.
8849 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
8850 subclass of gimple_opt_pass along with...
8851 (pass_data_phiprop): ...new pass_data instance and...
8852 (make_pass_phiprop): ...new function.
8853 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
8854 subclass of gimple_opt_pass along with...
8855 (pass_data_pre): ...new pass_data instance and...
8856 (make_pass_pre): ...new function.
8857 (pass_fre): Convert from a global struct to a subclass of
8858 gimple_opt_pass along with...
8859 (pass_data_fre): ...new pass_data instance and...
8860 (make_pass_fre): ...new function.
8861 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
8862 subclass of gimple_opt_pass along with...
8863 (pass_data_reassoc): ...new pass_data instance and...
8864 (make_pass_reassoc): ...new function.
8865 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
8866 subclass of gimple_opt_pass along with...
8867 (pass_data_sink_code): ...new pass_data instance and...
8868 (make_pass_sink_code): ...new function.
8869 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
8870 subclass of gimple_opt_pass along with...
8871 (pass_data_strlen): ...new pass_data instance and...
8872 (make_pass_strlen): ...new function.
8873 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
8874 struct to a subclass of gimple_opt_pass along with...
8875 (pass_data_build_alias): ...new pass_data instance and...
8876 (make_pass_build_alias): ...new function.
8877 (pass_build_ealias): Convert from a global struct to a subclass of
8878 gimple_opt_pass along with...
8879 (pass_data_build_ealias): ...new pass_data instance and...
8880 (make_pass_build_ealias): ...new function.
8881 (pass_ipa_pta): Convert from a global struct to a subclass of
8882 simple_ipa_opt_pass along with...
8883 (pass_data_ipa_pta): ...new pass_data instance and...
8884 (make_pass_ipa_pta): ...new function.
8885 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
8886 subclass of gimple_opt_pass along with...
8887 (pass_data_uncprop): ...new pass_data instance and...
8888 (make_pass_uncprop): ...new function.
8889 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
8890 global struct to a subclass of gimple_opt_pass along with...
8891 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
8892 (make_pass_late_warn_uninitialized): ...new function.
8893 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
8894 to a subclass of gimple_opt_pass along with...
8895 (pass_data_init_datastructures): ...new pass_data instance and...
8896 (make_pass_init_datastructures): ...new function.
8897 (pass_early_warn_uninitialized): Convert from a global struct to a
8898 subclass of gimple_opt_pass along with...
8899 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
8900 (make_pass_early_warn_uninitialized): ...new function.
8901 (pass_update_address_taken): Convert from a global struct to a
8902 subclass of gimple_opt_pass along with...
8903 (pass_data_update_address_taken): ...new pass_data instance and...
8904 (make_pass_update_address_taken): ...new function.
8905 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
8906 struct to a subclass of gimple_opt_pass along with...
8907 (pass_data_release_ssa_names): ...new pass_data instance and...
8908 (make_pass_release_ssa_names): ...new function.
8909 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
8910 subclass of gimple_opt_pass along with...
8911 (pass_data_stdarg): ...new pass_data instance and...
8912 (make_pass_stdarg): ...new function.
8913 * tree-switch-conversion.c (pass_convert_switch): Convert from a
8914 global struct to a subclass of gimple_opt_pass along with...
8915 (pass_data_convert_switch): ...new pass_data instance and...
8916 (make_pass_convert_switch): ...new function.
8917 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
8918 to a subclass of gimple_opt_pass along with...
8919 (pass_data_tail_recursion): ...new pass_data instance and...
8920 (make_pass_tail_recursion): ...new function.
8921 (pass_tail_calls): Convert from a global struct to a subclass of
8922 gimple_opt_pass along with...
8923 (pass_data_tail_calls): ...new pass_data instance and...
8924 (make_pass_tail_calls): ...new function.
8925 * tree-vect-generic.c (pass_lower_vector): Convert from a global
8926 struct to a subclass of gimple_opt_pass along with...
8927 (pass_data_lower_vector): ...new pass_data instance and...
8928 (make_pass_lower_vector): ...new function.
8929 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
8930 gimple_opt_pass along with...
8931 (pass_data_lower_vector_ssa): ...new pass_data instance and...
8932 (make_pass_lower_vector_ssa): ...new function.
8933 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
8934 to a subclass of gimple_opt_pass along with...
8935 (pass_data_slp_vectorize): ...new pass_data instance and...
8936 (make_pass_slp_vectorize): ...new function.
8937 (pass_ipa_increase_alignment): Convert from a global struct to a
8938 subclass of simple_ipa_opt_pass along with...
8939 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
8940 (make_pass_ipa_increase_alignment): ...new function.
8941 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
8942 gimple_opt_pass along with...
8943 (pass_data_vrp): ...new pass_data instance and...
8944 (make_pass_vrp): ...new function.
8945 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
8946 subclass of simple_ipa_opt_pass along with...
8947 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
8948 (make_pass_ipa_free_lang_data): ...new function.
8949 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
8950 gimple_opt_pass along with...
8951 (pass_data_tsan): ...new pass_data instance and...
8952 (make_pass_tsan): ...new function.
8953 (pass_tsan_O0): Convert from a global struct to a subclass of
8954 gimple_opt_pass along with...
8955 (pass_data_tsan_O0): ...new pass_data instance and...
8956 (make_pass_tsan_O0): ...new function.
8957 * var-tracking.c (pass_variable_tracking): Convert from a global
8958 struct to a subclass of rtl_opt_pass along with...
8959 (pass_data_variable_tracking): ...new pass_data instance and...
8960 (make_pass_variable_tracking): ...new function.
8961 * web.c (pass_web): Convert from a global struct to a subclass of
8962 rtl_opt_pass along with...
8963 (pass_data_web): ...new pass_data instance and...
8964 (make_pass_web): ...new function.
8965 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
8966 declaration with that of...
8967 (make_pass_mode_switch_use): ...new function.
8968 (pass_resolve_sw_modes): Replace declaration with that of...
8969 (make_pass_resolve_sw_modes): ...new function.
8970 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
8971 from a global struct to a subclass of rtl_opt_pass along with...
8972 (pass_data_mode_switch_use): ...new pass_data instance and...
8973 (make_pass_mode_switch_use): ...new function.
8974 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
8975 from a global struct to a subclass of rtl_opt_pass along with...
8976 (pass_data_resolve_sw_modes): ...new pass_data instance and...
8977 (make_pass_resolve_sw_modes): ...new function.
8978 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
8979 struct to a subclass of rtl_opt_pass along with...
8980 (pass_data_insert_vzeroupper): ...new pass_data instance and...
8981 (make_pass_insert_vzeroupper): ...new function.
8982 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
8983 global struct to a subclass of rtl_opt_pass along with...
8984 (pass_data_work_around_errata): ...new pass_data instance and...
8985 (make_pass_work_around_errata): ...new function.
8986 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
8987 struct to a subclass of rtl_opt_pass along with...
8988 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
8989 (make_pass_mips_machine_reorg2): ...new function.
8991 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8993 * passes.c (pass_manager::operator new): New.
8995 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8997 Handwritten part of conversion of passes to C++ classes.
8999 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
9000 (toplev.o): Add dep on PASS_MANAGER_H.
9001 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
9002 of early local pases to reflect this moving from a global to a
9003 member of gcc::pass_manager.
9004 (cgraph_add_new_function): Likewise.
9005 * lto-cgraph.c (lto_output_node): Update for conversion of
9006 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
9007 * passes.c (opt_pass::clone): New.
9008 (opt_pass::gate): New.
9009 (opt_pass::execute): New.
9010 (opt_pass::opt_pass): New.
9011 (pass_manager::execute_early_local_passes): New.
9012 (pass_manager::execute_pass_mode_switching): new.
9013 (finish_optimization_passes): Convert to...
9014 (pass_manager::finish_optimization_passes): ...this.
9015 (finish_optimization_passes): Update for conversion of passes to
9017 (register_dump_files_1): Use has_gate since we cannot portably
9018 check a vtable entry against NULL.
9019 (dump_one_pass): Likewise.
9020 (ipa_write_summaries_2): Likewise.
9021 (ipa_write_optimization_summaries_1): Likewise.
9022 (ipa_read_summaries_1): Likewise.
9023 (ipa_read_optimization_summaries_1): Likewise.
9024 (execute_ipa_stmt_fixups): Likewise.
9025 (pass_manager::pass_manager): Rewrite pass-creation, invoking
9026 pass-creation functions rather than wiring up globals, and
9027 storing the results in fields of pass_manager generated using
9029 (pass_manager::dump_profile_report): Update for conversion of
9030 passes to C++ classes.
9031 (pass_manager::execute_ipa_summary_passes): Likewise.
9032 (execute_one_ipa_transform_pass): Likewise.
9033 (execute_one_pass): Use has_gate and has_execute since we cannot
9034 portably check a vtable entry against NULL.
9035 * pass_manager.h (pass_manager::finish_optimization_passes): New.
9036 (pass_manager): Use pass-instances.def to add fields for the
9037 various pass instances.
9038 * toplev.c (finalize): Update for move of
9039 finish_optimization_passes to a method of gcc::pass_manager.
9040 * toplev.h (finish_optimization_passes): Move to method of class
9042 * tree-pass.h (struct pass_data): New.
9043 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
9044 (opt_pass::gate): Convert to virtual function.
9045 (opt_pass::~opt_pass): New.
9046 (opt_pass::clone): New.
9047 (opt_pass::execute): Convert to virtual function.
9048 (opt_pass::opt_pass): New.
9049 (opt_pass::ctxt_): new.
9050 (gimple_opt_pass): Convert to subclass of opt_pass.
9051 (gimple_opt_pass::gimple_opt_pass): New.
9052 (rtl_opt_pass): Convert to subclass of opt_pass.
9053 (rtl_opt_pass::rtl_opt_pass): New.
9054 (ipa_opt_pass_d): Convert to subclass of opt_pass.
9055 (ipa_opt_pass_d::ipa_opt_pass_d): New.
9056 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
9057 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
9058 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
9059 invocation of pass_mode_switching to reflect this moving from a
9060 global to a member of gcc::pass_manager.
9061 (ix86_option_override): Rework how pass_insert_vzeroupper is
9062 added to the pass_manager to reflect autogenerated changes.
9063 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
9065 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
9067 PR rtl-optimization/57708
9068 * recog.c (peep2_find_free_register): Validate all regs in a
9071 2013-08-05 Jan Hubicka <jh@suse.cz>
9074 * cgraph.c (verify_cgraph_node): Accept local flags from other
9076 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
9077 (function_and_variable_visibility): Likewise.
9078 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
9080 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9082 * graph.c (init_graph_slim_pretty_print): Remove.
9083 (print_graph_cfg): Do not call it. Use local pretty printer.
9084 (start_graph_dump): Likewise.
9086 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9088 * gimple-pretty-print.c (buffer): Remove.
9089 (initialized): Likewise.
9090 (maybe_init_pretty_print): Likewise.
9091 (print_gimple_stmt): Do not call it. Use non-static local
9092 pretty_printer variable.
9093 (print_gimple_expr): Likewise.
9094 (print_gimple_seq): Likewise.
9095 (gimple_dump_bb): Likewise.
9097 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9099 * asan.c (asan_pp): Remove.
9100 (asan_pp_initialized): Likewise.
9101 (asan_pp_initialize): Likewise.
9102 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
9103 (asan_emit_stack_protection): Tidy. Use local pretty printer.
9104 (asan_add_global): Likewise.
9106 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
9108 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
9109 * diagnostic.h (diagnostic_flush_buffer): Adjust.
9110 * pretty-print.c (pp_formatted_text_data): Likewise.
9111 (pp_indent): Rename from pp_base_indent.
9112 (pp_format): Rename from pp_base_format.
9113 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
9114 (pp_format_verbatim): Rename from pp_base_format_verbatim.
9115 (pp_flush): Rename from pp_base_flush.
9116 (pp_set_line_maximum_length): Rename from
9117 pp_base_set_line_maximum_length.
9118 (pp_clear_output_area): Rename from pp_base_clear_output_area.
9119 (pp_set_prefix): Rename from pp_base_set_prefix.
9120 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
9121 (pp_emit_prefix): Rename from pp_base_emit_prefix.
9122 (pp_append_text): Rename from pp_base_append_text.
9123 (pp_formatted_text): Rename from pp_base_formatted_text.
9124 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
9125 (pp_remaining_character_count_for_line): Rename from
9126 pp_base_remaining_character_count_for_line.
9127 (pp_newline): Rename from pp_base_newline.
9128 (pp_character): Rename from pp_base_character.
9129 (pp_string): Rename from pp_base_string.
9130 (pp_maybe_space): Rename from pp_base_maybe_space.
9131 * asan.c (asan_pp_string): Adjust.
9132 (asan_emit_stack_protection): Likewise.
9133 (asan_add_global): Likewise.
9134 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
9135 * tree-mudflap.c (mf_varname_tree): Likewise.
9136 * tree-pretty-print.c (pp_tree_identifier): Rename from
9137 pp_base_tree_identifier.
9138 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
9139 Declare as function.
9141 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
9143 * pretty-print.h (pp_bar_bar): New.
9144 (pp_ampersand_ampersand): Likewise.
9145 (pp_less_equal): Likewise.
9146 (pp_greater_equal): Likewise.
9147 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
9148 printer functions instead of pp_string or operators and punctuators.
9149 (dump_gimple_call): Likewise.
9150 (dump_gimple_omp_for): Likewise.
9151 (dump_gimple_transaction): Likewise.
9152 (dump_gimple_phi): Likewise.
9153 (pp_gimple_stmt_1): Likewise.
9154 * sched-vis.c (print_insn): Likewise.
9155 * tree-mudflap.c (mf_varname_tree): Likewise.
9156 * tree-pretty-print.c (dump_block_node): Likewise.
9157 (dump_generic_node): Likewise.
9159 2013-08-02 Jan Hubicka <jh@suse.cz>
9161 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
9163 * lto-streamer-out.c (tree_is_indexable): Results decls and
9164 parm decls are not indexable.
9165 (DFS_write_tree_body): Do not follow args and results.
9166 (hash_tree): Likewise.
9167 (output_functions): Rearrange so struct function is needed
9168 only when real body is output; be able to also ouptut abstract
9169 functions; output DECL_ARGUMENTS and DECL_RESULT.
9170 (lto_output): When not in WPA, ale store abstract functions.
9171 (write_symbol): Do not care about RESULT_DECL.
9172 (output_symbol_p): Handle correctly sbtract decls.
9173 * lto-streamer-in.c (input_function): Rearrange so struct
9174 function can be NULL at entry; allow streaming of
9175 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
9176 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
9177 sanity check during LTO.
9178 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
9179 RESULT_DECl and DECL_ARGUMENTS.
9180 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
9183 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
9185 * pretty-print.h (pp_underscore): New.
9187 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
9188 printer functions instead of pp_character.
9189 (dump_binary_rhs): Likewise.
9190 (dump_ternary_rhs): Likewise.
9191 (dump_gimple_call_args): Likewise.
9192 (pp_points_to_solution): Likewise.
9193 (dump_gimple_call): Likewise.
9194 (dump_gimple_switch): Likewise.
9195 (dump_gimple_cond): Likewise.
9196 (dump_gimple_bind): Likewise.
9197 (dump_gimple_try): Likewise.
9198 (dump_gimple_omp_for): Likewise.
9199 (dump_gimple_omp_continue): Likewise.
9200 (dump_gimple_omp_single): Likewise.
9201 (dump_gimple_omp_sections): Likewise.
9202 (dump_gimple_omp_block): Likewise.
9203 (dump_gimple_omp_critical): Likewise.
9204 (dump_gimple_transaction): Likewise.
9205 (dump_gimple_asm): Likewise.
9206 (dump_gimple_phi): Likewise.
9207 (dump_gimple_omp_parallel): Likewise.
9208 (dump_gimple_omp_task): Likewise.
9209 (dump_gimple_omp_atomic_load): Likewise.
9210 (dump_gimple_omp_atomic_store): Likewise.
9211 (dump_gimple_mem_ops): Likewise.
9212 (pp_gimple_stmt_1): Likewise.
9213 (pp_cfg_jump): Likewise.
9214 (dump_implicit_edges): Likewise.
9215 (gimple_dump_bb_for_graph): Likewise.
9216 * graph.c (draw_cfg_node): Likewise.
9217 * langhooks.c (lhd_print_error_function): Likewise.
9218 * sched-vis.c (print_exp): Likewise.
9219 (print_value): Likewise.
9220 (print_pattern): Likewise.
9221 (print_insn): Likewise.
9222 (rtl_dump_bb_for_graph): Likewise.
9223 * tree-pretty-print.c (dump_function_declaration): Likewise.
9224 (dump_array_domain): Likewise.
9225 (dump_omp_clause): Likewise.
9226 (dump_location): Likewise.
9227 (dump_generic_node): Likewise.
9228 (print_struct_decl): Likewise.
9229 * diagnostic.c (diagnostic_show_locus): Use pp_space.
9231 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9233 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
9234 candidate table when replacing a candidate statement.
9235 (replace_rhs_if_not_dup): Likewise.
9236 (replace_one_candidate): Likewise.
9238 2013-08-02 Jan Hubicka <jh@suse.cz>
9239 Martin Liska <marxin.liska@gmail.com>
9241 * cgraphunit.c (add_new_function): Fix logic when adding from
9243 (assemble_thunk): Rename to ...
9244 (expand_thunk); .. this one; export; get it working with
9245 general functions; make produced gimple valid.
9246 * cgraph.h (expand_thunk): Declare.
9248 2013-08-02 Jan Hubicka <jh@suse.cz>
9250 * ipa-cp.c (gather_context_independent_values): Use
9251 ipa_get_param_move_cost.
9252 (get_replacement_map): Remove PARAM; move parameter folding
9254 (create_specialized_node): Update.
9255 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
9256 assert that we have gimple body; update move_cost.
9257 (count_formal_params): Assert that we have gimple body.
9258 (ipa_dump_param): New function.
9259 (ipa_alloc_node_params): Break out from ...
9260 (ipa_initialize_node_params): ... here.
9261 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
9262 (ipa_write_node_info): Stream move costs.
9263 (ipa_read_node_info): Read move costs.
9264 (ipa_update_after_lto_read): Do not recompute node params.
9265 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
9266 (ipa_get_param): Check we are not in WPA.
9267 (ipa_get_param_move_cost): New.
9268 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
9269 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
9270 parm numbers to be present.
9272 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
9274 PR rtl-optimization/58048
9275 * lra-constraints.c (process_alt_operands): Don't check asm
9276 operand on register.
9278 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
9280 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
9281 the implied StoreLoad barrier for atomic operations if before.
9283 2013-08-02 Jan Hubicka <jh@suse.cz>
9284 Martin Liska <marxin.liska@gmail.com>
9286 * cgraph.c (cgraph_function_body_availability): Do not check
9288 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
9289 symtab_node_availability): Declare.
9290 * ipa.c (can_replace_by_local_alias): New.
9291 (function_and_variable_visibility): Use it.
9292 * symtab.c (symtab_for_node_and_aliases,
9293 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
9295 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
9297 PR rtl-optimization/57963
9298 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
9299 (lra_constraints): Use them.
9301 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
9303 * config/arm/types.md (define_attr "type"): Add "load_acq"
9305 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
9307 (cortex_a53_store1): Likewise.
9309 2013-08-01 Jan Hubicka <jh@suse.cz>
9311 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
9312 partitions are not needed.
9314 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
9316 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
9317 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
9318 MAYBE_NON_Q_CLASS_P where appropriate.
9320 2013-08-01 Jan Hubicka <jh@suse.cz>
9322 * cgraph.h (release_function_body): Declare.
9323 * tree.c (free_lang_data_in_decl): Free, parameters and return values
9324 of unused delcarations.
9326 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9328 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
9329 RTL form when subtracting a constant.
9331 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9333 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
9334 Generate canonical plus rtx with negated immediate instead of minus
9336 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
9338 2013-08-01 Jan Hubicka <jh@suse.cz>
9340 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
9341 (cgraph_release_function_body): Likewise.
9342 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
9343 * cgraph.h (cgrpah_node): Rename abstract_and_needed
9344 to used_as_abstract_origin.
9345 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
9346 symbols used as abstract origins.
9347 * cgraphunit.c (analyze_functions): Update.
9348 * ipa.c (symtab_remove_unreachable_nodes): Recompute
9349 used_as_abstract_origin.
9350 * tree-inline.c (tree_function_versioning): Update
9351 used_as_abstract_origin; be ready for DECL_RESULT and
9352 DECL_ARGUMENTS to be NULL.
9354 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
9355 for abstract functions.
9356 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
9359 2013-08-01 Jan Hubicka <jh@suse.cz>
9361 * profile.c (compute_value_histograms): Fix thinko.
9363 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
9365 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
9366 aarch-common-protos.h to extra_headers.
9367 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
9368 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
9369 * config/aarch64/t-aarch64 (aarch-common.o): Define.
9371 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
9373 * config/aarch64/aarch64.md (define_attr "type"): Delete.
9374 Include "../arm/types.md". Define "type" attribute for all patterns.
9375 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
9378 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
9380 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
9381 to support power8 load fusion.
9382 (fusion_gpr_mem_load): Likewise.
9384 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
9386 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
9387 declarations for power8 load fusion.
9388 (emit_fusion_gpr_load): Likewise.
9390 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
9391 tuning for power8, turn on fusion mode by default. Turn on sign
9392 extending fusion mode if normal fusion mode is on, and we are at
9394 (fusion_gpr_load_p): New function, return true if we can fuse an
9395 addis instruction with a dependent load to a GPR.
9396 (emit_fusion_gpr_load): Emit the instructions for power8 load
9399 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
9400 (VSX load fusion peepholes): New peepholes to fuse together an
9401 addi instruction with a VSX load instruction.
9403 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
9404 peepholes to fuse an addis instruction with a load to a GPR base
9405 register. If we are supporting sign extending fusions, convert
9406 sign extending loads to zero extending loads and add an explicit
9409 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9411 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
9412 aarch-common-protos.h to extra_headers.
9413 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
9414 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
9415 (arm_early_store_addr_dep): Likewise.
9416 (arm_no_early_alu_shift_dep): Likewise.
9417 (arm_no_early_alu_shift_value_dep): Likewise.
9418 (arm_no_early_mul_dep): Likewise.
9419 (arm_no_early_store_addr_dep): Likewise.
9420 (arm_mac_accumulator_is_mul_result): Likewise.
9421 (arm_mac_accumulator_is_result): Likewise.
9422 * config/arm/aarch-common.c: ... here. New file.
9423 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
9425 (arm_early_store_addr_dep): Likewise.
9426 (arm_no_early_alu_shift_dep): Likewise.
9427 (arm_no_early_alu_shift_value_dep): Likewise.
9428 (arm_no_early_mul_dep): Likewise.
9429 (arm_no_early_store_addr_dep): Likewise.
9430 (arm_mac_accumulator_is_mul_result): Likewise.
9431 (arm_mac_accumulator_is_result): Likewise.
9432 * config/arm/aarch-common-protos.h: ... here. New file.
9433 * config/arm/t-arm (aarch-common.o): Define.
9435 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9437 * config/arm/arm.md: Include new file "types.md".
9438 (define_attr "type"): Move from here to ...
9439 (define_attr "mul32"): Likewise.
9440 (define_attr "mul64"): Likewise.
9441 * config/arm/types.md: ... here. New file.
9443 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
9445 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
9446 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
9448 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9450 * gen-pass-instances.awk: Fix offset of substr().
9452 2013-07-31 David Malcolm <dmalcolm@redhat.com>
9454 * Makefile.in (pass-instances.def): New.
9455 (passes.o): Replace dependency on passes.def with one on
9458 * gen-pass-instances.awk: New.
9460 * passes.c (pass_manager::pass_manager): Use pass-instances.def
9461 rather than passes.def, updating local definition of NEXT_PASS
9462 macro to add an extra NUM parameter (currently unused).
9464 2013-07-30 David Malcolm <dmalcolm@redhat.com>
9466 * Makefile.in (PASS_MANAGER_H): New.
9467 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
9468 (passes.o): Likewise.
9469 (statistics.o): Likewise.
9470 (cgraphunit.o): Likewise.
9471 (context.o): Depend on PASS_MANAGER_H.
9473 * pass_manager.h: New.
9475 * cgraphunit.c (cgraph_add_new_function): Update for moves
9476 of globals to fields of pass_manager.
9477 (analyze_function): Likewise.
9478 (expand_function): Likewise.
9479 (ipa_passes): Likewise.
9480 (compile): Likewise.
9482 * context.c (context::context): New.
9483 * context.h (context::context): New.
9484 (context::get_passes): New.
9485 (context::passes_): New.
9487 * lto-cgraph.c (input_node): Update for moves of globals to
9488 fields of pass_manager.
9490 * passes.c (all_passes): Remove, in favor of a field of the
9491 same name within the new class pass_manager.
9492 (all_small_ipa_passes): Likewise.
9493 (all_lowering_passes): Likewise.
9494 (all_regular_ipa_passes): Likewise.
9495 (all_late_ipa_passes): Likewise.
9496 (all_lto_gen_passes): Likewise.
9497 (passes_by_id): Likewise.
9498 (passes_by_id_size): Likewise.
9499 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
9500 the new class pass_manager.
9501 (set_pass_for_id): Convert to...
9502 (pass_manager::set_pass_for_id): ...method.
9503 (get_pass_for_id): Convert to...
9504 (pass_manager::get_pass_for_id): ...method.
9505 (register_one_dump_file): Move body of implementation into...
9506 (pass_manager::register_one_dump_file): ...here.
9507 (register_dump_files_1): Convert to...
9508 (pass_manager::register_dump_files_1): ...method.
9509 (register_dump_files): Convert to...
9510 (pass_manager::register_dump_files): ...method.
9511 (create_pass_tab): Update for moves of globals to fields of
9513 (dump_passes): Move body of implementation into...
9514 (pass_manager::dump_passes): ...here.
9515 (register_pass): Move body of implementation into...
9516 (pass_manager::register_pass): ...here.
9517 (init_optimization_passes): Convert into...
9518 (pass_manager::pass_manager): ...constructor for new
9519 pass_manager class, and initialize the pass_lists array.
9520 (check_profile_consistency): Update for moves of globals to
9521 fields of pass_manager.
9522 (dump_profile_report): Move body of implementation into...
9523 (pass_manager::dump_profile_report): ...here.
9524 (ipa_write_summaries_1): Update for moves of pass lists from
9525 being globals to fields of pass_manager.
9526 (ipa_write_optimization_summaries): Likewise.
9527 (ipa_read_summaries): Likewise.
9528 (ipa_read_optimization_summaries): Likewise.
9529 (execute_all_ipa_stmt_fixups): Likewise.
9531 * statistics.c (statistics_fini): Update for moves of globals to
9532 fields of pass_manager.
9534 * toplev.c (general_init): Replace call to
9535 init_optimization_passes with construction of the pass_manager
9538 * tree-pass.h (all_passes): Remove, in favor of a field of the
9539 same name within the new class pass_manager.
9540 (all_small_ipa_passes): Likewise.
9541 (all_lowering_passes): Likewise.
9542 (all_regular_ipa_passes): Likewise.
9543 (all_lto_gen_passes): Likewise.
9544 (all_late_ipa_passes): Likewise.
9545 (passes_by_id): Likewise.
9546 (passes_by_id_size): Likewise.
9547 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
9548 the new class pass_manager.
9549 (get_pass_for_id): Remove.
9551 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
9553 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
9556 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
9558 * arm.md (mulhi3): New expand pattern.
9560 2013-07-30 Jan Hubicka <jh@suse.cz>
9561 Martin Liska <marxin.liska@gmail.com>
9563 * profile.c (compute_value_histograms): Do not ICE when
9564 there is mismatch only on some counters.
9566 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9568 PR rtl-optimization/57637
9569 * function.c (move_insn_for_shrink_wrap): Also check the
9570 GEN set of the LIVE problem for the liveness analysis
9571 if it exists, otherwise give up.
9573 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9575 PR tree-optimization/57993
9576 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
9577 replaced statement in the candidate table.
9578 (phi_add_costs): Return infinite cost when the hidden basis does
9579 not dominate all phis on which the candidate is dependent.
9580 (replace_one_candidate): Record replaced statement in the
9583 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
9585 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
9586 (ashlv2si3): New expander.
9587 (*ashlv2si3_i): New define_insn_and_split.
9588 * predicates.md (float_operation): Allow patterns with three
9591 PR rtl-optimization/58021
9592 * mode-switching.c (create_pre_exit): Always split off preceding
9593 insns if we are not at the basic block head.
9595 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
9597 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
9598 (UCLIBC_DYNAMIC_LINKER): New macro.
9599 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
9601 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
9602 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
9604 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
9605 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
9606 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
9607 for SF and DF modes. Use ieee_quad_format for TF mode.
9608 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
9609 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
9610 (mips_option_override): Handle `-mnan=legacy'.
9611 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
9612 `-mabs=2008' and `-mnan=2008'.
9613 (OPTION_DEFAULT_SPECS): Add "nan" default.
9614 (ASM_SPEC): Handle `-mnan='.
9615 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
9616 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
9617 comment accordingly.
9618 (neg<mode>2): Likewise.
9619 * config/mips/mips.opt (mabs, mnan): New options.
9620 * doc/install.texi (Configuration): Document `--with-nan=' option.
9621 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
9623 (MIPS Options): Document them.
9624 * config.gcc <mips*-*-*>: Handle `--with-nan='.
9625 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
9626 * configure: Regenerate.
9627 * config.in: Regenerate.
9629 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
9631 * config/i386/i386.md (float post-reload splitters): Do not check
9632 for subregs of SSE registers.
9634 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
9635 H.J. Lu <hongjiu.lu@intel.com>
9639 * config/i386/i386.md (post-reload splitter
9640 to avoid partial SSE reg dependency stalls): New pattern.
9642 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
9644 * config/s390/s390.md ("movcc"): Swap load and store instructions.
9646 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
9648 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
9649 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
9651 2013-07-26 Cary Coutant <ccoutant@google.com>
9653 * dwarf2out.c (die_checksum_ordered): Don't include template
9654 instantiations in signature.
9655 (is_template_parameter): New function.
9656 (is_template_instantiation): New function.
9657 (generate_skeleton_bottom_up): Don't include template instantiations
9659 (generate_skeleton): Likewise.
9660 (break_out_comdat_types): Move recursive call to break out nested
9662 (prune_unused_types_mark_generic_parms_dies): Call
9663 is_template_parameter.
9665 2013-07-26 Ian Bolton <ian.bolton@arm.com>
9667 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
9668 uses vector registers.
9669 * config/aarch64/iterators.md: Add attributes rtn and vas.
9671 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9672 Richard Earnshaw <richard.earnshaw@arm.com>
9674 * combine.c (simplify_comparison): Re-canonicalize operands
9676 * config/arm/arm.md (movcond_addsi): New splitter.
9678 2013-07-25 Sterling Augustine <saugustine@google.com>
9680 * dwarf2out.c (size_of_pubnames): Move code to...
9681 (include_pubname_in_output): ...here. New.
9682 (want_pubnames): Rearrange.
9683 (output_pubnames): Call include_pubname_in_output. Move assertion.
9685 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
9687 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
9689 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
9692 * doc/extend.texi: Remove obsolete builtins. Fix
9693 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
9695 2013-07-25 Jan Hubicka <jh@suse.cz>
9697 * cgraph.c (release_function_body): Break out from ...
9698 (cgraph_release_function_body): ... this one; also release DECL_RESULT
9700 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
9701 old_tree in the map.
9702 (create_specialized_node): Update.
9703 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
9705 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
9706 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
9707 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
9708 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
9711 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9713 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
9714 addsi3_carryin_alt2_<optab>): Correct output template.
9716 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9718 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
9719 Adjust for arm_restrict_it.
9720 Remove trailing whitespace.
9722 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
9724 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
9725 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
9727 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
9729 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
9731 PR rtl-optimization/57960
9732 * lra-constraints.c (process_alt_operands): Use the right mode
9733 when checking strict_low.
9735 2013-07-25 Jan Hubicka <jh@suse.cz>
9737 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
9738 * cgraph.c (cgraph_remove_node): Do not release function body
9739 when in cgraph streaming.
9740 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
9741 in other partitions are not considered reachable; fix handling of
9744 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9746 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
9748 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9753 * config/arm/arm.md ("*sibcall_insn): Replace use of
9754 Ss with US. Adjust output for v5 and v4t.
9755 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
9757 * config/arm/constraints.md ("Ss"): Rename to US.
9759 2013-07-25 Terry Guo <terry.guo@arm.com>
9761 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
9762 shift_add/shift_sub0/shift_sub1 RTXs.
9764 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
9765 Anton Blanchard <anton@au1.ibm.com>
9767 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
9768 (altivec_vpks<VI_char>ss): Likewise.
9769 (altivec_vpks<VI_char>us): Likewise.
9770 (altivec_vpku<VI_char>us): Likewise.
9771 (altivec_vpku<VI_char>um): Likewise.
9773 2013-07-24 David Malcolm <dmalcolm@redhat.com>
9775 Introduce context class.
9777 * Makefile.in (CONTEXT_H): New.
9778 (OBJS): Add context.o.
9779 (toplev.o): Add CONTEXT_H to dependencies.
9782 * toplev.c (general_init): Create the singleton gcc::context instance.
9788 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
9790 PR rtl-optimization/57968
9791 * mode-switching.c (create_pre_exit): Allow instructions that
9792 don't set a return register to need a non-exit mode.
9794 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9795 Anton Blanchard <anton@au1.ibm.com>
9797 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
9798 operands to vperm for little endian.
9799 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
9800 of lvsl to create the control mask for a vperm for little endian.
9802 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9803 Anton Blanchard <anton@au1.ibm.com>
9805 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
9806 two operands for little-endian.
9808 2013-07-23 Steve Ellcey <sellcey@mips.com>
9810 * config/mips/mips.c (mips_case_values_threshold): New.
9811 (TARGET_CASE_VALUES_THRESHOLD): Define.
9813 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9814 Anton Blanchard <anton@au1.ibm.com>
9816 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
9817 selection of field for vector splat in little endian mode.
9819 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9821 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
9822 expanders to rs6000.md.
9823 (ior<mode>3): Likewise.
9824 (and<mode>3): Likewise.
9825 (one_cmpl<mode>2): Likewise.
9826 (nor<mode>3): Likewise.
9827 (andc<mode>3): Likewise.
9828 (eqv<mode>3): Likewise.
9829 (nand<mode>3): Likewise.
9830 (orc<mode>3): Likewise.
9832 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
9835 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
9836 to split multi-word logical operations.
9837 (rs6000_split_logical_di): Likewise.
9838 (rs6000_split_logical): Likewise.
9840 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
9841 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
9842 and allow TImode operations in 32-bit.
9843 (vsx_and<mode>3_64bit): Likewise.
9844 (vsx_ior<mode>3_32bit): Likewise.
9845 (vsx_ior<mode>3_64bit): Likewise.
9846 (vsx_xor<mode>3_32bit): Likewise.
9847 (vsx_xor<mode>3_64bit): Likewise.
9848 (vsx_one_cmpl<mode>2_32bit): Likewise.
9849 (vsx_one_cmpl<mode>2_64bit): Likewise.
9850 (vsx_nor<mode>3_32bit): Likewise.
9851 (vsx_nor<mode>3_64bit): Likewise.
9852 (vsx_andc<mode>3_32bit): Likewise.
9853 (vsx_andc<mode>3_64bit): Likewise.
9854 (vsx_eqv<mode>3_32bit): Likewise.
9855 (vsx_eqv<mode>3_64bit): Likewise.
9856 (vsx_nand<mode>3_32bit): Likewise.
9857 (vsx_nand<mode>3_64bit): Likewise.
9858 (vsx_orc<mode>3_32bit): Likewise.
9859 (vsx_orc<mode>3_64bit): Likewise.
9861 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
9862 logical types in GPRs.
9864 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
9865 logical insns to rs6000.md, and allow TImode operations in
9867 (altivec_ior<mode>3): Likewise.
9868 (altivec_xor<mode>3): Likewise.
9869 (altivec_one_cmpl<mode>2): Likewise.
9870 (altivec_nor<mode>3): Likewise.
9871 (altivec_andc<mode>3): Likewise.
9873 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
9874 attributes for moving the 128-bit logical operations into
9876 (BOOL_REGS_OUTPUT): Likewise.
9877 (BOOL_REGS_OP1): Likewise.
9878 (BOOL_REGS_OP2): Likewise.
9879 (BOOL_REGS_UNARY): Likewise.
9880 (BOOL_REGS_AND_CR0): Likewise.
9881 (one_cmpl<mode>2): Add support for DI logical operations on
9882 32-bit, splitting the operations to 32-bit.
9886 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
9887 changes to combine the 32/64-bit code, allow logical operations on
9888 TI mode in 32-bit, and to use similar match_operator patterns like
9889 scalar mode uses. Combine the Altivec and VSX code for logical
9890 operations, and move it here.
9891 (ior<mode>3, 128-bit types): Likewise.
9892 (xor<mode>3, 128-bit types): Likewise.
9893 (one_cmpl<mode>3, 128-bit types): Likewise.
9894 (nor<mode>3, 128-bit types): Likewise.
9895 (andc<mode>3, 128-bit types): Likewise.
9896 (eqv<mode>3, 128-bit types): Likewise.
9897 (nand<mode>3, 128-bit types): Likewise.
9898 (orc<mode>3, 128-bit types): Likewise.
9899 (and<mode>3_internal): Likewise.
9900 (bool<mode>3_internal): Likewise.
9901 (boolc<mode>3_internal1): Likewise.
9902 (boolc<mode>3_internal2): Likewise.
9903 (boolcc<mode>3_internal1): Likewise.
9904 (boolcc<mode>3_internal2): Likewise.
9905 (eqv<mode>3_internal1): Likewise.
9906 (eqv<mode>3_internal2): Likewise.
9907 (one_cmpl1<mode>3_internal): Likewise.
9909 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9911 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9912 Rename flag_stack_usage to flag_stack_usage_info.
9914 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9916 * config/microblaze/sync.md: New file.
9917 * config/microblaze/microblaze.md: Include sync.md
9918 * config/microblaze/microblaze.c: Add print_operand 'y'.
9919 * config/microblaze/constraints.md: Add memory_contraint
9920 'Q' which is a single register.
9922 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
9924 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
9926 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
9928 * reload.c (find_reloads): Exit loop once we find this operand
9929 cannot be reloaded somehow for this alternative.
9931 * reload.c (find_reloads): Exit loop once we find a hard register.
9933 * rtlanal.c (computed_jump_p): Exit loop once we find label
9936 * i386.c (ix86_pad_returns): Exit loop after setting replace.
9938 * cfgloopmanip.c (remove_path): Exit loop after setting
9941 * gensupport.c (subst_dup): Avoid loop if code is not
9942 MATCH_DUP nor MATCH_OP_DUP.
9944 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
9946 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
9948 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9950 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
9951 true for SP_REGNUM if mode == ptr_mode.
9952 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
9953 with value R0_REGNUM + 31.
9955 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9957 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
9958 pad pointer-typed argument downward.
9960 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9962 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
9963 and __ILP32__ when the ILP32 model is in use.
9965 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9967 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
9968 (aarch64_load_symref_appropriately): In the case of
9969 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
9970 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
9971 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
9972 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
9973 if 'mode' doesn't equal to 'ptr_mode'.
9974 (aarch64_output_mi_thunk): Add an assertion on the alignment of
9975 'vcall_offset'; change to call aarch64_emit_move differently depending
9976 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
9977 to calculate the upper bound of 'vcall_offset'.
9978 (aarch64_cannot_force_const_mem): Change to also return true if
9980 (aarch64_legitimize_reload_address): In the case of large
9981 displacements, add new local variable 'xmode' and an assertion
9982 based on it; change to use 'xmode' to generate the new rtx and
9984 (aarch64_asm_trampoline_template): Change to generate the template
9985 differently depending on TARGET_ILP32 or not; change to use
9986 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
9987 (aarch64_trampoline_size): Removed.
9988 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
9989 and replace immediate literals with it. Change to use 'ptr_mode'
9990 instead of 'DImode' and call convert_memory_address if the mode
9991 of 'fnaddr' doesn't equal to 'ptr_mode'.
9992 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
9994 (aarch64_elf_asm_destructor): Likewise.
9995 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
9996 on TARGET_ILP32 instead of aarch64_trampoline_size.
9997 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
9998 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
9999 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
10000 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
10001 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
10002 (storewb_pair<GPI:mode>_<P:mode>): ... this.
10003 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
10004 depending on the value of 'mode'.
10005 (add_losym_<mode>): New.
10006 (ldr_got_small_<mode>): New, based on ldr_got_small.
10007 (ldr_got_small): Remove.
10008 (ldr_got_small_sidi): New.
10009 * config/aarch64/iterators.md (P): New.
10010 (PTR): Change to 'ptr_mode' in the condition.
10012 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10014 * config.gcc (aarch64*-*-*): Support --with-abi.
10015 (aarch64*-*-elf): Support --with-multilib-list.
10016 (aarch64*-*-linux*): Likewise.
10017 (supported_defaults): Add abi to aarch64*-*-*.
10018 * configure.ac: Mention AArch64 for --with-multilib-list.
10019 * configure: Re-generated.
10020 * config/aarch64/biarchilp32.h: New file.
10021 * config/aarch64/biarchlp64.h: New file.
10022 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
10024 (MULTILIB_DEFAULTS): Ditto.
10025 (DRIVER_SELF_SPECS): Ditto.
10026 (ASM_SPEC): Update to also substitute -mabi.
10027 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
10028 file whose name depends on -mabi= and -mbig-endian.
10029 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
10031 (POINTER_SIZE): New define.
10032 (POINTERS_EXTEND_UNSIGNED): Ditto.
10033 (enum aarch64_abi_type): New enumeration tag.
10034 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
10035 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
10036 (TARGET_ILP32): New define.
10037 * config/aarch64/aarch64.opt (mabi): New.
10038 (aarch64_abi): New.
10039 (ilp32, lp64): New values for -mabi.
10040 * config/aarch64/t-aarch64 (comma): New define.
10041 (MULTILIB_OPTIONS): Ditto.
10042 (MULTILIB_DIRNAMES): Ditto.
10043 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
10044 * doc/invoke.texi: Document -mabi for AArch64.
10046 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
10048 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
10050 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10051 Anton Blanchard <anton@au1.ibm.com>
10053 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
10054 endianness when selecting field to splat.
10056 2013-07-22 Eric Christopher <echristo@gmail.com>
10058 * dwarf2out.c (die_odr_checksum): New function to use
10059 CHECKSUM_ macros and ULEB128 for DIE tag.
10060 (generate_type_signature): Use.
10062 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
10064 * config.gcc (sparc*-*-*): Accept leon3 processor.
10065 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
10066 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
10067 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
10068 * config/sparc/sparc.opt (enum processor_type): Add leon3.
10069 (mfix-ut699): Adjust comment.
10070 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
10071 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
10072 (CPP_CPU_SPEC): Likewise.
10073 (ASM_CPU_SPEC): Likewise.
10074 * config/sparc/sparc.c (leon3_cost): New constant.
10075 (sparc_option_override): Add leon3 support.
10076 (mem_ref): New function.
10077 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
10078 (sparc_do_work_around_errata): Look into the instruction in the delay
10079 slot and adjust accordingly. Add fix for the data cache nullify issues
10080 of the UT699. Change insertion position for the NOP.
10081 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
10082 (leon3_load): New reservation.
10083 (leon_store): Bump latency to 2.
10084 (grfpu): New automaton.
10085 (grfpu_alu): New unit.
10086 (grfpu_ds): Likewise.
10087 (leon_fp_alu): Adjust.
10088 (leon_fp_mult): Delete.
10089 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
10090 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
10091 * config/sparc/sparc.md (cpu): Add leon3.
10092 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
10093 (swapsi): Likewise.
10094 (atomic_test_and_set): Likewise.
10095 (ldstub): Likewise.
10097 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
10099 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
10100 default for R5900 targets.
10101 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
10102 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
10103 * config/mips/mips.c (mips_option_override): Report an error for
10104 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
10105 for -march=r5900 -mhard-float.
10107 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
10109 * df-problems.c (can_move_insns_across): Exit loop once we
10110 find a non-fixed, non-global register.
10112 * ipa-pure-const.c (propagate_nothrow): Exit loop after
10115 * omega.c (omega_eliminate_red): Break after setting red_found.
10116 (omega_problem_has_red_equations): Similarly after setting found.
10117 (omega_query_variable): Similarly after setting coupled.
10119 2013-07-22 Marek Polacek <polacek@redhat.com>
10121 * gimplify.c: Don't include gimple.h twice.
10123 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10125 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
10126 instead of TARGET_THUMB1.
10127 (Pz): New constraint.
10128 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
10130 (compare_negsi_si): Likewise.
10131 (compare_addsi2_op0): Likewise.
10132 (compare_addsi2_op1): Likewise.
10133 (addsi3_carryin_<optab>): Likewise.
10134 (addsi3_carryin_alt2_<optab>): Likewise.
10135 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
10136 for arm_restrict_it.
10137 (subsi3_carryin): Likewise.
10138 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
10139 (minmax_arithsi): Disable for arm_restrict_it.
10140 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
10141 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
10142 (satsi_<SAT:code>_shift): Likewise.
10143 (arm_shiftsi3): Add alternative for 16-bit encoding.
10144 (arm32_movhf): Disable for arm_restrict_it.
10145 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
10146 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
10148 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
10150 * config/arm/arm.md (attribute "insn"): Delete.
10151 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
10152 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
10153 (not_shiftsi): Update for attribute change.
10154 (not_shiftsi_compare0): Likewise.
10155 (not_shiftsi_compare0_scratch): Likewise.
10156 (arm_one_cmplsi2): Likewise.
10157 (thumb1_one_cmplsi2): Likewise.
10158 (notsi_compare0): Likewise.
10159 (notsi_compare0_scratch): Likewise.
10160 (thumb1_movdi_insn): Likewise.
10161 (arm_movsi_insn): Likewise.
10162 (movhi_insn_arch4): Likewise.
10163 (movhi_bytes): Likewise.
10164 (arm_movqi_insn): Likewise.
10165 (thumb1_movqi_insn): Likewise.
10166 (arm32_movhf): Likewise.
10167 (thumb1_movhf): Likewise.
10168 (arm_movsf_soft_insn): Likewise.
10169 (thumb1_movsf_insn): Likewise.
10170 (thumb_movdf_insn): Likewise.
10171 (movsicc_insn): Likewise.
10172 (movsfcc_soft_insn): Likewise.
10173 (and_scc): Likewise.
10174 (cond_move): Likewise.
10175 (if_move_not): Likewise.
10176 (if_not_move): Likewise.
10177 (if_shift_move): Likewise.
10178 (if_move_shift): Likewise.
10179 (if_shift_shift): Likewise.
10180 (if_not_arith): Likewise.
10181 (if_arith_not): Likewise.
10182 (cond_move_not): Likewise.
10183 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
10184 (neon_mov<mode>): Likewise.
10185 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
10186 (thumb2_movsi_vfp): Likewise.
10187 (movsf_vfp): Likewise.
10188 (thumb2_movsf_vfp): Likewise.
10189 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
10191 (cortexa7_older_only): Likewise.
10192 (cortexa7_younger): Likewise.
10193 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
10194 (1020alu_shift_op): Likewise.
10195 (1020alu_shift_reg_op): Likewise.
10196 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
10197 (alu_shift_op): Likewise.
10198 (alu_shift_reg_op): Likewise.
10199 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
10200 (11_alu_shift_op): Likewise.
10201 (11_alu_shift_reg_op): Likewise.
10202 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
10203 (9_alu_shift_reg_op): Likewise.
10204 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10206 (cortex_a15_alu_shift): Likewise.
10207 (cortex_a15_alu_shift_reg): Likewise.
10208 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
10209 (cortex_a5_alu_shift): Likewise.
10210 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
10212 (cortex_a53_alu_shift): Likewise.
10213 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
10215 (cortex_a7_alu_reg): Likewise.
10216 (cortex_a7_alu_shift): Likewise.
10217 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10218 (cortex_a8_alu_shift): Likewise.
10219 (cortex_a8_alu_shift_reg): Likewise.
10220 (cortex_a8_mov): Likewise.
10221 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10222 (cortex_a9_dp_shift): Likewise.
10223 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
10224 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
10225 (cortex_r4_mov): Likewise.
10226 (cortex_r4_alu_shift): Likewise.
10227 (cortex_r4_alu_shift_reg): Likewise.
10228 * config/arm/fa526.md (526_alu_op): Update for attribute change.
10229 (526_alu_shift_op): Likewise.
10230 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
10231 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
10232 (626te_alu_shift_op): Likewise.
10233 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
10234 (726te_alu_op): Likewise.
10235 (726te_alu_shift_op): Likewise.
10236 (726te_alu_shift_reg_op): Likewise.
10237 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
10238 (mp626_alu_shift_op): Likewise.
10239 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
10240 (pj4_alu_e1_conds): Likewise.
10241 (pj4_alu): Likewise.
10242 (pj4_alu_conds): Likewise.
10243 (pj4_shift): Likewise.
10244 (pj4_shift_conds): Likewise.
10245 (pj4_alu_shift): Likewise.
10246 (pj4_alu_shift_conds): Likewise.
10248 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10250 * config/arm/predicates.md (shiftable_operator_strict_it):
10252 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
10253 Disable cond_exec version for arm_restrict_it.
10254 (thumb2_smaxsi3): Convert to generate cond_exec.
10255 (thumb2_sminsi3): Likewise.
10256 (thumb32_umaxsi3): Likewise.
10257 (thumb2_uminsi3): Likewise.
10258 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
10259 (thumb2_neg_abssi2): Likewise.
10260 (thumb2_mov_scc): Add alternative for 16-bit encoding.
10261 (thumb2_movsicc_insn): Adjust alternatives.
10262 (thumb2_mov_negscc): Disable for arm_restrict_it.
10263 (thumb2_mov_negscc_strict_it): New pattern.
10264 (thumb2_mov_notscc_strict_it): New pattern.
10265 (thumb2_mov_notscc): Disable for arm_restrict_it.
10266 (thumb2_ior_scc): Likewise.
10267 (thumb2_ior_scc_strict_it): New pattern.
10268 (thumb2_cond_move): Adjust for arm_restrict_it.
10269 (thumb2_cond_arith): Disable for arm_restrict_it.
10270 (thumb2_cond_arith_strict_it): New pattern.
10271 (thumb2_cond_sub): Adjust for arm_restrict_it.
10272 (thumb2_movcond): Likewise.
10273 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
10274 (thumb2_zero_extendhisi2_v6): Likewise.
10275 (thumb2_zero_extendqisi2_v6): Likewise.
10276 (orsi_notsi_si): Likewise.
10277 (orsi_not_shiftsi_si): Likewise.
10279 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
10281 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
10282 instruction sequence is 1 byte shorter.
10284 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
10286 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
10287 it is not needed after split.
10289 2013-07-20 Iain Sandoe <iain@codesourcery.com>
10292 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
10293 second label for nonlocal goto receivers. Don't output pic base labels
10294 unless we're producing PIC; mark that action unreachable().
10295 (ix86_save_reg): If the function contains a nonlocal label, save the
10297 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
10298 * config/darwin.c (emitted_pic_label_num): New GTY.
10299 (update_pic_label_number_if_needed): New.
10300 (machopic_output_function_base_name): Adjust for nonlocal receiver
10302 (machopic_should_output_picbase_label): New.
10303 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
10304 (nonlocal_goto_receiver): New insn and split.
10306 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
10308 * config/aarch64/aarch64-builtins.c
10309 (aarch64_fold_builtin): Fold abs in all modes.
10310 * config/aarch64/aarch64-simd-builtins.def
10311 (abs): Enable for all modes.
10312 * config/aarch64/arm_neon.h
10313 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
10314 (vabs_f64): Add missing intrinsic.
10316 2013-07-19 Ian Bolton <ian.bolton@arm.com>
10318 * config/aarch64/arm_neon.h (vabs_s64): New function
10320 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
10323 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
10324 * config/avr/avr.md (adjust_len): Add `round'.
10325 * config/avr/avr-protos.h (avr_out_round): New prototype.
10326 (avr_out_plus): Add `out_label' argument.
10327 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
10328 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
10329 Handle the case where `insn' is just a pattern.
10330 (avr_out_bitop): Handle the case where `insn' is just a pattern.
10331 (avr_out_round): New function.
10332 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
10334 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
10336 * config/microblaze/microblaze.c (microblaze_expand_prologue):
10337 Add check for flag_stack_usage to handle -fstack-usage support
10339 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
10341 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
10342 interaction for new Power8 flags and VSX.
10344 2013-07-18 Sriraman Tallam <tmsriram@google.com>
10346 PR middle-end/57698
10347 * tree-inline.c (expand_call_inline): Emit errors during
10348 early_inlining only if optimization is not turned on.
10350 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10354 * passes.c (init_optimization_passes): Move the construction of
10355 the pass hierarchy into a new passes.def file.
10357 * Makefile.in (passes.o): Add dependency on passes.def.
10359 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10361 * passes.c (init_optimization_passes): Introduce macros for
10362 constructing the tree of passes (INSERT_PASSES_AFTER,
10363 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
10364 TERMINATE_PASS_LIST).
10366 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
10367 Wei Mi <wmi@google.com>
10369 PR rtl-optimization/57878
10370 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
10372 (reload_pseudo_compare_func): Check nregs first for reload
10375 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10377 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
10379 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
10381 * read-rtl.c (validate_const_int): Once an invalid character is
10382 seen, quit the loop.
10384 * gengtype.c (write_roots): Similarly once we find the "deletable"
10385 or "if_marked" option.
10387 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
10389 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
10390 "xtab" and "sat". Move value "clz" from here to ...
10391 (attriubte "type"): ... here.
10392 (satsi_<SAT:code>): Delete "insn" attribute.
10393 (satsi_<SAT:code>_shift): Likewise.
10394 (arm_zero_extendqisi2addsi): Likewise.
10395 (arm_extendqisi2addsi): Likewise.
10396 (clzsi2): Update for attribute changes.
10397 (rbitsi2): Likewise.
10398 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
10400 (arm_usatsihi): Likewise.
10401 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10403 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
10405 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
10406 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
10407 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
10408 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
10409 in alphabetical order.
10410 (attribute "core_cycles"): Update for attribute changes.
10411 (arm_addsi3): Likewise.
10412 (addsi3_compare0): Likewise.
10413 (addsi3_compare0_scratch): Likewise.
10414 (addsi3_compare_op1): Likewise.
10415 (addsi3_compare_op2): Likewise.
10416 (compare_addsi2_op0): Likewise.
10417 (compare_addsi2_op1): Likewise.
10418 (addsi3_carryin_shift_<optab>): Likewise.
10419 (subsi3_carryin_shift): Likewise.
10420 (rsbsi3_carryin_shift): Likewise.
10421 (arm_subsi3_insn): Likewise.
10422 (subsi3_compare0): Likewise.
10423 (subsi3_compare): Likewise.
10424 (arm_andsi3_insn): Likewise.
10425 (thumb1_andsi3_insn): Likewise.
10426 (andsi3_compare0): Likewise.
10427 (andsi3_compare0_scratch): Likewise.
10428 (zeroextractsi_compare0_scratch
10429 (andsi_not_shiftsi_si): Likewise.
10430 (iorsi3_insn): Likewise.
10431 (iorsi3_compare0): Likewise.
10432 (iorsi3_compare0_scratch): Likewise.
10433 (arm_xorsi3): Likewise.
10434 (thumb1_xorsi3_insn): Likewise.
10435 (xorsi3_compare0): Likewise.
10436 (xorsi3_compare0_scratch): Likewise.
10437 (satsi_<SAT:code>_shift): Likewise.
10439 (arm_shiftsi3): Likewise.
10440 (shiftsi3_compare0): Likewise.
10441 (not_shiftsi): Likewise.
10442 (not_shiftsi_compare0): Likewise.
10443 (not_shiftsi_compare0_scratch): Likewise.
10444 (arm_one_cmplsi2): Likewise.
10445 (thumb_one_complsi2): Likewise.
10446 (notsi_compare0): Likewise.
10447 (notsi_compare0_scratch): Likewise.
10448 (thumb1_zero_extendhisi2): Likewise.
10449 (arm_zero_extendhisi2): Likewise.
10450 (arm_zero_extendhisi2_v6): Likewise.
10451 (arm_zero_extendhisi2addsi): Likewise.
10452 (thumb1_zero_extendqisi2): Likewise.
10453 (thumb1_zero_extendqisi2_v6): Likewise.
10454 (arm_zero_extendqisi2): Likewise.
10455 (arm_zero_extendqisi2_v6): Likewise.
10456 (arm_zero_extendqisi2addsi): Likewise.
10457 (thumb1_extendhisi2): Likewise.
10458 (arm_extendhisi2): Likewise.
10459 (arm_extendhisi2_v6): Likewise.
10460 (arm_extendqisi): Likewise.
10461 (arm_extendqisi_v6): Likewise.
10462 (arm_extendqisi2addsi): Likewise.
10463 (thumb1_extendqisi2): Likewise.
10464 (thumb1_movdi_insn): Likewise.
10465 (arm_movsi_insn): Likewise.
10466 (movsi_compare0): Likewise.
10467 (movhi_insn_arch4): Likewise.
10468 (movhi_bytes): Likewise.
10469 (arm_movqi_insn): Likewise.
10470 (thumb1_movqi_insn): Likewise.
10471 (arm32_movhf): Likewise.
10472 (thumb1_movhf): Likewise.
10473 (arm_movsf_soft_insn): Likewise.
10474 (thumb1_movsf_insn): Likewise.
10475 (movdf_soft_insn): Likewise.
10476 (thumb_movdf_insn): Likewise.
10477 (arm_cmpsi_insn): Likewise.
10478 (cmpsi_shiftsi): Likewise.
10479 (cmpsi_shiftsi_swp): Likewise.
10480 (arm_cmpsi_negshiftsi_si): Likewise.
10481 (movsicc_insn): Likewise.
10482 (movsfcc_soft_insn): Likewise.
10483 (arith_shiftsi): Likewise.
10484 (arith_shiftsi_compare0
10485 (arith_shiftsi_compare0_scratch
10486 (sub_shiftsi): Likewise.
10487 (sub_shiftsi_compare0
10488 (sub_shiftsi_compare0_scratch
10489 (and_scc): Likewise.
10490 (cond_move): Likewise.
10491 (if_plus_move): Likewise.
10492 (if_move_plus): Likewise.
10493 (if_move_not): Likewise.
10494 (if_not_move): Likewise.
10495 (if_shift_move): Likewise.
10496 (if_move_shift): Likewise.
10497 (if_shift_shift): Likewise.
10498 (if_not_arith): Likewise.
10499 (if_arith_not): Likewise.
10500 (cond_move_not): Likewise.
10501 (thumb1_ashlsi3): Set type attribute.
10502 (thumb1_ashrsi3): Likewise.
10503 (thumb1_lshrsi3): Likewise.
10504 (thumb1_rotrsi3): Likewise.
10505 (shiftsi3_compare0_scratch): Likewise.
10506 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
10507 (neon_mov<mode>): Likewise.
10508 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
10510 (thumb2_movsi_insn): Likewise.
10511 (thumb2_cmpsi_neg_shiftsi): Likewise.
10512 (thumb2_extendqisi_v6): Likewise.
10513 (thumb2_zero_extendhisi2_v6): Likewise.
10514 (thumb2_zero_extendqisi2_v6): Likewise.
10515 (thumb2_shiftsi3_short): Likewise.
10516 (thumb2_addsi3_compare0_scratch): Likewise.
10517 (orsi_not_shiftsi_si): Likewise.
10518 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10519 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
10521 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
10522 (1020alu_shift_op): Likewise.
10523 (1020alu_shift_reg_op): Likewise.
10524 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
10525 (alu_shift_op): Likewise.
10526 (alu_shift_reg_op): Likewise.
10527 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
10528 (11_alu_shift_op): Likewise.
10529 (11_alu_shift_reg_op): Likewise.
10530 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
10531 (9_alu_shift_reg_op): Likewise.
10532 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10534 (cortex_a15_alu_shift): Likewise.
10535 (cortex_a15_alu_shift_reg): Likewise.
10536 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
10538 (cortex_a5_alu_shift): Likewise.
10539 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
10541 (cortex_a53_alu_shift): Likewise.
10542 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
10544 (cortex_a7_alu_reg): Likewise.
10545 (cortex_a7_alu_shift): Likewise.
10546 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
10548 (cortex_a8_alu_shift): Likewise.
10549 (cortex_a8_alu_shift_reg): Likewise.
10550 (cortex_a8_mov): Likewise.
10551 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
10552 (cortex_a9_dp_shift): Likewise.
10553 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
10555 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
10557 (cortex_r4_mov): Likewise.
10558 (cortex_r4_alu_shift): Likewise.
10559 (cortex_r4_alu_shift_reg): Likewise.
10560 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
10561 (526_alu_shift_op): Likewise.
10562 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
10563 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
10564 (626te_alu_shift_op): Likewise.
10565 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
10566 (726te_alu_op): Likewise.
10567 (726te_alu_shift_op): Likewise.
10568 (726te_alu_shift_reg_op): Likewise.
10569 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
10570 (mp626_alu_shift_op): Likewise.
10571 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
10572 (pj4_alu_e1_conds): Likewise.
10573 (pj4_alu): Likewise.
10574 (pj4_alu_conds): Likewise.
10575 (pj4_shift): Likewise.
10576 (pj4_shift_conds): Likewise.
10577 (pj4_alu_shift): Likewise.
10578 (pj4_alu_shift_conds): Likewise.
10579 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
10581 (cortexa7_older_only): Likewise.
10582 (cortexa7_younger): Likewise.
10584 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10586 * ipa-pure-const.c (generate_summary): Rename to...
10587 (pure_const_generate_summary): ... this.
10589 2013-07-17 Iain Sandoe <iain@codesourcery.com>
10591 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
10593 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
10596 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
10599 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10601 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
10602 enabled without -march=zEC12.
10603 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
10606 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10608 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
10609 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
10610 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
10611 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
10612 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
10613 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
10614 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
10616 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
10617 and ISA_HAS_NMADD3_NMSUB3.
10618 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
10619 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
10620 (nmsub4<mode>, nmsub3<mode>): Likewise.
10621 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
10623 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10625 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
10626 TARGET_MIPS5400 checking.
10628 2013-07-16 Jakub Jelinek <jakub@redhat.com>
10629 Peter Bergner <bergner@vnet.ibm.com>
10631 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
10632 registers in the comment.
10633 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
10634 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
10635 rather than FIRST_PSEUDO_REGISTERS.
10637 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
10639 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
10640 enable extra ISA flags with TARGET_HTM.
10642 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10644 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
10647 2013-07-15 Cong Hou <congh@google.com>
10649 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
10650 in compare function for sorting.
10652 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
10654 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
10655 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
10656 * config/rs6000/rs6000.opt: Add -mhtm option.
10657 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
10658 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
10659 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
10660 __HTM__ if the HTM instructions are available.
10661 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
10662 htm_spr_reg_operand): New define_predicates.
10663 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
10664 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
10666 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
10667 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
10668 HTM builtin functions.
10669 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
10670 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
10671 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
10672 (rs6000_builtin_mask_calculate): Likewise.
10673 (rs6000_option_override_internal): Likewise.
10674 (bdesc_htm): Add new HTM builtin support.
10675 (htm_spr_num): New function.
10676 (htm_spr_regno): Likewise.
10677 (rs6000_htm_spr_icode): Likewise.
10678 (htm_expand_builtin): Likewise.
10679 (htm_init_builtins): Likewise.
10680 (rs6000_expand_builtin): Add support for HTM builtin functions.
10681 (rs6000_init_builtins): Likewise.
10682 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
10684 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
10685 (TARGET_HTM, MASK_HTM): Define macros.
10686 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
10687 (FIXED_REGISTERS): Likewise.
10688 (CALL_USED_REGISTERS): Likewise.
10689 (CALL_REALLY_USED_REGISTERS): Likewise.
10690 (REG_ALLOC_ORDER): Likewise.
10691 (enum reg_class): Likewise.
10692 (REG_CLASS_NAMES): Likewise.
10693 (REG_CLASS_CONTENTS): Likewise.
10694 (REGISTER_NAMES): Likewise.
10695 (ADDITIONAL_REGISTER_NAMES): Likewise.
10696 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
10697 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
10698 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
10699 * config/rs6000/htm.md: New file.
10700 * config/rs6000/htmintrin.h: New file.
10701 * config/rs6000/htmxlintrin.h: New file.
10703 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
10705 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
10706 Define SYMBOL_TINY_GOT, update comment.
10707 * config/aarch64/aarch64.c
10708 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
10709 (aarch64_expand_mov_immediate): Likewise.
10710 (aarch64_print_operand): Likewise.
10711 (aarch64_classify_symbol): Likewise.
10712 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
10713 (ldr_got_tiny): Define.
10715 2013-07-13 Tobias Grosser <tobias@grosser.es>
10717 PR tree-optimization/54094
10718 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
10719 scheduling dimension for the parallelism check from the polyhedral
10720 information in the AST.
10721 * graphite-dependences.c (carries_deps): Do not assume the schedule is
10724 2013-07-13 Jason Merrill <jason@redhat.com>
10726 * print-tree.c (debug_vec_tree): Use debug_raw.
10727 (debug_raw (vec<tree, va_gc> &)): New.
10728 (debug_raw (vec<tree, va_gc> *)): New.
10729 * tree.h: Declare them.
10731 2013-07-13 Bin Cheng <bin.cheng@arm.com>
10733 * ifcvt.c (ifcvt_after_combine): New static variable.
10734 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
10736 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
10737 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
10738 rest_of_handle_if_after_reload): Pass new argument for if_convert.
10740 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
10742 * config/mips/mips.c (mips_expand_call): Remove empty statement.
10744 2013-07-12 Michael Matz <matz@suse.de>
10746 PR middle-end/55771
10747 * convert.c (convert_to_real): Reject non-float inner types.
10749 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
10751 * config/aarch64/aarch64-protos.h
10752 (aarch64_simd_immediate_valid_for_move): Remove.
10753 * config/aarch64/aarch64.c (simd_immediate_info): New member.
10754 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
10756 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
10758 2013-07-11 Steve Ellcey <sellcey@mips.com>
10760 * config/mips/mips.c (mips_conditional_register_usage): Do not
10761 use t[0-7] registers in MIPS16 mode when optimizing for size.
10763 2013-07-11 Sriraman Tallam <tmsriram@google.com>
10765 * config/i386/i386.c (dispatch_function_versions): Fix array
10766 indexing of function_version_info to match actual_versions.
10768 2013-07-11 Teresa Johnson <tejohnson@google.com>
10770 * vec.h (struct va_gc): Move release out-of-line.
10771 (va_gc::release): Call ggc_free on released vec.
10773 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10775 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
10776 Require GOT register as additional operand in UNSPEC.
10777 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
10778 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
10779 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
10780 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
10781 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
10782 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
10783 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
10785 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
10788 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
10789 name seen by assembler/linker rather if available.
10791 2013-07-11 Andreas Schwab <schwab@suse.de>
10793 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
10795 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
10797 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
10799 2013-07-10 Joseph Myers <joseph@codesourcery.com>
10801 * doc/tm.texi.in: Move hook documentation to ....
10802 * target.def: ... here.
10804 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
10805 text on @hook line.
10806 * doc/tm.texi: Regenerate.
10808 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
10811 * doc/invoke.texi: Document -Wconditionally-supported.
10813 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10816 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
10819 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10822 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
10823 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
10824 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
10825 Remove duplicate devices.
10826 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
10827 * config/avr/t-multilib: Regenerate.
10828 * config/avr/avr-tables.opt: Regenerate.
10829 * doc/avr-mmcu.texi: Regenerate.
10831 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10834 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
10836 2013-07-10 Graham Stott <graham.stott@btinternet.com>
10838 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
10839 the cost of MULT when optimizing for size.
10841 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10843 * config/cr16/cr16-protos.h: Don't include target.h.
10845 2013-07-09 Joseph Myers <joseph@codesourcery.com>
10847 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
10848 adjust register size for TDmode and TFmode for VSX registers.
10850 2013-07-08 Kai Tietz <ktietz@redhat.com>
10853 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
10854 hook_bool_const_tree_true.
10856 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10858 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
10859 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
10860 * config/s390/s390.md: Define FPR*_REGNUM constants.
10861 Fix FPR2_REGNUM constant (18 -> 17).
10862 ("*trunc<BFP:mode><DFP_ALL:mode>2")
10863 ("*trunc<DFP_ALL:mode><BFP:mode>2")
10864 ("trunc<BFP:mode><DFP_ALL:mode>2")
10865 ("trunc<DFP_ALL:mode><BFP:mode>2")
10866 ("*extend<BFP:mode><DFP_ALL:mode>2")
10867 ("*extend<DFP_ALL:mode><BFP:mode>2")
10868 ("extend<BFP:mode><DFP_ALL:mode>2")
10869 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
10872 2013-07-08 Graham Stott <graham.stott@btinternet.com>
10874 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
10876 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10878 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
10879 and cfun_fpr_bit_p to cfun_fpr_save_p.
10880 (s390_frame_area, s390_register_info, s390_frame_info)
10881 (s390_emit_prologue, s390_emit_epilogue)
10882 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
10884 * config/s390/s390.h: Define *_REGNUM macros for floating point
10887 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
10889 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
10891 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
10893 PR rtl-optimization/57786
10894 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
10895 and break out of the loop when it is set to false.
10897 2013-07-08 Jakub Jelinek <jakub@redhat.com>
10900 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
10901 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
10902 (const_int 63)) 0)).
10903 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
10904 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
10905 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
10907 PR rtl-optimization/57829
10908 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
10909 mask bits outside of mode are just sign-extension from mode to HWI.
10911 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
10913 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
10914 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
10915 adjust_address instead of change_address to keep info about alignment.
10916 (emit_strmov): Remove.
10917 (emit_memmov): New function.
10918 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
10919 (expand_movmem_epilogue): Likewise and return updated rtx for
10921 (expand_constant_movmem_prologue): Likewise and return updated rtx for
10922 destination and source.
10923 (decide_alignment): Refactor, handle vector_loop.
10924 (ix86_expand_movmem): Likewise.
10925 (ix86_expand_setmem): Likewise.
10926 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
10928 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
10930 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
10931 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
10933 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
10935 * config/i386/sse.md (sse_movlhps): Change alternative 3
10936 of operand 2 to "m".
10938 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
10941 * config/i386/sse.md (iptr): New mode attribute.
10942 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
10943 (<sse>_vm<plusminus_insn><mode>3): Ditto.
10944 (<sse>_vmmul<mode>3): Ditto.
10945 (<sse>_vmdiv<mode>3): Ditto.
10946 (sse_vmrcpv4sf2): Ditto.
10947 (<sse>_vmsqrt<mode>2): Ditto.
10948 (sse_vmrsqrtv4sf2): Ditto.
10949 (<sse>_vm<code><mode>3): Ditto.
10950 (avx_vmcmp<mode>3): Ditto.
10951 (<sse>_vmmaskcmp<mode>3): Ditto.
10952 (<sse>_comi): Ditto.
10953 (<sse>_ucomi): Ditto.
10954 (*xop_vmfrcz_<mode>): Ditto.
10955 (*fmai_fmadd_<mode>): Ditto.
10956 (*fmai_fmsub_<mode>): Ditto.
10957 (*fmai_fnmadd_<mode>): Ditto.
10958 (*fmai_fnmsub_<mode>): Ditto.
10959 (*fma4i_vmfmadd_<mode>): Ditto.
10960 (*fma4i_vmfmsub_<mode>): Ditto.
10961 (*fma4i_vmfnmadd_<mode>): Ditto.
10962 (*fma4i_vmfnmsub_<mode>): Ditto.
10963 (*xop_vmfrcz_<mode>): Ditto.
10964 (sse_cvtps2pi): Ditto.
10965 (sse_cvttps2pi): Ditto.
10966 (sse_cvtss2si): Ditto.
10967 (sse_cvtss2si_2): Ditto.
10968 (sse_cvtss2siq_2): Ditto.
10969 (sse_cvttss2si): Ditto.
10970 (sse_cvttss2siq): Ditto.
10971 (sse_cvtsd2si): Ditto.
10972 (sse_cvtsd2si_2): Ditto.
10973 (sse_cvtsd2siq_2): Ditto.
10974 (sse_cvttsd2si): Ditto.
10975 (sse_cvttsd2siq): Ditto.
10976 (sse_cvtsd2ss): Ditto.
10977 (sse_cvtss2sd): Ditto.
10978 (avx2_pbroadcast<mode>): Ditto.
10979 (avx2_pbroadcast<mode>_1): Ditto.
10980 (*avx_vperm_broadcast_v4sf): Ditto.
10982 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
10983 (sse_movlhps): Ditto.
10984 (sse_storehps): Ditto.
10985 (sse_loadhps): Ditto.
10986 (sse_storelps): Ditto.
10987 (sse_loadlps): Ditto.
10988 (*vec_concatv4sf): Ditto.
10989 (*vec_interleave_highv2df): Ditto.
10990 (*vec_interleave_lowv2df): Ditto.
10991 (*vec_extractv2df_1_sse): Ditto.
10992 (*vec_extractv2df_0_sse): Ditto.
10993 (sse2_storelpd): Ditto.
10994 (sse2_loadlpd): Ditto.
10995 (sse2_movsd): Ditto.
10996 (*vec_concatv4si): Ditto.
10997 (vec_concatv2di): Ditto.
10999 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
11000 for Intel asm dialect.
11001 (mmx_punpcklwd): Ditto.
11002 (mmx_punpckldq): Ditto.
11004 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
11005 for intel assembler dialect.
11007 2013-07-06 Jakub Jelinek <jakub@redhat.com>
11010 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
11011 for BUILT_IN_C{LZ,LRSB}*.
11012 * tree.h (CASE_INT_FN): Add FN##IMAX case.
11013 * tree-vrp.c (extract_range_basic): Handle
11014 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
11015 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
11016 fall thru to code calling set_value*.
11017 * builtins.c (expand_builtin): Remove *IMAX cases.
11018 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
11019 if width is bigger than 2*HWI.
11021 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
11023 PR rtl-optimization/55342
11024 * lra-int.h (lra_subreg_reload_pseudos): New.
11025 * lra.c: Add undoing optional reloads to the block diagram.
11026 (lra_subreg_reload_pseudos): New.
11027 (lra_optional_reload_pseudos): Change comments.
11028 (lra): Init and clear lra_subreg_reload_pseudos. Clear
11029 lra_optional_reload_pseudos after undo transformations.
11030 * lra-assigns.c (pseudo_prefix_title): New.
11031 (lra_setup_reg_renumber): Use it.
11032 (spill_for): Ditto. Check subreg reload pseudos too.
11033 (assign_by_spills): Consider subreg reload pseudos too.
11034 * lra-constraints.c (simplify_operand_subreg): Use
11035 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
11036 (curr_insn_transform): Recognize and do optional reloads.
11037 (undo_optional_reloads): New.
11038 (lra_undo_inheritance): Call undo_optional_reloads.
11040 2013-07-05 Thomas Quinot <quinot@adacore.com>
11042 * tree-complex.c (expand_complex_operations_1): Fix typo.
11044 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
11046 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
11047 (tune_params): New member 'const vec_costs'.
11048 * config/aarch64/aarch64.c (generic_vector_cost): New.
11049 (generic_tunings): New member 'generic_vector_cost'.
11050 (aarch64_builtin_vectorization_cost): New.
11051 (aarch64_add_stmt_cost): New.
11052 (TARGET_VECTORIZE_ADD_STMT_COST): New.
11053 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
11055 2013-07-03 Jakub Jelinek <jakub@redhat.com>
11058 * config/i386/predicates.md (vsib_address_operand): Disallow
11059 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
11061 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
11063 PR middle-end/55030
11064 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
11065 expand_builtin_setjmp_receiver.
11066 (expand_label): Adjust, call expand_builtin_setjmp_receiver
11067 with NULL for the label parameter.
11068 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
11069 the frame-pointer. Adjust comments.
11070 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
11071 only if LABEL is non-NULL.
11073 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
11075 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
11076 (ARM_ABI_AAPCS64): Ditto.
11078 (ARM_DEFAULT_ABI): Ditto.
11080 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
11082 * config/aarch64/aarch64-builtins.c
11083 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
11084 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
11086 * config/aarch64/aarch64-simd.md
11087 (aarch64_ld1<VALL:mode>): New.
11088 (aarch64_st1<VALL:mode>): Likewise.
11089 * config/aarch64/arm_neon.h
11090 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
11092 2013-07-02 Sriraman Tallam <tmsriram@google.com>
11094 * config/i386/i386.c (gate_insert_vzeroupper): Check if
11096 (ix86_option_override_internal):Turn on all -mavx target flags by
11097 default as they are dependent on AVX anyway.
11099 2013-07-02 Cary Coutant <ccoutant@google.com>
11101 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
11102 deterministic hash.
11103 (loc_checksum_ordered): Likewise.
11104 (hash_loc_operands): Remove inline keyword.
11106 2013-07-02 Jakub Jelinek <jakub@redhat.com>
11108 PR tree-optimization/57741
11109 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
11110 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
11111 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
11112 Allow REAL_CST step_exprs if flag_associative_math.
11113 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
11115 2013-07-02 Ian Bolton <ian.bolton@arm.com>
11117 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
11119 2013-07-02 Ian Bolton <ian.bolton@arm.com>
11121 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
11123 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11125 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
11127 (iorsi3_insn): Likewise.
11128 (arm_xorsi3): Likewise.
11130 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
11132 * arm.md (attribute "wtype"): Delete. Move attribute values from here
11134 (attribute "type"): ... here, and prefix with "wmmx_".
11135 (attribute "core_cycles"): Update for attribute changes.
11136 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
11137 (tbcstv4hi): Likewise.
11138 (tbcstv2si): Likewise.
11139 (iwmmxt_iordi3): Likewise.
11140 (iwmmxt_xordi3): Likewise.
11141 (iwmmxt_anddi3): Likewise.
11142 (iwmmxt_nanddi3): Likewise.
11143 (iwmmxt_arm_movdi): Likewise.
11144 (iwmmxt_movsi_insn): Likewise.
11145 (mov<mode>_internal): Likewise.
11146 (and<mode>3_iwmmxt): Likewise.
11147 (ior<mode>3_iwmmxt): Likewise.
11148 (xor<mode>3_iwmmxt): Likewise.
11149 (add<mode>3_iwmmxt): Likewise.
11150 (ssaddv8qi3): Likewise.
11151 (ssaddv4hi3): Likewise.
11152 (ssaddv2si3): Likewise.
11153 (usaddv8qi3): Likewise.
11154 (usaddv4hi3): Likewise.
11155 (usaddv2si3): Likewise.
11156 (sub<mode>3_iwmmxt): Likewise.
11157 (sssubv8qi3): Likewise.
11158 (sssubv4hi3): Likewise.
11159 (sssubv2si3): Likewise.
11160 (ussubv8qi3): Likewise.
11161 (ussubv4hi3): Likewise.
11162 (ussubv2si3): Likewise.
11163 (mulv4hi3_iwmmxt): Likewise.
11164 (smulv4hi3_highpart): Likewise.
11165 (umulv4hi3_highpart): Likewise.
11166 (iwmmxt_wmacs): Likewise.
11167 (iwmmxt_wmacsz): Likewise.
11168 (iwmmxt_wmacu): Likewise.
11169 (iwmmxt_wmacuz): Likewise.
11170 (iwmmxt_clrdi): Likewise.
11171 (iwmmxt_clrv8qi): Likewise.
11172 (iwmmxt_clr4hi): Likewise.
11173 (iwmmxt_clr2si): Likewise.
11174 (iwmmxt_uavgrndv8qi3): Likewise.
11175 (iwmmxt_uavgrndv4hi3): Likewise.
11176 (iwmmxt_uavgv8qi3): Likewise.
11177 (iwmmxt_uavgv4hi3): Likewise.
11178 (iwmmxt_tinsrb): Likewise.
11179 (iwmmxt_tinsrh): Likewise.
11180 (iwmmxt_tinsrw): Likewise.
11181 (iwmmxt_textrmub): Likewise.
11182 (iwmmxt_textrmsb): Likewise.
11183 (iwmmxt_textrmuh): Likewise.
11184 (iwmmxt_textrmsh): Likewise.
11185 (iwmmxt_textrmw): Likewise.
11186 (iwmxxt_wshufh): Likewise.
11187 (eqv8qi3): Likewise.
11188 (eqv4hi3): Likewise.
11189 (eqv2si3): Likewise.
11190 (gtuv8qi3): Likewise.
11191 (gtuv4hi3): Likewise.
11192 (gtuv2si3): Likewise.
11193 (gtv8qi3): Likewise.
11194 (gtv4hi3): Likewise.
11195 (gtv2si3): Likewise.
11196 (smax<mode>3_iwmmxt): Likewise.
11197 (umax<mode>3_iwmmxt): Likewise.
11198 (smin<mode>3_iwmmxt): Likewise.
11199 (umin<mode>3_iwmmxt): Likewise.
11200 (iwmmxt_wpackhss): Likewise.
11201 (iwmmxt_wpackwss): Likewise.
11202 (iwmmxt_wpackdss): Likewise.
11203 (iwmmxt_wpackhus): Likewise.
11204 (iwmmxt_wpackwus): Likewise.
11205 (iwmmxt_wpackdus): Likewise.
11206 (iwmmxt_wunpckihb): Likewise.
11207 (iwmmxt_wunpckihh): Likewise.
11208 (iwmmxt_wunpckihw): Likewise.
11209 (iwmmxt_wunpckilb): Likewise.
11210 (iwmmxt_wunpckilh): Likewise.
11211 (iwmmxt_wunpckilw): Likewise.
11212 (iwmmxt_wunpckehub): Likewise.
11213 (iwmmxt_wunpckehuh): Likewise.
11214 (iwmmxt_wunpckehuw): Likewise.
11215 (iwmmxt_wunpckehsb): Likewise.
11216 (iwmmxt_wunpckehsh): Likewise.
11217 (iwmmxt_wunpckehsw): Likewise.
11218 (iwmmxt_wunpckelub): Likewise.
11219 (iwmmxt_wunpckeluh): Likewise.
11220 (iwmmxt_wunpckeluw): Likewise.
11221 (iwmmxt_wunpckelsb): Likewise.
11222 (iwmmxt_wunpckelsh): Likewise.
11223 (iwmmxt_wunpckelsw): Likewise.
11224 (ror<mode>3): Likewise.
11225 (ashr<mode>3_iwmmxt): Likewise.
11226 (lshr<mode>3_iwmmxt): Likewise.
11227 (ashl<mode>3_iwmmxt): Likewise.
11228 (ror<mode>3_di): Likewise.
11229 (ashr<mode>3_di): Likewise.
11230 (lshr<mode>3_di): Likewise.
11231 (ashl<mode>3_di): Likewise.
11232 (iwmmxt_wmadds): Likewise.
11233 (iwmmxt_wmaddu): Likewise.
11234 (iwmmxt_tmia): Likewise.
11235 (iwmmxt_tmiaph): Likewise.
11236 (iwmmxt_tmiabb): Likewise.
11237 (iwmmxt_tmiatb): Likewise.
11238 (iwmmxt_tmiabt): Likewise.
11239 (iwmmxt_tmiatt): Likewise.
11240 (iwmmxt_tmovmskb): Likewise.
11241 (iwmmxt_tmovmskh): Likewise.
11242 (iwmmxt_tmovmskw): Likewise.
11243 (iwmmxt_waccb): Likewise.
11244 (iwmmxt_wacch): Likewise.
11245 (iwmmxt_waccw): Likewise.
11246 (iwmmxt_waligni): Likewise.
11247 (iwmmxt_walignr): Likewise.
11248 (iwmmxt_walignr0): Likewise.
11249 (iwmmxt_walignr1): Likewise.
11250 (iwmmxt_walignr2): Likewise.
11251 (iwmmxt_walignr3): Likewise.
11252 (iwmmxt_wsadb): Likewise.
11253 (iwmmxt_wsadh): Likewise.
11254 (iwmmxt_wsadbz): Likewise.
11255 (iwmmxt_wsadhz): Likewise.
11256 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
11257 (iwmmxt_wabsdiffb): Likewise.
11258 (iwmmxt_wabsdiffh): Likewise.
11259 (iwmmxt_wabsdiffw): Likewise.
11260 (iwmmxt_waddsubhx): Likewise
11261 (iwmmxt_wsubaddhx): Likewise.
11262 (addc<mode>3): Likewise.
11263 (iwmmxt_avg4): Likewise.
11264 (iwmmxt_avg4r): Likewise.
11265 (iwmmxt_wmaddsx): Likewise.
11266 (iwmmxt_wmaddux): Likewise.
11267 (iwmmxt_wmaddsn): Likewise.
11268 (iwmmxt_wmaddun): Likewise.
11269 (iwmmxt_wmulwsm): Likewise.
11270 (iwmmxt_wmulwum): Likewise.
11271 (iwmmxt_wmulsmr): Likewise.
11272 (iwmmxt_wmulumr): Likewise.
11273 (iwmmxt_wmulwsmr): Likewise.
11274 (iwmmxt_wmulwumr): Likewise.
11275 (iwmmxt_wmulwl): Likewise.
11276 (iwmmxt_wqmulm): Likewise.
11277 (iwmmxt_wqmulwm): Likewise.
11278 (iwmmxt_wqmulmr): Likewise.
11279 (iwmmxt_wqmulwmr): Likewise.
11280 (iwmmxt_waddbhusm): Likewise.
11281 (iwmmxt_waddbhusl): Likewise.
11282 (iwmmxt_wqmiabb): Likewise.
11283 (iwmmxt_wqmiabt): Likewise.
11284 (iwmmxt_wqmiatb): Likewise.
11285 (iwmmxt_wqmiatt): Likewise.
11286 (iwmmxt_wqmiabbn): Likewise.
11287 (iwmmxt_wqmiabtn): Likewise.
11288 (iwmmxt_wqmiatbn): Likewise.
11289 (iwmmxt_wqmiattn): Likewise.
11290 (iwmmxt_wmiabb): Likewise.
11291 (iwmmxt_wmiabt): Likewise.
11292 (iwmmxt_wmiatb): Likewise.
11293 (iwmmxt_wmiatt): Likewise.
11294 (iwmmxt_wmiabbn): Likewise.
11295 (iwmmxt_wmiabtn): Likewise.
11296 (iwmmxt_wmiatbn): Likewise.
11297 (iwmmxt_wmiattn): Likewise.
11298 (iwmmxt_wmiawbb): Likewise.
11299 (iwmmxt_wmiawbt): Likewise.
11300 (iwmmxt_wmiawtb): Likewise.
11301 (iwmmxt_wmiawtt): Likewise.
11302 (iwmmxt_wmiawbbn): Likewise.
11303 (iwmmxt_wmiawbtn): Likewise.
11304 (iwmmxt_wmiawtbn): Likewise.
11305 (iwmmxt_wmiawttn): Likewise.
11306 (iwmmxt_wmerge): Likewise.
11307 (iwmmxt_tandc<mode>3): Likewise.
11308 (iwmmxt_torc<mode>3): Likewise.
11309 (iwmmxt_torvsc<mode>3): Likewise.
11310 (iwmmxt_textrc<mode>3): Likewise.
11311 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
11312 (wmmxt_pack): Likewise.
11313 (wmmxt_mult_c1): Likewise.
11314 (wmmxt_mult_c2): Likewise.
11315 (wmmxt_alu_c1): Likewise.
11316 (wmmxt_alu_c2): Likewise.
11317 (wmmxt_alu_c3): Likewise.
11318 (wmmxt_transfer_c1): Likewise.
11319 (wmmxt_transfer_c2): Likewise.
11320 (wmmxt_transfer_c3): Likewise.
11321 (marvell_f_iwmmxt_wstr): Likewise.
11322 (marvell_f_iwmmxt_wldr): Likewise.
11324 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
11326 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
11328 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11331 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11332 * lra-constraints.c (need_for_split_p): Check call used hard regs
11333 living through calls.
11335 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
11336 call used regs for call insn.
11338 2013-06-28 Jakub Jelinek <jakub@redhat.com>
11341 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
11342 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
11344 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
11346 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
11348 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11350 * lra-constraints.c (need_for_split_p): Check call used hard regs
11351 living through calls.
11353 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
11356 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
11357 to tie with any other modes. Eliminate Altivec vector mode tests,
11358 since these are a subset of ALTIVEC or VSX vector modes. Simplify
11359 code, to return 0 if testing MODE2 for a condition, if we've
11360 already tested MODE1 for the same condition.
11362 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11364 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
11367 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11369 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
11370 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
11372 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11374 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
11376 (aarch64_symbolic_constant_p): Remove.
11377 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
11378 static. Fix line length and white space.
11379 (aarch64_symbolic_constant_p): Remove.
11380 * config/aarch64/predicates.md (aarch64_valid_symref):
11381 Use aarch64_classify_symbol_expression.
11383 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11385 * config/arm/constraints.md (Ts): New constraint.
11386 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
11388 (compare_scc): Use "Ts" constraint for operand 0.
11389 (ior_scc_scc): Likewise.
11390 (and_scc_scc): Likewise.
11391 (and_scc_scc_nodom): Likewise.
11392 (ior_scc_scc_cmp): Likewise for operand 7.
11393 (and_scc_scc_cmp): Likewise.
11394 * config/arm/thumb2.md (thumb2_movsi_insn):
11395 Add alternatives for 16-bit encodings.
11396 (thumb2_movhi_insn): Likewise.
11397 (thumb2_movsicc_insn): Likewise.
11398 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
11399 (thumb2_negscc): Use "Ts" constraint.
11400 Move mvn instruction outside cond_exec block.
11401 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
11402 for 16-bit encodings.
11404 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11406 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
11408 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
11409 (mulsi3subsi): Likewise.
11410 (mulsidi3adddi): Likewise.
11411 (mulsidi3_v6): Likewise.
11412 (umulsidi3_v6): Likewise.
11413 (umulsidi3adddi_v6): Likewise.
11414 (smulsi3_highpart_v6): Likewise.
11415 (umulsi3_highpart_v6): Likewise.
11416 (mulhisi3tb): Likewise.
11417 (mulhisi3bt): Likewise.
11418 (mulhisi3tt): Likewise.
11419 (maddhisi4): Likewise.
11420 (maddhisi4tb): Likewise.
11421 (maddhisi4tt): Likewise.
11422 (maddhidi4): Likewise.
11423 (maddhidi4tb): Likewise.
11424 (maddhidi4tt): Likewise.
11425 (zeroextractsi_compare0_scratch): Likewise.
11426 (insv_zero): Likewise.
11427 (insv_t2): Likewise.
11428 (anddi_notzesidi_di): Likewise.
11429 (anddi_notsesidi_di): Likewise.
11430 (andsi_notsi_si): Likewise.
11431 (iordi_zesidi_di): Likewise.
11432 (xordi_zesidi_di): Likewise.
11433 (andsi_iorsi3_notsi): Likewise.
11434 (smax_0): Likewise.
11435 (smax_m1): Likewise.
11436 (smin_0): Likewise.
11437 (not_shiftsi): Likewise.
11438 (unaligned_loadsi): Likewise.
11439 (unaligned_loadhis): Likewise.
11440 (unaligned_loadhiu): Likewise.
11441 (unaligned_storesi): Likewise.
11442 (unaligned_storehi): Likewise.
11443 (extv_reg): Likewise.
11444 (extzv_t2): Likewise.
11445 (divsi3): Likewise.
11446 (udivsi3): Likewise.
11447 (arm_zero_extendhisi2addsi): Likewise.
11448 (arm_zero_extendqisi2addsi): Likewise.
11449 (compareqi_eq0): Likewise.
11450 (arm_extendhisi2_v6): Likewise.
11451 (arm_extendqisi2addsi): Likewise.
11452 (arm_movt): Likewise.
11453 (thumb2_ldrd): Likewise.
11454 (thumb2_ldrd_base): Likewise.
11455 (thumb2_ldrd_base_neg): Likewise.
11456 (thumb2_strd): Likewise.
11457 (thumb2_strd_base): Likewise.
11458 (thumb2_strd_base_neg): Likewise.
11459 (arm_negsi2): Add alternative for 16-bit encoding.
11460 (arm_one_cmplsi2): Likewise.
11462 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11464 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
11465 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
11466 (movdfcc): Likewise.
11467 * config/arm/vfp.md (*thumb2_movsf_vfp):
11468 Disable predication for arm_restrict_it.
11469 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
11470 (*thumb2_movdfcc_vfp): Likewise.
11471 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
11472 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
11473 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
11474 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
11475 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
11476 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
11477 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
11478 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
11479 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
11480 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
11481 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
11482 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
11483 Disable predication for arm_restrict_it.
11485 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
11487 * config/i386/bmiintrin.h (_bextr_u32): New.
11488 (_bextr_u64): Ditto.
11490 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
11492 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
11493 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
11494 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
11495 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
11496 * config/mips/n32-elf.h: ...this new file.
11498 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
11501 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
11502 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
11504 2013-06-27 Catherine Moore <clm@codesourcery.com>
11506 * config/mips/mips-tables.opt: Regenerate.
11507 * config/mips/mips-cpus.def: Add m14ke and m14kec.
11508 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
11509 * doc/invoke.texi: Add -m14kc.
11511 2013-06-27 Jakub Jelinek <jakub@redhat.com>
11514 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
11515 constraints of operand 1 and 2.
11518 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
11519 to match RTL canonicalization. Swap predicates and
11520 constraints of operand 1 and 2.
11522 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
11524 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
11525 Process OP_INOUT regs for splitting too.
11527 2013-06-27 Jakub Jelinek <jakub@redhat.com>
11529 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
11530 decl before the loop, initialize to NULL.
11531 (vectorizable_load): Initialize ptr_incr to NULL.
11533 2013-06-27 Martin Jambor <mjambor@suse.cz>
11536 * ipa-ref.h (ipa_maybe_record_reference): Declare.
11537 * ipa-ref.c (ipa_maybe_record_reference): New function.
11538 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
11539 * ipa-cp.c (create_specialized_node): Record potential references from
11541 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
11543 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
11545 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
11546 parameter 'mode' of type 'enum machine_mode mode'; change to pass
11547 'mode' to force_reg.
11548 (aarch64_add_offset): Update calls to aarch64_force_temporary.
11549 (aarch64_expand_mov_immediate): Likewise.
11551 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
11553 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
11554 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
11556 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11558 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
11559 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
11560 (struct machine_function): Add tbegin_p.
11561 (s390_canonicalize_comparison): Fold CC mode compares to
11562 conditional jump if possible.
11563 (s390_emit_jump): Return the emitted jump.
11564 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
11565 Handle CCRAWmode compares.
11566 (s390_option_override): Default to -mhtm if available.
11567 (s390_reg_clobbered_rtx): Handle floating point regs as well.
11568 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
11569 FPRs instead of df_regs_ever_live_p.
11570 (s390_optimize_nonescaping_tx): New function.
11571 (s390_init_frame_layout): Extend clobbered_regs array to cover
11573 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
11574 (s390_expand_tbegin): New function.
11575 (enum s390_builtin): New enum definition.
11576 (code_for_builtin): New array definition.
11577 (s390_init_builtins): New function.
11578 (s390_expand_builtin): New function.
11579 (TARGET_INIT_BUILTINS): Define.
11580 (TARGET_EXPAND_BUILTIN): Define.
11581 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
11582 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
11583 (s390_alc_comparison): Likewise.
11584 * config/s390/s390-modes.def: Add CCRAWmode.
11585 * config/s390/s390.h (processor_flags): Add PF_TX.
11586 (TARGET_CPU_HTM): Define macro.
11587 (TARGET_HTM): Define macro.
11588 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
11589 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
11590 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
11591 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
11592 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
11593 (TBEGIN_MASK, TBEGINC_MASK): New constants.
11594 ("*cc_to_int"): Move up.
11595 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
11596 constants other than 0.
11597 ("*ccraw_to_int"): New insn and splitter definition.
11598 ("tbegin", "tbegin_nofloat", "tbegin_retry")
11599 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
11600 ("tx_assist"): New expander.
11601 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
11602 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
11603 * config/s390/s390.opt: Add -mhtm option.
11604 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
11605 * config/s390/htmxlintrin.h: New file.
11606 * config/s390/htmintrin.h: New file.
11607 * config/s390/s390intrin.h: New file.
11608 * doc/extend.texi: Document htm builtins.
11609 * config.gcc: Add the new header files to extra_headers.
11611 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
11613 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
11614 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
11616 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
11617 Pat Haugen <pthaugen@us.ibm.com>
11618 Peter Bergner <bergner@vnet.ibm.com>
11620 * config/rs6000/power8.md: New.
11621 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
11622 setting for power8 entry.
11623 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
11624 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
11625 test for Power4/Power5 only.
11626 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
11628 (force_new_group): Adjust comment.
11629 * config/rs6000/rs6000.md: Include power8.md.
11631 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
11633 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
11634 * config/arm/arm-protos.h (arm_max_conditional_execute): New
11636 (tune_params): Update comment.
11637 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
11638 (arm_max_conditional_execute): New function.
11639 (thumb2_final_prescan_insn): Use max_insn_skipped and
11640 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
11642 2013-06-25 Jakub Jelinek <jakub@redhat.com>
11644 PR tree-optimization/57705
11645 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
11646 SSA_NAME step, provided that it is not defined inside the loop.
11647 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
11648 (get_initial_def_for_induction): Handle SSA_NAME IV step.
11650 2013-06-25 Martin Jambor <mjambor@suse.cz>
11652 PR middle-end/57670
11653 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
11654 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
11656 (ipa_note_param_call): Initialize member_ptr flag.
11657 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
11658 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
11659 (ipa_write_indirect_edge_info): Stream member_ptr flag.
11660 (ipa_read_indirect_edge_info): Likewise.
11662 2013-06-25 Richard Biener <rguenther@suse.de>
11664 PR middle-end/56977
11665 * passes.c (init_optimization_passes): Move pass_fold_builtins
11666 and pass_dce earlier with -Og.
11668 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
11670 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
11671 <BIT_FIELD_REF>: Remove trailing TAB.
11672 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
11675 2013-06-24 Martin Jambor <mjambor@suse.cz>
11677 PR tree-optimization/57358
11678 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
11679 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
11680 (ipa_analyze_params_uses): Generate pessimistic info when true.
11682 2013-06-24 Martin Jambor <mjambor@suse.cz>
11684 PR tree-optimization/57539
11685 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
11686 global.inlined_to of the new node to it. All callers changed.
11687 * ipa-inline-transform.c (clone_inlined_nodes): New variable
11688 inlining_into, pass it to cgraph_clone_node.
11689 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
11690 ipa_free_edge_args_substructures.
11691 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
11692 rdesc linked list. Do not assert rdesc edges have inlined caller.
11693 Assert we have found an rdesc in the rdesc list.
11695 2013-06-24 Richard Biener <rguenther@suse.de>
11697 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
11698 (pointer_set_lookup): Declare.
11699 (class pointer_map): New template class implementing a
11700 generic pointer to T map.
11701 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
11702 pointer_map<T>::contains, pointer_map<T>::insert,
11703 pointer_map<T>::traverse): New functions.
11704 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
11705 (pointer_set_lookup): New function.
11706 (pointer_set_contains): Use pointer_set_lookup.
11707 (pointer_set_insert): Likewise.
11708 (insert_aux): Remove.
11709 (struct pointer_map_t): Embed a pointer_set_t.
11710 (pointer_map_create): Adjust.
11711 (pointer_map_destroy): Likewise.
11712 (pointer_map_contains): Likewise.
11713 (pointer_map_insert): Likewise.
11714 (pointer_map_traverse): Likewise.
11715 * tree-streamer.h (struct streamer_tree_cache_d): Use a
11716 pointer_map<unsigned> instead of a pointer_map_t.
11717 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
11718 (streamer_tree_cache_lookup): Likewise.
11719 (streamer_tree_cache_create): Likewise.
11720 (streamer_tree_cache_delete): Likewise.
11721 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
11722 pointer_map<unsigned> instead of a pointer_map_t.
11723 (lto_init_tree_ref_encoder): Adjust.
11724 (lto_destroy_tree_ref_encoder): Likewise.
11725 * lto-section-out.c (lto_output_decl_index): Likewise.
11726 (lto_record_function_out_decl_state): Likewise.
11727 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
11729 2013-06-24 Richard Biener <rguenther@suse.de>
11731 PR tree-optimization/57488
11732 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
11734 2013-06-24 Alan Modra <amodra@gmail.com>
11736 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
11737 (gen_easy_altivec_constant): Likewise.
11738 * config/rs6000/predicates.md (easy_vector_constant_add_self,
11739 easy_vector_constant_msb): Likewise.
11741 2013-06-23 Jakub Jelinek <jakub@redhat.com>
11744 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
11745 add missing return true.
11747 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
11750 * config/sh/predicates.md (general_extend_operand): Invoke
11751 general_movsrc_operand for memory operands.
11752 (general_movsrc_operand): Allow reg+reg addressing, do not use
11753 general_operand for memory operands.
11755 2013-06-23 Sriraman Tallam <tmsriram@google.com>
11757 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
11758 when current target options does not apply.
11759 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
11760 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
11761 * config/i386/bmiintrin.h: Pass appropriate target
11762 attributes to header.
11763 * config/i386/mmintrin.h: Ditto.
11764 * config/i386/nmmintrin.h: Ditto.
11765 * config/i386/avx2intrin.h: Ditto.
11766 * config/i386/fxsrintrin.h: Ditto.
11767 * config/i386/tbmintrin.h: Ditto.
11768 * config/i386/xsaveintrin.h: Ditto.
11769 * config/i386/f16cintrin.h: Ditto.
11770 * config/i386/xtestintrin.h: Ditto.
11771 * config/i386/xsaveoptintrin.h: Ditto.
11772 * config/i386/bmi2intrin.h: Ditto.
11773 * config/i386/lzcntintrin.h: Ditto.
11774 * config/i386/smmintrin.h: Ditto.
11775 * config/i386/wmmintrin.h: Ditto.
11776 * config/i386/x86intrin.h: Remove all header include guards.
11777 * config/i386/prfchwintrin.h: Ditto.
11778 * config/i386/pmmintrin.h: Ditto.
11779 * config/i386/tmmintrin.h: Ditto.
11780 * config/i386/xmmintrin.h: Ditto.
11781 * config/i386/popcntintrin.h: Ditto.
11782 * config/i386/rdseedintrin.h: Ditto.
11783 * config/i386/ammintrin.h: Ditto.
11784 * config/i386/emmintrin.h: Ditto.
11785 * config/i386/immintrin.h: Remove all header include guards.
11786 * config/i386/fma4intrin.h: Ditto.
11787 * config/i386/lwpintrin.h: Ditto.
11788 * config/i386/xopintrin.h: Ditto.
11789 * config/i386/ia32intrin.h: Ditto.
11790 * config/i386/avxintrin.h: Ditto.
11791 * config/i386/rtmintrin.h: Ditto.
11792 * config/i386/fmaintrin.h: Ditto.
11793 * config/i386/mm3dnow.h: Ditto.
11795 2013-06-22 Sriraman Tallam <tmsriram@google.com>
11797 * common/config/i386/i386-common.c: Handle LZCNT.
11799 2013-06-22 Andi Kleen <ak@linux.intel.com>
11801 * doc/extend.texi: Use __atomic_store_n instead of
11802 __atomic_store in HLE example.
11804 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
11806 * config/sh/sh.c: Remove <cstdlib> workaround.
11808 2013-06-21 Andi Kleen <ak@linux.intel.com>
11810 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
11812 2013-06-21 Andi Kleen <ak@linux.intel.com>
11814 * doc/extend.texi: Document that __atomic_clear and
11815 __atomic_test_and_set should only be used with bool.
11817 2013-06-20 Jan Hubicka <jh@suse.cz>
11819 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
11820 types_same_for_odr.
11821 * tree.c (decls_same_for_odr): New function.
11822 (same_for_edr): New function.
11823 (types_same_for_odr): New function.
11824 (get_binfo_at_offset): Use it.
11825 * tree.h (types_same_for_odr): Declare.
11827 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
11828 Jason Merrill <jason@redhat.com>
11830 * system.h: Include <cstdlib> as well as <stdlib.h>.
11832 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
11835 * config/i386/i386.c (construct_container): Report error if
11836 long double is used with disabled x87 float returns.
11838 2013-06-20 Jan Hubicka <jh@suse.cz>
11840 * lto-cgraph.c (input_symtab): Do not set cgraph state.
11842 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
11844 PR rtl-optimization/57425
11845 PR rtl-optimization/57569
11846 * alias.c (write_dependence_p): Remove parameters mem_mode and
11847 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
11848 Changed all callers.
11849 (canon_anti_dependence): Get comments and semantics in sync.
11850 Add parameter mem_canonicalized. Changed all callers.
11851 * rtl.h (canon_anti_dependence): Update prototype.
11853 2013-06-20 Richard Biener <rguenther@suse.de>
11855 * data-streamer-in.c (streamer_read_uhwi): Optimize single
11856 byte case, inline streamer_read_uchar and defer section
11859 2013-06-20 Richard Biener <rguenther@suse.de>
11861 PR tree-optimization/57584
11862 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
11863 SSA names into the expanded expression that take part in
11864 abnormal coalescing.
11866 2013-06-19 Sharad Singhai <singhai@google.com>
11868 * gcov.c (print_usage): Handle new option.
11869 (process_args): Ditto.
11870 (get_gcov_intermediate_filename): New function.
11871 (output_intermediate_file): New function.
11872 (output_gcov_file): New function
11873 (generate_results): Handle new option.
11874 (release_function): Relase demangled name.
11875 (read_graph_file): Handle demangled name.
11876 (output_lines): Ditto.
11877 * doc/gcov.texi: Document gcov intermediate format.
11879 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
11882 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
11883 (lra_emit_add): Use the functions. Add comment about Y as an
11886 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
11889 * collect2.c (collect_atexit): New.
11890 (collect_exit): Delete.
11891 (main): Register collect_atexit with atexit.
11892 (collect_wait): Change collect_exit to exit.
11894 * collect2.h (collect_exit): Delete.
11895 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
11897 2013-06-19 Wei Mi <wmi@google.com>
11899 PR rtl-optimization/57518
11900 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
11901 if regno is used in paradoxical subreg.
11902 (update_equiv_regs): Check pdx_subregs[regno] before
11903 set a reg to be equivalent with a mem.
11905 2013-06-19 Matthias Klose <doko@ubuntu.com>
11908 * file-find.h (find_a_file): Add a mode parameter.
11909 * file-find.c (find_a_file): Likewise.
11910 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
11911 with X_OK for the executables.
11912 * collect2.c (main): Call find_a_file with X_OK.
11914 2013-06-19 Steve Ellcey <sellcey@mips.com>
11917 * config/mips/mips.md (casesi_internal_mips16_<mode>):
11918 Use NEXT_INSN instead of next_real_insn.
11920 2013-06-19 Jan Hubicka <jh@suse.cz>
11922 * cgraph.h (const_value_known_p): Replace by ...
11923 (ctor_for_folding): .. this one.
11924 * cgraphunit.c (process_function_and_variable_attributes): Use it.
11925 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
11926 * expr.c (expand_expr_real_1): Likewise.
11927 (string_constant): Likewise.
11928 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
11929 * ipa.c (process_references): Likewise.
11930 (symtab_remove_unreachable_nodes): Likewise.
11931 * ipa-inline-analysis.c (param_change_prob): Likewise.
11932 * gimple-fold.c (canonicalize_constructor_val): Likewise.
11933 (get_base_constructor): Likwise.
11934 * varpool.c (varpool_remove_node): Likewise.
11935 (varpool_remove_initializer): LIkewise.
11936 (dump_varpool_node): LIkwise.
11937 (const_value_known_p): Rewrite to ...
11938 (ctor_for_folding): ... this one.
11940 2013-06-19 Jakub Jelinek <jakub@redhat.com>
11943 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
11944 PERSONALITY in $PATH derived prefixes.
11946 2013-06-19 Jeff Law <law@redhat.com>
11948 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
11951 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
11952 (simplify_bitwise_binary): Use it to simpify certain binary ops on
11955 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
11957 * config/arm/vfp.md: Move VFP instruction classification documentation
11959 * config/arm/arm.md: ... here. Update instruction classification
11962 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
11964 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
11965 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
11966 pattern. Use more efficient sequences on ARMv5 and Thumb2.
11968 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
11971 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11972 with NEXT_INSN. Use tablejump_p to check for jump table data
11975 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
11978 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
11979 that now in C++ the value is correct per the C++ standards.
11981 2013-06-19 Richard Biener <rguenther@suse.de>
11983 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
11984 for global context.
11986 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11989 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11992 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11993 with next_active_insn.
11995 2013-06-18 Sriraman Tallam <tmsriram@google.com>
11997 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
11998 functions are inlined during failures to flag an error.
11999 * tree-inline.c (expand_call_inline): Allow the error to be flagged
12000 in early inline pass.
12002 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
12004 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
12007 2013-06-18 Julian Brown <julian@codesourcery.com>
12009 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
12010 Permit virtual register pre-reload if !strict.
12011 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
12013 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
12015 * config/arm/neon.md (movmisalign<mode>): Use
12016 neon_perm_struct_or_reg_operand instead of
12017 neon_struct_or_register_operand.
12018 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
12019 neon_permissive_struct_operand instead of neon_struct_operand.
12020 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
12021 neon_vector_mem_operand.
12022 * config/arm/predicates.md (neon_struct_operand): Adjust call to
12023 neon_vector_mem_operand.
12024 (neon_permissive_struct_operand): New.
12025 (neon_struct_or_register_operand): Rename to...
12026 (neon_perm_struct_or_reg_operand): This. Adjust call to
12027 neon_vector_mem_operand.
12029 2013-06-18 Richard Biener <rguenther@suse.de>
12031 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
12032 * lto-streamer.h: Include pointer-set.h.
12033 (struct lto_decl_slot): Remove.
12034 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
12035 Remove next_index entry.
12036 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
12037 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
12038 (lto_init_tree_ref_encoder): Adjust.
12039 (lto_destroy_tree_ref_encoder): Likewise.
12040 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
12041 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
12042 (lto_output_decl_index): Adjust.
12043 (lto_new_out_decl_state): Likewise.
12044 (lto_record_function_out_decl_state): Likewise.
12045 * lto-streamer-out.c (copy_function): Likewise.
12047 2013-06-18 Richard Biener <rguenther@suse.de>
12049 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
12050 * cgraphunit.c: Include cfgloop.h.
12051 (init_lowered_empty_function): Initialize the loop tree.
12052 (assemble_thunk): Insert new BBs into loops.
12054 2013-06-18 Richard Biener <rguenther@suse.de>
12056 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
12057 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
12058 the map from cache entry to cache index optional.
12059 (streamer_tree_cache_replace_tree): Adjust accordingly.
12060 (streamer_tree_cache_append): Likewise.
12061 (streamer_tree_cache_delete): Likewise.
12062 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
12063 streamer cache map from cache entry to cache index.
12064 * lto-streamer-out.c (create_output_block): Adjust.
12066 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
12068 * config/arm/arm.md (attribute "insn"): Move multiplication and
12069 division attributes to...
12070 (attribute "type"): ... here. Remove mult.
12071 (attribute "mul32"): New attribute.
12072 (attribute "mul64"): Add umaal.
12073 (*arm_mulsi3): Update attributes.
12074 (*arm_mulsi3_v6): Likewise.
12075 (*thumb_mulsi3): Likewise.
12076 (*thumb_mulsi3_v6): Likewise.
12077 (*mulsi3_compare0): Likewise.
12078 (*mulsi3_compare0_v6): Likewise.
12079 (*mulsi_compare0_scratch): Likewise.
12080 (*mulsi_compare0_scratch_v6): Likewise.
12081 (*mulsi3addsi): Likewise.
12082 (*mulsi3addsi_v6): Likewise.
12083 (*mulsi3addsi_compare0): Likewise.
12084 (*mulsi3addsi_compare0_v6): Likewise.
12085 (*mulsi3addsi_compare0_scratch): Likewise.
12086 (*mulsi3addsi_compare0_scratch_v6): Likewise.
12087 (*mulsi3subsi): Likewise.
12088 (*mulsidi3adddi): Likewise.
12089 (*mulsi3addsi_v6): Likewise.
12090 (*mulsidi3adddi_v6): Likewise.
12091 (*mulsidi3_nov6): Likewise.
12092 (*mulsidi3_v6): Likewise.
12093 (*umulsidi3_nov6): Likewise.
12094 (*umulsidi3_v6): Likewise.
12095 (*umulsidi3adddi): Likewise.
12096 (*umulsidi3adddi_v6): Likewise.
12097 (*smulsi3_highpart_nov6): Likewise.
12098 (*smulsi3_highpart_v6): Likewise.
12099 (*umulsi3_highpart_nov6): Likewise.
12100 (*umulsi3_highpart_v6): Likewise.
12101 (mulhisi3): Likewise.
12102 (*mulhisi3tb): Likewise.
12103 (*mulhisi3bt): Likewise.
12104 (*mulhisi3tt): Likewise.
12105 (maddhisi4): Likewise.
12106 (*maddhisi4tb): Likewise.
12107 (*maddhisi4tt): Likewise.
12108 (maddhidi4): Likewise.
12109 (*maddhidi4tb): Likewise.
12110 (*maddhidi4tt): Likewise.
12111 (divsi3): Likewise.
12112 (udivsi3): Likewise.
12113 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
12114 (thumb2_mulsi_short_compare0): Likewise.
12115 (thumb2_mulsi_short_compare0_scratch): Likewise.
12116 * config/arm/arm1020e.md (1020mult1): Update attribute change.
12117 (1020mult2): Likewise.
12118 (1020mult3): Likewise.
12119 (1020mult4): Likewise.
12120 (1020mult5): Likewise.
12121 (1020mult6): Likewise.
12122 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
12124 (cortex_a15_mult64): Likewise.
12125 (cortex_a15_sdiv): Likewise.
12126 (cortex_a15_udiv): Likewise.
12127 * config/arm/arm1026ejs.md (mult1): Update attribute change.
12133 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
12134 (pj4_ir_div): Likewise.
12135 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
12136 (11_mult2): Likewise.
12137 (11_mult3): Likewise.
12138 (11_mult4): Likewise.
12139 (11_mult5): Likewise.
12140 (11_mult6): Likewise.
12141 (11_mult7): Likewise.
12142 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
12143 (cortex_a8_mla): Likewise.
12144 (cortex_a8_mull): Likewise.
12145 (cortex_a8_smulwy): Likewise.
12146 (cortex_a8_smlald): Likewise.
12147 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
12148 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
12149 (cortex_r4_mul_3): Likewise.
12150 (cortex_r4_mla_4): Likewise.
12151 (cortex_r4_mla_3): Likewise.
12152 (cortex_r4_smlald): Likewise.
12153 (cortex_r4_mull): Likewise.
12154 (cortex_r4_sdiv): Likewise.
12155 (cortex_r4_udiv): Likewise.
12156 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
12157 (cortex_a7_idiv): Likewise.
12158 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
12159 (9_mult2): Likewise.
12160 (9_mult3): Likewise.
12161 (9_mult4): Likewise.
12162 (9_mult5): Likewise.
12163 (9_mult6): Likewise.
12164 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
12165 (cortex_a53_sdiv): Likewise.
12166 (cortex_a53_udiv): Likewise.
12167 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
12168 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
12169 (mp626_mult2): Likewise.
12170 (mp626_mult3): Likewise.
12171 (mp626_mult4): Likewise.
12172 * config/arm/fa526.md (526_mult1): Update attribute change.
12173 (526_mult2): Likewise.
12174 * config/arm/arm-generic.md (mult): Update attribute change.
12175 (mult_ldsched_strongarm): Likewise.
12176 (mult_ldsched): Likewise.
12177 (multi_cycle): Likewise.
12178 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
12179 * config/arm/fa606te.md (606te_mult1): Update attribute change.
12180 (606te_mult2): Likewise.
12181 (606te_mult3): Likewise.
12182 (606te_mult4): Likewise.
12183 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
12184 (cortex_a9_mac16): Likewise.
12185 (cortex_a9_multiply): Likewise.
12186 (cortex_a9_mac): Likewise.
12187 (cortex_a9_multiply_long): Likewise.
12188 * config/arm/fa626te.md (626te_mult1): Update attribute change.
12189 (626te_mult2): Likewise.
12190 (626te_mult3): Likewise.
12191 (626te_mult4): Likewise.
12193 2013-06-18 Richard Biener <rguenther@suse.de>
12196 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
12198 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12201 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
12202 with next_active_insn.
12204 2013-06-18 Alan Modra <amodra@gmail.com>
12206 * config/rs6000/rs6000.h (enum data_align): New.
12207 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
12208 (DATA_ABI_ALIGNMENT): Define.
12209 (CONSTANT_ALIGNMENT): Correct comment.
12210 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
12211 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
12213 2013-06-17 David Malcolm <dmalcolm@redhat.com>
12215 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
12216 ATTRIBUTE_UNUSED marking.
12218 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
12220 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
12221 alternative and update.
12222 (aarch64_dup_lanedi): Delete.
12223 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
12224 * config/aarch64/aarch64-simd-builtins.def: Update.
12226 2013-06-17 Richard Biener <rguenther@suse.de>
12228 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
12229 (lto_input_scc): Declare.
12230 (lto_input_tree_1): Likewise.
12231 (struct lto_stats_d): Add num_tree_bodies_output and
12232 num_pickle_refs_output.
12233 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
12234 (lto_read_tree_1): Split out from ...
12235 (lto_read_tree): ... this.
12236 (lto_input_scc): New function.
12237 (lto_input_tree_1): Split out from ...
12238 (lto_input_tree): ... this. Handle LTO_tree_scc.
12239 (lto_data_in_create): Create the streamer cache without hashes.
12240 * lto-streamer-out.c (create_output_block): Create the streamer
12241 cache with hashes when not doing WPA.
12242 (lto_write_tree_1): Split out from ...
12243 (lto_write_tree): ... this.
12244 (get_symbol_initial_value): New function.
12245 (lto_output_tree_1): Split out from ...
12246 (lto_output_tree): ... this. Write trees as series of SCCs
12247 using a DFS walk via DFS_write_tree.
12248 (struct sccs, struct scc_entry): New types.
12249 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
12250 (DFS_write_tree_body): New function.
12251 (DFS_write_tree): Likewise.
12252 (hash_tree): Likewise.
12253 (scc_entry_compare): Likewise.
12254 (hash_scc): Likewise.
12255 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
12256 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
12257 TREE_CHAIN as regular reference.
12258 (streamer_read_integer_cst): Remove.
12259 (streamer_get_pickled_tree): Adjust.
12260 * tree-streamer-out.c (streamer_write_chain): Disable streaming
12261 of DECL_EXTERNALs in BLOCK_VARS for now.
12262 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
12264 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
12265 Add hash value argument and record that if hashes are recorded
12267 (streamer_tree_cache_insert_1): Adjust.
12268 (streamer_tree_cache_insert): Likewise.
12269 (streamer_tree_cache_insert_at): Rename to ...
12270 (streamer_tree_cache_replace_tree): ... this and adjust.
12271 (streamer_tree_cache_append): Adjust.
12272 (record_common_node): Likewise.
12273 (streamer_tree_cache_create): Add argument whether to
12274 record hash values together with trees.
12275 (streamer_tree_cache_delete): Adjust.
12276 * tree-streamer.h (struct streamer_tree_cache_d): Add
12278 (streamer_read_integer_cst): Remove.
12279 (streamer_tree_cache_insert): Adjust.
12280 (streamer_tree_cache_append): Likewise.
12281 (streamer_tree_cache_insert_at): Rename to ...
12282 (streamer_tree_cache_replace_tree): ... this and adjust.
12283 (streamer_tree_cache_create): Add argument whether to record hashes.
12284 (streamer_tree_cache_get): Rename to ...
12285 (streamer_tree_cache_get_tree): ... this.
12286 (streamer_tree_cache_get_hash): New function.
12287 * tree.c (cache_integer_cst): New function.
12288 * tree.h (cache_integer_cst): Declare.
12289 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
12290 * lto-symtab.c (lto_varpool_replace_node): Only release
12291 DECL_INITIAL of non-prevailing decls.
12292 * varpool.c (varpool_remove_initializer): Do not release
12293 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
12295 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
12297 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
12298 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
12299 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
12300 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
12301 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
12302 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
12303 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
12304 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
12305 instead of TARGET_64BIT.
12306 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
12307 Require ISA_HAS_<D>DIV.
12309 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
12311 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
12312 (mips*-*-linux*): Move default with_llsc setting to where other
12314 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
12316 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
12317 Likewise. Remove default with_tune setting. Move default float
12318 setting to its own block. Handle with_llsc in the same block as above.
12320 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
12322 PR rtl-optimization/57425
12323 PR rtl-optimization/57569
12324 * alias.c (write_dependence_p): Add new parameters mem_mode,
12325 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
12326 Changed all callers.
12327 (canon_anti_dependence): New function.
12328 * cse.c (check_dependence): Use canon_anti_dependence.
12329 * cselib.c (cselib_invalidate_mem): Likewise.
12330 * rtl.h (canon_anti_dependence): Declare.
12332 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
12334 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
12335 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
12336 ".set mips3" for 64-bit targets.
12338 2013-06-15 Dehao Chen <dehao@google.com>
12340 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
12341 * gimple-low.c (gimple_check_call_matching_types): Likewise.
12342 (gimple_check_call_args): Likewise.
12343 * value-prof.c (check_ic_target): Likewise.
12344 * ipa-inline.c (early_inliner): Likewise.
12345 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
12346 * cgraph.c (cgraph_create_edge_1): Likewise.
12347 (cgraph_make_edge_direct): Likewise.
12349 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
12352 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
12353 rs6000_output_move_128bit to handle emitting quad memory
12354 operations. Set attribute length to 8 bytes.
12356 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
12358 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
12360 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
12361 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
12362 (aarch64_<su>mlsl<mode>): Likewise.
12364 2013-06-14 Mike Stump <mikestump@comcast.net>
12366 * Makefile.in (TARGET_H): Add insn-codes.h.
12368 2013-06-14 Alan Modra <amodra@gmail.com>
12370 PR middle-end/57134
12371 PR middle-end/57586
12372 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
12373 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
12374 bitfield expansion when EXPAND_MEMORY.
12375 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
12377 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
12379 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
12380 test for clearing quad memory on 32-bit later.
12382 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
12384 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
12385 (fold_negate_expr): Likewise.
12386 (fold_real_zero_addition_p): Handle vectors.
12387 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
12389 2013-06-14 Alan Modra <amodra@gmail.com>
12391 * varasm.c (force_const_mem): Revert 2013-06-07 change.
12393 2013-06-13 Jan Hubicka <jh@suse.cz>
12395 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12396 Local comdats are not externally visible.
12397 * symtab.c (dump_symtab_base): Dump externally visible.
12398 (verify_symtab_base): Verify back links in the symtab hash.
12400 2013-06-13 Bin Cheng <bin.cheng@arm.com>
12402 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
12403 CONVERT_EXPR as equal nodes.
12405 2013-06-13 Bin Cheng <bin.cheng@arm.com>
12407 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
12409 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
12411 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
12412 Generalize to complex and vector.
12413 * tree.c (build_all_ones_cst): New function.
12414 * tree.h (build_all_ones_cst): Declare it.
12416 2013-06-13 Alan Modra <amodra@gmail.com>
12418 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
12419 * config/rs6000/rs6000.md (signbittf2): New insn.
12420 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
12421 (abstf2_internal, cmptf_internal2): Likewise.
12422 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
12424 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12425 Pat Haugen <pthaugen@us.ibm.com>
12426 Peter Bergner <bergner@vnet.ibm.com>
12428 * config/rs6000/rs6000.c (emit_load_locked): Add support for
12429 power8 byte, half-word, and quad-word atomic instructions.
12430 (emit_store_conditional): Likewise.
12431 (rs6000_expand_atomic_compare_and_swap): Likewise.
12432 (rs6000_expand_atomic_op): Likewise.
12434 * config/rs6000/sync.md (larx): Add new modes for power8.
12436 (AINT): New mode iterator to include TImode as well as normal
12437 integer modes on power8.
12438 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
12439 that VSX registers are not considered. Use AINT mode iterator
12440 instead of INT1 to allow inclusion of quad word atomic operations
12442 (load_locked<mode>): Likewise.
12443 (store_conditional<mode>): Likewise.
12444 (atomic_compare_and_swap<mode>): Likewise.
12445 (atomic_exchange<mode>): Likewise.
12446 (atomic_nand<mode>): Likewise.
12447 (atomic_fetch_<fetchop_name><mode>): Likewise.
12448 (atomic_nand_fetch<mode>): Likewise.
12449 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
12451 (ATOMIC): On power8, add QImode, HImode modes.
12452 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
12453 modes that promote to SImode.
12454 (load_lockedti): Convert TImode arguments to PTImode, so that we
12455 get a guaranteed even/odd register pair.
12456 (load_lockedpti): Likewise.
12457 (store_conditionalti): Likewise.
12458 (store_conditionalpti): Likewise.
12460 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
12461 atomic load/store instructions.
12464 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
12466 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
12468 (insn_count): New attribute, with most cases extracted from...
12469 (length): ...here. Redefine most cases in terms of insn_count.
12470 (single_insn): Delete.
12471 (can_delay): Use insn_count to check for single instructions.
12472 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
12473 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
12474 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
12475 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
12476 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
12477 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
12478 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
12479 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
12480 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
12481 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
12482 rather than "length".
12483 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
12484 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
12485 Use "insn_count" rather than "length".
12486 * config/mips/mips-dsp.md
12487 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
12488 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
12491 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
12493 PR tree-optimization/57361
12494 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
12496 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
12498 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
12500 (aarch64_simd_combine<mode>): New instruction expansion.
12501 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
12502 function prototype.
12503 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
12504 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
12506 2013-06-12 Jan Hubicka <jh@suse.cz>
12508 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
12509 decl has when in streaming stage.
12510 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
12511 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
12513 2013-06-12 Roland Stigge <stigge@antcom.de>
12516 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
12518 2013-06-12 Jakub Jelinek <jakub@redhat.com>
12520 PR tree-optimization/57537
12521 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
12522 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
12524 2013-06-12 Richard Biener <rguenther@suse.de>
12526 * data-streamer.h (streamer_write_char_stream): CSE
12527 obs->current_pointer.
12528 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
12529 streamer_write_char_stream manually and optimize the resulting loop.
12530 (streamer_write_hwi_stream): Likewise.
12532 2013-06-12 Jan Hubicka <jh@suse.cz>
12534 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
12535 * cgraph.h (varpool_create_empty_node): Declare.
12536 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
12538 * symtab.c (symtab_unregister_node): Be lax about missin entries
12540 (symtab_get_node): Update comment.
12541 * varpool.c (varpool_create_empty_node): Break out from ...
12542 (varpool_node_for_decl): ... here.
12543 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
12545 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
12547 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
12548 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
12549 part. Use straight-line flow at the end.
12550 <COMPONENT_REF>: Remove superfluous else.
12551 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
12553 2013-06-12 Jakub Jelinek <jakub@redhat.com>
12556 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
12557 target hook even for !TREE_PUBLIC decls. If no resolution info
12558 is available, return false for common and external decls.
12560 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12562 * config/rl78/constraints.md (U): New constraint.
12563 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
12566 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
12569 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
12570 to allow returning address to AT_PLATFORM name.
12572 2013-06-11 Jan Hubicka <jh@suse.cz>
12574 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
12575 * cgraph.h (symtab_node_base): Add weakref flag.
12576 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
12577 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
12578 (output_weakrefs): Use weakref flag.
12579 * fold-const.c (simple_operand_p): Handle WEAK.
12580 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
12581 * ipa.c (varpool_externally_visible_p): Drop weakref.
12582 (function_and_variable_visibility): Update comment; fix weakref
12583 sanity checks; do not clear DECL_WEAK on them.
12584 * lto-cgraph.c (lto_output_node): update.
12585 (lto_output_varpool_node): Update.
12586 (input_overwrite_node): Update.
12587 (input_node): Update.
12588 (input_varpool_node): Update.
12589 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
12590 (lto_symtab_merge_symbols): Add sanity check.
12591 (lto_symtab_prevailing_decl): Do not special case weakrefs.
12592 * passes.c (rest_of_decl_compilation): Set static flag, too.
12593 * symtab.c (dump_symtab_base): Dump weakref.
12594 (verify_symtab_base): Sanity check weakrefs.
12595 (symtab_make_decl_local): Remove duplicated code.
12596 (symtab_alias_ultimate_target): Simplify.
12597 * varpool.c (varpool_create_variable_alias): Set weakref flag.
12599 2013-06-11 Tom de Vries <tom@codesourcery.com>
12601 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
12602 sequence_vect == NULL.
12604 2013-06-11 DJ Delorie <dj@redhat.com>
12606 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
12607 (rl78_unwind_word_mode): New.
12609 2013-06-11 David Malcolm <dmalcolm@redhat.com>
12611 * final.c (debug_prefix_maps): Make static.
12613 2013-06-11 David Malcolm <dmalcolm@redhat.com>
12615 * function.c (initial_trampoline): Remove stray copy.
12617 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
12619 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
12621 2013-06-11 Martin Jambor <mjambor@suse.cz>
12623 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
12624 within bounds at the beginning of the function.
12626 2013-06-11 Alan Modra <amodra@gmail.com>
12628 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
12630 (get_named_section): Don't NULL !DECL_P decl.
12632 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
12634 * doc/invoke.texi (core-avx2): Document.
12636 (atom): Updated with MOVBE.
12638 2013-06-11 Richard Biener <rguenther@suse.de>
12640 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
12642 2013-06-11 Anton Blanchard <anton@samba.org>
12644 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
12645 correct shift value in little-endian mode.
12647 2013-06-11 Jakub Jelinek <jakub@redhat.com>
12650 * varasm.c (get_variable_align): Move #endif to the right place.
12652 2013-06-10 Cary Coutant <ccoutant@google.com>
12654 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
12655 for hash so that hash table traversal order is deterministic.
12657 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12658 Pat Haugen <pthaugen@us.ibm.com>
12659 Peter Bergner <bergner@vnet.ibm.com>
12661 * config/rs6000/vector.md (GPR move splitter): Do not split moves
12662 of vectors in GPRS if they are direct moves or quad word load or
12665 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
12667 (direct_move_p): Likewise.
12668 (quad_load_store_p): Likewise.
12670 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
12671 classes into bins based on the physical register type.
12672 (reg_class_to_reg_type): Likewise.
12673 (IS_STD_REG_TYPE): Likewise.
12674 (IS_FP_VECT_REG_TYPE): Likewise.
12675 (reload_fpr_gpr): Arrays to determine what insn to use if we can
12676 use direct move instructions.
12677 (reload_gpr_vsx): Likewise.
12678 (reload_vsx_gpr): Likewise.
12679 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
12680 information that is a simplification of register classes. Also
12681 precalculate direct move reload helpers.
12682 (direct_move_p): New function to return true if the operation can
12683 be done as a direct move instruciton.
12684 (quad_load_store_p): New function to return true if the operation
12685 is a quad memory operation.
12686 (rs6000_legitimize_address): If quad memory, only allow register
12687 indirect for TImode addresses.
12688 (rs6000_legitimate_address_p): Likewise.
12689 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
12690 (rs6000_reload_register_type): Likewise.
12691 (register_to_reg_type): Return register type.
12692 (rs6000_secondary_reload_simple_move): New helper function for
12693 secondary reload and secondary memory needed to identify anything
12694 that is a simple move, and does not need reloading.
12695 (rs6000_secondary_reload_direct_move): New helper function for
12696 secondary reload to identify cases that can be done with several
12697 instructions via the direct move instructions.
12698 (rs6000_secondary_reload_move): New helper function for secondary
12699 reload to identify moves between register types that can be done.
12700 (rs6000_secondary_reload): Add support for quad memory operations
12701 and for direct move.
12702 (rs6000_secondary_memory_needed): Likewise.
12703 (rs6000_debug_secondary_memory_needed): Change argument names.
12704 (rs6000_output_move_128bit): New function to return the move to
12705 use for 128-bit moves, including knowing about the various
12706 limitations of quad memory operations.
12708 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
12709 memory operations. call rs6000_output_move_128bit for the actual
12710 instruciton(s) to generate.
12711 (vsx_movti_64bit): Likewise.
12713 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
12714 (UNSPEC_P8V_MTVSRWZ): Likewise.
12715 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
12716 (UNSPEC_P8V_MTVSRD): Likewise.
12717 (UNSPEC_P8V_XXPERMDI): Likewise.
12718 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
12719 (UNSPEC_FUSION_GPR): Likewise.
12720 (FMOVE128_GPR): New iterator for direct move.
12721 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
12722 (f32_sv): Likewise.
12723 (f32_dm): Likewise.
12724 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
12725 loads and direct move instructions.
12726 (zero_extendsidi2_lfiwzx): Likewise.
12727 (extendsidi2_lfiwax): Likewise.
12728 (extendsidi2_nocell): Likewise.
12729 (floatsi<mode>2_lfiwax): Likewise.
12730 (lfiwax): Likewise.
12731 (floatunssi<mode>2_lfiwzx): Likewise.
12732 (lfiwzx): Likewise.
12733 (fix_trunc<mode>_stfiwx): Likewise.
12734 (fixuns_trunc<mode>_stfiwx): Likewise.
12735 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
12736 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
12737 (parity<mode>2_cmpb): Set length/type attr.
12738 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
12739 for 'mr.' to fast_compare.
12740 (bpermd_<mode>): Change type attr to popcnt.
12741 (p8_fmrgow_<mode>): New insns for power8 direct move support.
12742 (p8_mtvsrwz_1): Likewise.
12743 (p8_mtvsrwz_2): Likewise.
12744 (reload_fpr_from_gpr<mode>): Likewise.
12745 (p8_mtvsrd_1): Likewise.
12746 (p8_mtvsrd_2): Likewise.
12747 (p8_xxpermdi_<mode>): Likewise.
12748 (reload_vsx_from_gpr<mode>): Likewise.
12749 (reload_vsx_from_gprsf): Likewise.
12750 (p8_mfvsrd_3_<mode>): LIkewise.
12751 (reload_gpr_from_vsx<mode>): Likewise.
12752 (reload_gpr_from_vsxsf): Likewise.
12753 (p8_mfvsrd_4_disf): Likewise.
12754 (multi-word GPR splits): Do not split direct moves or quad memory
12757 2013-06-10 David Malcolm <dmalcolm@redhat.com>
12759 * tree-into-ssa.c (interesting_blocks): Make static.
12761 2013-06-10 Jakub Jelinek <jakub@redhat.com>
12764 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
12765 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
12766 Use DATA_ABI_ALIGNMENT for that case instead if defined.
12767 (get_variable_align): New function.
12768 (get_variable_section, emit_bss, emit_common,
12769 assemble_variable_contents, place_block_symbol): Use
12770 get_variable_align instead of DECL_ALIGN.
12771 (assemble_noswitch_variable): Add align argument, use it
12772 instead of DECL_ALIGN.
12773 (assemble_variable): Adjust caller. Use get_variable_align
12774 instead of DECL_ALIGN.
12775 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
12777 (DATA_ABI_ALIGNMENT): Define.
12778 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
12779 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
12780 opt is false, only return the psABI mandated alignment increase.
12781 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
12782 (DATA_ABI_ALIGNMENT): ... this.
12783 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
12784 (DATA_ABI_ALIGNMENT): ... this.
12785 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
12786 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
12787 (DATA_ABI_ALIGNMENT): ... this.
12788 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
12789 * doc/tm.texi: Regenerated.
12791 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
12793 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
12794 cmp_code to construct REG_EQUAL note.
12796 2013-06-09 Jakub Jelinek <jakub@redhat.com>
12799 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
12800 that operands[2] doesn't overlap with operands[0].
12802 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
12803 Jan Hubicka <jh@suse.cz>
12805 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
12806 hack to mark symbols as used.
12808 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
12810 PR rtl-optimization/57559
12811 * lra-constraints.c (process_alt_operands): Don't discourage
12812 memory with known offset for offsetable memory constraint.
12813 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
12815 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
12817 * varasm.c (struct oc_local_state): Reorder fields.
12818 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
12819 and adjust accordingly.
12820 (output_constructor): Reorder initialization code and adjust call to
12821 output_constructor_bitfield.
12823 2013-06-07 Jan Hubicka <jh@suse.cz>
12825 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
12827 2013-06-07 David Malcolm <dmalcolm@redhat.com>
12829 * tree-object-size.c (unknown): Make const.
12831 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12833 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
12834 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
12835 for last alternative in the cpu_facility attribute.
12837 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12840 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
12841 (xordi3): Change operand 2 constraint to arm_xordi_operand.
12842 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
12843 * config/arm/constraints.md (Dg): New constraint.
12844 * config/arm/neon.md (xordi3_neon): Remove.
12845 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
12846 * config/arm/predicates.md (arm_xordi_operand): New predicate.
12848 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12850 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
12851 Clean up alternatives.
12853 2013-06-07 Alan Modra <amodra@gmail.com>
12855 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
12858 2013-06-07 Alan Modra <amodra@gmail.com>
12860 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
12862 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12864 * config/arm/constraints.md (Df): New constraint.
12865 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
12866 Correct length attribute for last two alternatives.
12868 2013-06-07 Alan Modra <amodra@gmail.com>
12870 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
12871 override user -mfp-in-toc.
12872 (offsettable_ok_by_alignment): Consider just the current access
12873 rather than the whole object, unless BLKmode. Handle
12874 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
12875 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
12876 for -mcmodel=medium.
12877 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
12878 override user -mfp-in-toc or -msum-in-toc. Default to
12879 -mno-fp-in-toc for -mcmodel=medium.
12881 2013-06-06 DJ Delorie <dj@redhat.com>
12883 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
12884 TARGET_VALID_POINTER_MODE.
12886 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12887 Pat Haugen <pthaugen@us.ibm.com>
12888 Peter Bergner <bergner@vnet.ibm.com>
12890 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
12891 Document new power8 builtins.
12893 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
12894 condition code register, to allow 128-bit logical operations to be
12895 done in the VSX or GPR registers.
12896 (nor<mode>3): Use the canonical form for nor.
12897 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
12898 vclz*, and vpopcnt* vector instructions.
12899 (nand<mode>3): Likewise.
12900 (orc<mode>3): Likewise.
12901 (clz<mode>2): LIkewise.
12902 (popcount<mode>2): Likewise.
12904 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
12905 that only the GPRs are recognized.
12907 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12908 support for new power8 builtins.
12910 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
12912 (xscvdpspn): Likewise.
12918 (vpopcnt): Likewise.
12919 (vpopcntb): Likewise.
12920 (vpopcnth): Likewise.
12921 (vpopcntw): Likewise.
12922 (vpopcntd): Likewise.
12924 (vmrgew): Likewise.
12925 (vmrgow): Likewise.
12927 (eqv_v16qi3): Likewise.
12928 (eqv_v8hi3): Likewise.
12929 (eqv_v4si3): Likewise.
12930 (eqv_v2di3): Likewise.
12931 (eqv_v4sf3): Likewise.
12932 (eqv_v2df3): Likewise.
12934 (nand_v16qi3): Likewise.
12935 (nand_v8hi3): Likewise.
12936 (nand_v4si3): Likewise.
12937 (nand_v2di3): Likewise.
12938 (nand_v4sf3): Likewise.
12939 (nand_v2df3): Likewise.
12941 (orc_v16qi3): Likewise.
12942 (orc_v8hi3): Likewise.
12943 (orc_v4si3): Likewise.
12944 (orc_v2di3): Likewise.
12945 (orc_v4sf3): Likewise.
12946 (orc_v2df3): Likewise.
12948 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
12949 allow power8 quad mode in 64-bit.
12950 (rs6000_builtin_vectorized_function): Add support to vectorize
12951 ISA 2.07 count leading zeros, population count builtins.
12952 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
12953 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
12954 (builtin_function_type): Add vgbbd builtin function which takes an
12956 (altivec_expand_vec_perm_const): Add support for new power8 merge
12959 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
12960 that does not include TImdoe for use with 32-bit.
12961 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
12963 (UNSPEC_VSX_CVDPSPN): Likewise.
12964 (vsx_xscvdpspn): Likewise.
12965 (vsx_xscvspdpn): Likewise.
12966 (vsx_xscvdpspn_scalar): Likewise.
12967 (vsx_xscvspdpn_directmove): Likewise.
12968 (vsx_and<mode>3): Split logical operations into 32-bit and
12969 64-bit. Add support to do logical operations on TImode as well as
12970 VSX vector types. Allow logical operations to be done in either
12971 VSX registers or in general purpose registers in 64-bit mode. Add
12972 splitters if GPRs were used. For AND, add clobber of CCmode to
12973 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
12975 (vsx_and<mode>3_32bit): Likewise.
12976 (vsx_and<mode>3_64bit): Likewise.
12977 (vsx_ior<mode>3): Likewise.
12978 (vsx_ior<mode>3_32bit): Likewise.
12979 (vsx_ior<mode>3_64bit): Likewise.
12980 (vsx_xor<mode>3): Likewise.
12981 (vsx_xor<mode>3_32bit): Likewise.
12982 (vsx_xor<mode>3_64bit): Likewise.
12983 (vsx_one_cmpl<mode>2): Likewise.
12984 (vsx_one_cmpl<mode>2_32bit): Likewise.
12985 (vsx_one_cmpl<mode>2_64bit): Likewise.
12986 (vsx_nor<mode>3): Likewise.
12987 (vsx_nor<mode>3_32bit): Likewise.
12988 (vsx_nor<mode>3_64bit): Likewise.
12989 (vsx_andc<mode>3): Likewise.
12990 (vsx_andc<mode>3_32bit): Likewise.
12991 (vsx_andc<mode>3_64bit): Likewise.
12992 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
12993 and xxlorc instructions.
12994 (vsx_eqv<mode>3_64bit): Likewise.
12995 (vsx_nand<mode>3_32bit): Likewise.
12996 (vsx_nand<mode>3_64bit): Likewise.
12997 (vsx_orc<mode>3_32bit): Likewise.
12998 (vsx_orc<mode>3_64bit): Likewise.
13000 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
13002 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
13004 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
13005 (p8_vmrgow): Likewise.
13006 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
13007 GPRs to be split under VSX.
13008 (p8v_clz<mode>2): Add power8 count leading zero support.
13009 (p8v_popcount<mode>2): Add power8 population count support.
13010 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
13013 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
13016 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
13018 (vec_nand): Likewise.
13019 (vec_vclz): Likewise.
13020 (vec_vclzb): Likewise.
13021 (vec_vclzd): Likewise.
13022 (vec_vclzh): Likewise.
13023 (vec_vclzw): Likewise.
13024 (vec_vgbbd): Likewise.
13025 (vec_vmrgew): Likewise.
13026 (vec_vmrgow): Likewise.
13027 (vec_vpopcnt): Likewise.
13028 (vec_vpopcntb): Likewise.
13029 (vec_vpopcntd): Likewise.
13030 (vec_vpopcnth): Likewise.
13031 (vec_vpopcntw): Likewise.
13033 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13035 PR rtl-optimization/57468
13036 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
13039 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13041 PR rtl-optimization/57459
13042 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
13043 type when setting live regs.
13045 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13047 * config/s390/s390.opt (mlra): New option.
13048 * config/s390/s390.c (s390_decompose_address): Check displacement
13049 for all registers for LRA.
13050 (s390_secondary_reload): Don't used secondary reloads for LRA.
13051 (s390_lra_p): New function.
13052 (TARGET_LRA_P): Define.
13053 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
13054 of attribute cpu_facility to zarch for the last alternative.
13055 (*cmpmem_short): Ditto.
13057 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
13059 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
13060 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
13061 (arm_expand_prologue): Likewise.
13063 2013-06-06 Teresa Johnson <tejohnson@google.com>
13066 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
13067 as this is now done by redirect_edge_and_branch_force.
13068 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
13069 barriers, and fix interaction with splitting.
13070 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
13071 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
13072 reflect changes made in the routine.
13073 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
13074 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
13075 since this is called in cfglayout mode, and replace partition fixup
13076 with assert as that is now done by force_nonfallthru_and_redirect.
13077 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
13078 already be marked with region crossing note.
13079 (insert_section_boundary_note): Make non-static, gate on flag
13080 has_bb_partition, rewrite to also check for multiple partitions.
13081 (rest_of_handle_reorder_blocks): Remove call to
13082 insert_section_boundary_note, now done later during free_cfg.
13083 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
13084 * bb-reorder.h (insert_section_boundary_note): Declare.
13085 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
13086 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
13087 invoke insert_section_boundary_note.
13088 (try_redirect_by_replacing_jump): Remove unnecessary
13089 check for region crossing note.
13090 (fixup_partition_crossing): New function.
13091 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
13092 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
13093 in non-cfglayout mode.
13094 (force_nonfallthru_and_redirect): Fixup partition boundaries,
13095 remove old code that tried to do this. Emit barrier correctly
13096 when we are in cfglayout mode.
13097 (last_bb_in_partition): New function.
13098 (rtl_split_edge): Correctly fixup partition boundaries.
13099 (commit_one_edge_insertion): Remove old code that tried to
13100 fixup region crossing edge since this is now handled in
13101 split_block, and set up insertion point correctly since
13102 block may now end in a jump.
13103 (verify_hot_cold_block_grouping): Guard against checking when not in
13104 linearized RTL mode.
13105 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
13107 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
13108 rtl_verify_flow_info, so not called in cfglayout mode.
13109 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
13110 (fixup_reorder_chain): Remove old code that attempted to fixup region
13111 crossing note as this is now handled in force_nonfallthru_and_redirect.
13112 (duplicate_insn_chain): Don't duplicate switch section notes.
13113 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
13115 * basic-block.h (emit_barrier_after_bb): Declare.
13117 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13119 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
13120 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
13121 arm_usatsihi): Adjust alternatives for arm_restrict_it.
13123 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13125 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
13127 * config/arm/ldmstm.md: Regenerate.
13129 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13131 * config/arm/sync.md (atomic_loaddi_1):
13132 Disable predication for arm_restrict_it.
13133 (arm_load_exclusive<mode>): Likewise.
13134 (arm_load_exclusivesi): Likewise.
13135 (arm_load_exclusivedi): Likewise.
13136 (arm_load_acquire_exclusive<mode>): Likewise.
13137 (arm_load_acquire_exclusivesi): Likewise.
13138 (arm_load_acquire_exclusivedi): Likewise.
13139 (arm_store_exclusive<mode>): Likewise.
13140 (arm_store_exclusive<mode>): Likewise.
13141 (arm_store_release_exclusivedi): Likewise.
13142 (arm_store_release_exclusive<mode>): Likewise.
13144 2013-06-06 Richard Biener <rguenther@suse.de>
13146 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
13148 (lto_tag_is_tree_code_p): Adjust.
13149 (lto_tag_is_gimple_code_p): Likewise.
13150 (lto_gimple_code_to_tag): Likewise.
13151 (lto_tag_to_gimple_code): Likewise.
13152 (lto_tree_code_to_tag): Likewise.
13153 (lto_tag_to_tree_code): Likewise.
13154 * data-streamer.h (streamer_write_hwi_in_range): Use
13155 uhwi streaming to stream the normalized range.
13156 (streamer_read_hwi_in_range): Likewise.
13158 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13160 * config/arm/arm.md (enabled_for_depr_it): New attribute.
13161 (predicable_short_it): Likewise.
13162 (predicated): Likewise.
13163 (enabled): Handle above.
13164 (define_cond_exec): Set predicated attribute to yes.
13166 2013-06-05 Mike Stump <mikestump@comcast.net>
13168 * gdbinit.in (__FUNCTION__): Add.
13170 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
13172 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
13173 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
13175 2013-06-05 Jan Hubicka <jh@suse.cz>
13177 * varasm.c (mark_decl_referenced): Revert the removal until targets
13180 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
13182 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
13183 instead of mark_decl_referenced.
13185 2013-06-05 Jan Hubicka <jh@suse.cz>
13187 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
13188 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
13189 and symtab_used_from_object_file_p.
13190 (cgraph_make_node_local_1): Clear forced_by_abi.
13191 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
13192 * cgraph.h (symtab_node_base): Add forced_by_abi.
13193 (decide_is_variable_needed): Remove.
13194 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
13195 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
13196 (decide_is_symbol_needed): ... this one; handle symbols in general;
13197 always analyze virtuals; honnor forced_by_abi.
13198 (cgraph_finalize_function): Update.
13199 (varpool_finalize_decl): Update.
13200 (symbol_defined_and_needed): Remove.
13201 (analyze_functions): Update.
13202 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13203 output_refs, input_overwrite_node): Handle forced_by_abi.
13204 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
13205 (address_taken_from_non_vtable_p): ... this one.
13206 (comdat_can_be_unshared_p_1): New function.
13207 (cgraph_comdat_can_be_unshared_p): Rename to ...
13208 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
13209 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
13210 (function_and_variable_visibility): Clear forced_by_abi as needed.
13211 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
13212 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
13213 * symtab.c (dump_symtab_base): Dump forced_by_abi.
13214 * varpool.c (decide_is_variable_needed): Remove.
13216 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13218 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
13219 (arm_option_override): Override arm_restrict_it where appropriate.
13220 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
13221 * config/arm/arm.opt (mrestrict-it): New command-line option.
13222 * doc/invoke.texi: Document -mrestrict-it.
13224 2013-06-05 David Malcolm <dmalcolm@redhat.com>
13226 * tsan.c (tsan_atomic_table): Make const.
13228 2013-06-05 Richard Biener <rguenther@suse.de>
13230 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
13231 index associated with the tree we are supposed to replace.
13232 * tree-streamer-out.c (pack_ts_base_value_fields): Output
13233 TREE_ASM_WRITTEN as zero for everything but SSA names.
13235 2013-06-05 David Malcolm <dmalcolm@redhat.com>
13237 * tree-ssa-structalias.c (call_stmt_vars): Make static.
13239 2013-06-04 Jan Hubicka <jh@suse.cz>
13241 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
13242 (input_node, input_varpool_node): Handle correctly external same
13244 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
13245 nodes at ltrans stage.
13247 2013-06-04 Jan Hubicka <jh@suse.cz>
13249 * ipa-inline.c (update_caller_keys): Fix availability test.
13250 (update_callee_keys): Likewise.
13251 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
13252 to follow ELF standard.
13254 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
13256 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
13257 (mips64r5900el-*-elf*): New configurations.
13258 * config/mips/mips-cpus.def (r5900): New processor.
13259 * config/mips/mips-tables.opt: Regenerate.
13260 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
13261 (mips_issue_rate): Handle PROCESSOR_R5900.
13262 (mips_reorg_process_insns): Force reorder mode for the R5900.
13263 * config/mips/mips.h (TARGET_MIPS5900): Define.
13264 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
13266 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
13268 * config/mips/mips.md (processor): Add r5900.
13269 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
13271 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13273 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
13274 into function to generate MOVI instruction.
13275 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
13276 (aarch64_preferred_simd_mode): Turn into wrapper.
13277 (aarch64_output_scalar_simd_mov_immediate): New function.
13278 * config/aarch64/aarch64-protos.h: Add prototype for above.
13280 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13282 * config/aarch64/aarch64.c (simd_immediate_info): Remove
13283 element_char member.
13284 (sizetochar): Return signed char.
13285 (aarch64_simd_valid_immediate): Remove elchar and other
13286 unnecessary variables.
13287 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
13288 Calculate element_char as required.
13289 * config/aarch64/aarch64-protos.h: Update and move prototype
13290 for aarch64_output_simd_mov_immediate.
13291 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
13294 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13296 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
13297 information completed by aarch64_simd_valid_immediate.
13298 (aarch64_legitimate_constant_p): Update arguments.
13299 (aarch64_simd_valid_immediate): Work with struct rather than many
13301 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
13302 (aarch64_simd_make_constant): Update arguments.
13303 (aarch64_output_simd_mov_immediate): Work with struct rather than
13304 many pointers. Output immediate directly rather than as operand.
13305 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
13307 * config/aarch64/constraints.md (Dn): Update arguments.
13309 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13311 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
13313 (aarch64_simd_immediate_valid_for_move): Remove.
13314 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
13315 (aarch64_simd_make_constant): Update call.
13316 (aarch64_output_simd_mov_immediate): Update call.
13317 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
13319 * config/aarch64/constraints.md (Dn): Update call.
13321 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13323 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
13324 return type to bool for prototype.
13325 (aarch64_legitimate_constant_p): Check for true instead of not -1.
13326 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
13327 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
13329 2013-06-04 Catherine Moore <clm@codesourcery.com>
13331 * config/mips/mips.opt (meva): New.
13332 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
13333 (ASM_SPEC): Handle -meva.
13334 * doc/invoke.texi (meva): Document.
13336 2013-06-04 Alan Modra <amodra@gmail.com>
13338 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
13341 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13343 * rtl.def: Add extra fourth optional field to define_cond_exec.
13344 * gensupport.c (process_one_cond_exec): Process attributes from
13346 * doc/md.texi: Document fourth field in define_cond_exec.
13348 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
13350 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
13351 out the processing order as in store_bit_field_1.
13353 2013-06-04 Jan Hubicka <jh@suse.cz>
13355 PR middle-end/57500
13356 * cgraphunit.c (cgraph_process_same_body_aliases): Create
13357 non-VAR_DECL node if it does not exist yet.
13359 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
13361 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
13362 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
13363 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
13364 target_cpu_default setting.
13366 2013-06-03 Teresa Johnson <tejohnson@google.com>
13368 * dumpfile.c (opt_info_switch_p): Change -fopt-info
13369 default to -fopt-info=optimized instead of all.
13370 * doc/invoke.texi: Ditto.
13371 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
13372 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
13373 (execute_vect_slp): Emit BB vectorization success under
13374 MSG_OPTIMIZED_LOCATIONS.
13375 * tree-vect-slp.c (vect_slp_transform_bb): Change
13376 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
13377 * tree-vect-loop.c (vect_transform_loop): Ditto.
13379 2013-06-03 Jason Merrill <jason@redhat.com>
13382 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
13383 Use TARGET_EXPR for C++.
13385 2013-06-03 Jakub Jelinek <jakub@redhat.com>
13387 PR rtl-optimization/57268
13388 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
13389 if DEBUG_INSN_P (insn).
13392 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
13394 PR rtl-optimization/57268
13395 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13396 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13398 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13400 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
13401 (ix86_avoid_lea_for_addr): Likewise.
13402 (exact_dependency_1): Likewise.
13403 (ix86_adjust_cost): Likewise.
13404 (swap_top_of_ready_list): Fix formatting and !reload_completed check
13406 (do_reorder_for_imul): Fix typo, formatting and
13407 !reload_completed check removed.
13408 (ix86_sched_reorder): Fix typo and formatting.
13409 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
13412 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
13414 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
13416 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
13418 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
13419 <STRING_CST>: Likewise.
13420 <VECTOR_CST>: Likewise.
13422 2013-06-01 Janus Weil <janus@gcc.gnu.org>
13423 Mikael Morin <mikael@gcc.gnu.org>
13425 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
13426 * config.in: Regenerated.
13427 * configure: Regenerated.
13429 2013-06-01 Jan Hubicka <jh@suse.cz>
13431 PR middle-end/57366
13432 * cgraphunit.c (compile): When weakref is not supported,
13433 set up transparent aliases before final output pass.
13434 * varasm.c (assemble_alias): Do not try to do it here.
13436 2013-06-01 Jan Hubicka <jh@suse.cz>
13438 PR middle-end/57467
13439 * passes.c (for_per_function): Skip unanalyzed functions.
13441 2013-06-01 Jan Hubicka <jh@suse.cz>
13443 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
13444 (lto_symtab_merge_symbols_1): ... this one.
13445 (lto_symtab_merge_cgraph_nodes): Rename to ...
13446 (lto_symtab_merge_symbols): ... this one; simplify.
13447 * cgraph.c (same_body_aliases_done): Rename to ...
13448 (cpp_implicit_aliases_done): ... this one.
13449 (cgraph_create_function_alias): Update.
13450 (cgraph_same_body_alias): Update.
13451 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
13452 (verify_edge_corresponds_to_fndecl): Simplify.
13453 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
13454 (cgraph_node): Remove same_body_alias.
13455 (varpool_node): Remove alias_of and extra_name_alias.
13456 (same_body_aliases_done): Rename to ..
13457 (cpp_implicit_aliases_done): ... this one.
13458 (symtab_alias_ultimate_target): Add default parameter.
13459 (symtab_resolve_alias): New function.
13460 (fixup_same_cpp_alias_visibility): Declare.
13461 (cgraph_function_node): Add default parameter.
13462 (cgraph_node_asm_name): Likewise.
13463 (cgraph_function_or_thunk_node): Add default parameter; do
13464 not ICE when it is NULL.
13465 (varpool_variable_node): Likewise.
13466 * tree-emutls.c (create_emultls_var): Update.
13467 (ipa_lower_emutls): Update.
13468 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
13469 (cgraph_reset_node): Reset alias info.
13470 (cgraph_finalize_function): Update.
13471 (fixup_same_cpp_alias_visibility): Move to symtab.c.
13472 (analyze_function): Simplify.
13473 (cgraph_process_same_body_aliases): Simplify.
13474 (analyze_functions): Fixup same body aliases.
13475 (handle_alias_pairs): Simplify.
13476 (assemble_thunk): Update.
13477 (assemble_thunks_and_aliases): Update.
13478 (output_weakrefs): Rewrite.
13479 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
13480 (lto_output_varpool_node): Likewise.
13481 (compute_ltrans_boundary): Remve assert.
13482 (get_alias_symbol): New functoin.
13483 (input_node): Rewrite alias handling.
13484 (input_varpool_node): Likewise.
13485 * ipa-pure-const.c (propagate_pure_const): Fix formating.
13486 * ipa.c (process_references): Handle weakrefs correctly.
13487 (symtab_remove_unreachable_nodes): Likewise.
13488 * trans-mem.c (get_cg_data): Update.
13489 (ipa_tm_create_version_alias): Update.
13490 (ipa_tm_execute): Update.
13491 * symtab.c (dump_symtab_base): Dump aliases.
13492 (verify_symtab_base): Verify aliases.
13493 (symtab_node_availability): New function.
13494 (symtab_alias_ultimate_target): Simplify.
13495 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
13496 handle all the fixup cases.
13497 (symtab_resolve_alias): New function.
13498 * passes.c (ipa_write_summaries): Handle weakrefs.
13499 * varpool.c (varpool_analyze_node): Simplify.
13500 (assemble_aliases): Update.
13501 (varpool_create_variable_alias): Simplify.
13502 (varpool_extra_name_alias): Simplify.
13503 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
13504 (lto_symtab_merge_symbols): ... this one.
13506 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
13509 PR rtl-optimization/57268
13510 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13511 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13513 2013-06-01 Tobias Burnus <burnus@net-b.de>
13515 Partially reverted:
13516 2013-05-31 Tobias Burnus <burnus@net-b.de>
13518 PR middle-end/57073
13519 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
13522 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
13524 PR rtl-optimization/57268
13525 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13526 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13528 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
13530 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
13531 unordered comparison operators when -fno-trapping-math is in effect
13533 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
13534 and implement unordered comparison operators properly on the e500.
13536 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
13538 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
13539 for constant scalar integers.
13540 (simplify_relational_operation_1): Likewise.
13542 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
13544 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
13545 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
13548 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13549 Igor Zamyatin <igor.zamyatin@intel.com>
13551 Silvermont (SLM) architecture performance tuning.
13552 * config/i386/i386.h (enum ix86_tune_indices): Add
13553 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
13554 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
13556 * config/i386/i386.c (initial_ix86_tune_features)
13557 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
13558 (ix86_lea_outperforms): Handle Silvermont tuning.
13559 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
13561 (ix86_use_lea_for_mov): Likewise.
13562 (ix86_avoid_lea_for_addr): Likewise.
13563 (ix86_lea_for_add_ok): Likewise.
13564 (exact_dependency_1): New function.
13565 (exact_store_load_dependency): Likewise.
13566 (ix86_adjust_cost): Handle Silvermont tuning.
13567 (do_reoder_for_imul): Likewise.
13568 (swap_top_of_ready_list): New function.
13569 (ix86_sched_reorder): Changed to handle Silvermont tuning.
13571 * config/i386/i386.md (peepholes that split memory operand in fp
13574 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
13576 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
13577 Remove un-necessary braces.
13579 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
13581 * config/aarch64/aarch64.c (aarch64_classify_symbol):
13582 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
13584 2013-05-31 Tobias Burnus <burnus@net-b.de>
13586 PR middle-end/57073
13587 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
13589 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13592 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
13593 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
13594 * config/arm/neon.md (iordi3_neon): Remove.
13595 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
13596 * config/arm/predicates.md (imm_for_neon_logic_operand):
13597 Move to earlier in the file.
13598 (neon_logic_op2): Likewise.
13599 (arm_iordi_operand_neon): New predicate.
13601 2013-05-31 Richard Biener <rguenther@suse.de>
13603 PR tree-optimization/57478
13604 PR tree-optimization/57453
13605 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
13608 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13610 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
13611 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
13613 2013-05-30 Tobias Burnus <burnus@net-b.de>
13614 Thomas Koenig <tkoenig@gcc.gnu.org>
13616 PR middle-end/57073
13617 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
13618 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
13620 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
13622 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
13624 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
13626 * target.def (register_usage_leveling_p): New hook.
13627 * targhooks.c (default_register_usage_leveling_p): New.
13628 * targhooks.h (default_register_usage_leveling_p): New prototype.
13629 * lra-assigns.c (register_usage_leveling_p): Use the hook.
13630 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
13631 * doc/tm.texi: Update.
13632 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
13634 2013-05-30 Ian Bolton <ian.bolton@arm.com>
13636 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
13637 (*insv_reg<mode>): New define_insn.
13639 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
13641 PR rtl-optimization/57439
13642 * postreload.c (move2add_valid_value_p): Check that we have
13643 a zero subreg_regno_offset when accessing the register in
13644 the requested mode.
13646 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13647 Igor Zamyatin <igor.zamyatin@intel.com>
13649 Silvermont (SLM) architecture pipeline model, tuning and
13651 * config.gcc: Add slm config options and target.
13653 * config/i386/slm.md: New.
13655 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
13657 * config/i386/i386-c.c (ix86_target_macros_internal): New case
13659 (ix86_target_macros_internal): Likewise.
13661 * config/i386/i386.c (slm_cost): New cost.
13662 (m_SLM): New macro flag.
13663 (initial_ix86_tune_features): Set m_SLM.
13664 (x86_accumulate_outgoing_args): Likewise.
13665 (x86_arch_always_fancy_math_387): Likewise.
13666 (processor_target_table): Add slm cost.
13667 (cpu_names): Add slm cpu name.
13668 (x86_option_override_internal): Set SLM ISA.
13669 (ix86_issue_rate): New case PROCESSOR_SLM.
13670 (ia32_multipass_dfa_lookahead): Likewise.
13671 (fold_builtin_cpu): Add slm.
13673 * config/i386/i386.h (TARGET_SLM): New target macro.
13674 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
13675 (processor_type): Add PROCESSOR_SLM.
13677 * config/i386/i386.md (cpu): Add new value "slm".
13678 (slm.md): Include slm.md.
13680 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
13681 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13683 * config/arm/arm-protos.h: Add and update function protos.
13684 * config/arm/arm.c (use_simple_return_p): New added.
13685 (thumb2_expand_return): Check simple_return flag.
13686 * config/arm/arm.md: Add simple_return and conditional simple_return.
13687 * config/arm/iterators.md: Add iterator for return and simple_return.
13689 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13691 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
13692 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
13693 (arm_emit_vfp_multi_reg_pop): Likewise.
13694 (thumb2_emit_ldrd_pop): Likewise.
13695 (arm_expand_epilogue): Add misc REG_CFA notes.
13696 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
13698 2013-05-29 Lawrence Crowl <crowl@google.com>
13700 * config/arm/t-arm: Update for below.
13702 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
13703 Change type to hash_table. Update dependent calls and types.
13705 * config/i386/t-cygming: Update for below.
13707 * config/i386/t-interix: Update for below.
13709 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
13710 Change type to hash_table. Update dependent calls and types.
13711 (i386_find_on_wrapper_list::wrappers): Likewise.
13713 * config/ia64/t-ia64: Update for below.
13715 * config/ia64/ia64.c (bundle_state_table):
13716 Change type to hash_table. Update dependent calls and types.
13718 * config/mips/mips.c (mips_reorg_process_insns::htab):
13719 Change type to hash_table. Update dependent calls and types.
13721 * config/sol2.c (solaris_comdat_htab):
13722 Change type to hash_table. Update dependent calls and types.
13724 * config/t-sol2: Update for above.
13726 2013-05-29 Teresa Johnson <tejohnson@google.com>
13728 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
13729 functions are not yet marked as defined.
13731 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
13732 Pat Haugen <pthaugen@us.ibm.com>
13733 Peter Bergner <bergner@vnet.ibm.com>
13735 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
13739 (vrotl<mode>3): Likewise.
13740 (vashl<mode>3): Likewise.
13741 (vlshr<mode>3): Likewise.
13742 (vashr<mode>3): Likewise.
13744 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
13745 support for power8 V2DI builtins.
13747 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
13748 power8 V2DI builtins.
13749 (vupkhsw): Likewise.
13750 (vupklsw): Likewise.
13751 (vaddudm): Likewise.
13752 (vminsd): Likewise.
13753 (vmaxsd): Likewise.
13754 (vminud): Likewise.
13755 (vmaxud): Likewise.
13756 (vpkudum): Likewise.
13757 (vpksdss): Likewise.
13758 (vpkudus): Likewise.
13759 (vpksdus): Likewise.
13764 (vsubudm): Likewise.
13765 (vcmpequd): Likewise.
13766 (vcmpgtsd): Likewise.
13767 (vcmpgtud): Likewise.
13768 (vcmpequd_p): Likewise.
13769 (vcmpgtsd_p): Likewise.
13770 (vcmpgtud_p): Likewise.
13771 (vupkhsw): Likewise.
13772 (vupklsw): Likewise.
13773 (vaddudm): Likewise.
13774 (vmaxsd): Likewise.
13775 (vmaxud): Likewise.
13776 (vminsd): Likewise.
13777 (vminud): Likewise.
13778 (vpksdss): Likewise.
13779 (vpksdus): Likewise.
13780 (vpkudum): Likewise.
13781 (vpkudus): Likewise.
13786 (vsubudm): Likewise.
13788 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
13789 support for power8 V2DI instructions.
13791 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
13792 power8 V2DI instructions. Combine pack and unpack insns to use an
13793 iterator for each mode. Check whether a particular mode supports
13794 Altivec instructions instead of just checking TARGET_ALTIVEC.
13795 (UNSPEC_VPKUWUM): Likewise.
13796 (UNSPEC_VPKSHSS): Likewise.
13797 (UNSPEC_VPKSWSS): Likewise.
13798 (UNSPEC_VPKUHUS): Likewise.
13799 (UNSPEC_VPKSHUS): Likewise.
13800 (UNSPEC_VPKUWUS): Likewise.
13801 (UNSPEC_VPKSWUS): Likewise.
13802 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
13803 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
13804 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
13805 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
13806 (UNSPEC_VUPKHSB): Likewise.
13807 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
13808 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
13809 (UNSPEC_VUPKHSH): Likewise.
13810 (UNSPEC_VUPKLSB): Likewise.
13811 (UNSPEC_VUPKLSH): Likewise.
13813 (VI_char): Likewise.
13814 (VI_scalar): Likewise.
13815 (VI_unit): Likewise.
13817 (VP_small): Likewise.
13818 (VP_small_lc): Likewise.
13819 (VU_char): Likewise.
13820 (add<mode>3): Likewise.
13821 (altivec_vaddcuw): Likewise.
13822 (altivec_vaddu<VI_char>s): Likewise.
13823 (altivec_vadds<VI_char>s): Likewise.
13824 (sub<mode>3): Likewise.
13825 (altivec_vsubcuw): Likewise.
13826 (altivec_vsubu<VI_char>s): Likewise.
13827 (altivec_vsubs<VI_char>s): Likewise.
13828 (altivec_vavgs<VI_char>): Likewise.
13829 (altivec_vcmpbfp): Likewise.
13830 (altivec_eq<mode>): Likewise.
13831 (altivec_gt<mode>): Likewise.
13832 (altivec_gtu<mode>): Likewise.
13833 (umax<mode>3): Likewise.
13834 (smax<mode>3): Likewise.
13835 (umin<mode>3): Likewise.
13836 (smin<mode>3): Likewise.
13837 (altivec_vpkuhum): Likewise.
13838 (altivec_vpkuwum): Likewise.
13839 (altivec_vpkshss): Likewise.
13840 (altivec_vpkswss): Likewise.
13841 (altivec_vpkuhus): Likewise.
13842 (altivec_vpkshus): Likewise.
13843 (altivec_vpkuwus): Likewise.
13844 (altivec_vpkswus): Likewise.
13845 (altivec_vpks<VI_char>ss): Likewise.
13846 (altivec_vpks<VI_char>us): Likewise.
13847 (altivec_vpku<VI_char>us): Likewise.
13848 (altivec_vpku<VI_char>um): Likewise.
13849 (altivec_vrl<VI_char>): Likewise.
13850 (altivec_vsl<VI_char>): Likewise.
13851 (altivec_vsr<VI_char>): Likewise.
13852 (altivec_vsra<VI_char>): Likewise.
13853 (altivec_vsldoi_<mode>): Likewise.
13854 (altivec_vupkhsb): Likewise.
13855 (altivec_vupkhs<VU_char>): Likewise.
13856 (altivec_vupkls<VU_char>): Likewise.
13857 (altivec_vupkhsh): Likewise.
13858 (altivec_vupklsb): Likewise.
13859 (altivec_vupklsh): Likewise.
13860 (altivec_vcmpequ<VI_char>_p): Likewise.
13861 (altivec_vcmpgts<VI_char>_p): Likewise.
13862 (altivec_vcmpgtu<VI_char>_p): Likewise.
13863 (abs<mode>2): Likewise.
13864 (vec_unpacks_hi_v16qi): Likewise.
13865 (vec_unpacks_hi_v8hi): Likewise.
13866 (vec_unpacks_lo_v16qi): Likewise.
13867 (vec_unpacks_hi_<VP_small_lc>): Likewise.
13868 (vec_unpacks_lo_v8hi): Likewise.
13869 (vec_unpacks_lo_<VP_small_lc>): Likewise.
13870 (vec_pack_trunc_v8h): Likewise.
13871 (vec_pack_trunc_v4si): Likewise.
13872 (vec_pack_trunc_<mode>): Likewise.
13874 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
13876 (vec_vmaxsd): Likewise.
13877 (vec_vmaxud): Likewise.
13878 (vec_vminsd): Likewise.
13879 (vec_vminud): Likewise.
13880 (vec_vpksdss): Likewise.
13881 (vec_vpksdus): Likewise.
13882 (vec_vpkudum): Likewise.
13883 (vec_vpkudus): Likewise.
13884 (vec_vrld): Likewise.
13885 (vec_vsld): Likewise.
13886 (vec_vsrad): Likewise.
13887 (vec_vsrd): Likewise.
13888 (vec_vsubudm): Likewise.
13889 (vec_vupkhsw): Likewise.
13890 (vec_vupklsw): Likewise.
13892 2013-05-29 Jan Hubicka <jh@suse.cz>
13894 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
13895 flags; reorder rest of fields in more consistent way.
13896 (varpool_node): Remove analyzed, finalized and alias.
13897 (cgraph_ndoe): Likewise.
13898 (symtab_alias_ultimate_target): New function.
13899 (cgraph_function_node): Move offline.
13900 (cgraph_reset_node): Declare.
13901 (cgraph_comdat_can_be_unshared_p): Remove.
13902 (varpool_remove_initializer): Declare.
13903 (varpool_first_defined_variable, varpool_next_defined_variable
13904 cgraph_first_defined_function, cgraph_next_defined_function): Update.
13905 (cgraph_function_with_gimple_body_p): Update.
13906 (varpool_all_refs_explicit_p): Update.
13907 (symtab_alias_target): New function.
13908 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
13909 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
13910 (cgraph_function_or_thunk_node): Simplify using
13911 symtab_alias_ultimate_target.
13912 (varpool_variable_node): Likewise.
13913 * cgraph.c (cgraph_create_function_alias): Update.
13914 (cgraph_add_thunk): Update.
13915 (cgraph_remove_node): Update.
13916 (dump_cgraph_node): Do not dump removed flags.
13917 (cgraph_function_body_availability): Update.
13918 (cgraph_propagate_frequency): Update.
13919 (verify_cgraph_node): Check sanity of local flag.
13920 (cgraph_function_node): Move here from cgraph.h; revamp for
13921 cgraph_function_or_thunk_node.
13922 * lto-symtab.c (lto_varpool_replace_node): Update.
13923 (lto_symtab_resolve_can_prevail_p): Update.
13924 (lto_symtab_merge_cgraph_nodes): Update.
13925 * ipa-cp.c (determine_versionability, initialize_node_lattices,
13926 propagate_constants_accross_call, devirtualization_time_bonus,
13927 ipcp_propagate_stage): Update.
13928 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
13929 * ipa-inline-transform.c (clone_inlined_nodes,
13930 preserve_function_body_p): Update.
13931 * ipa-reference.c (propagate): Update.
13932 (write_node_summary_p): Update.
13933 * toplev.c (wrapup_global_declaration_2): Update.
13934 * cgraphunit.c (cgraph_analyze_function): Rename to ...
13935 (analyze_function) ... this one.
13936 (cgraph_process_new_functions): Update.
13937 (cgraph_reset_node): Export.
13938 (cgraph_finalize_function): Update.
13939 (cgraph_add_new_function): Update.
13940 (process_function_and_variable_attributes): Update.
13941 (varpool_finalize_decl): Update.
13942 (symbol_finalized): Remove.
13943 (symbol_finalized_and_needed): Rename to ...
13944 (symbol_defined_and_needed): ... update.
13945 (cgraph_analyze_functions): Update.
13946 (handle_alias_pairs): Update.
13947 (mark_functions_to_output): Update.
13948 (assemble_thunk): Update.
13949 (output_in_order): Update.
13950 (output_weakrefs): Update.
13951 (finalize_compilation_unit): Update.
13952 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
13953 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
13954 input_node, input_varpool_node): Update.
13955 * dbxout.c (dbxout_expand_expr): Update.
13956 * cgraphclones.c (cgraph_clone_node): Update.
13957 (cgraph_copy_node_for_versioning): Update.
13958 (cgraph_materialize_clone): Update.
13959 (cgraph_materialize_all_clones): Update.
13960 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
13961 propagate_pure_const, propagate_nothrow): Update.
13962 * lto-streamer-out.c (lto_output, write_symbol): Update.
13963 * ipa-utils.c (ipa_reverse_postorder): Update.
13964 * ipa-inline.c (can_inline_edge_p): Update.
13965 (update_caller_keys, ipa_inline): Update.
13966 * dwarf2out.c (reference_to_unused,
13967 premark_types_used_by_global_vars_helper): Update.
13968 * tree-eh.c (tree_could_trap_p): Update.
13969 * ipa-split.c (consider_split, execute_split_functions): Update.
13970 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
13971 has_addr_references_p): Update; move ahead in file for better
13973 (process_references): Simplify.
13974 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
13975 bodies are removed.
13976 (cgraph_comdat_can_be_unshared_p): Make static.
13977 (cgraph_externally_visible_p): Update.
13978 (varpool_externally_visible_p): Update.
13979 (function_and_variable_visibility): Update.
13980 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
13981 ipa_tm_mark_force_output_node): Update.
13982 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
13983 estimate_edge_devirt_benefit, inline_generate_summary,
13984 inline_write_summary): Update.
13985 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
13986 * ipa-prop.c (ipa_compute_jump_functions): Update.
13987 (ipa_print_node_params, ipa_prop_read_section,
13988 ipa_update_after_lto_read, read_replacements_section): Update.
13989 * varasm.c (mark_decl_referenced): Update.
13990 (assemble_alias, dump_tm_clone_pairs): Update.
13991 * tree-inline.c (copy_bb): Update.
13992 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
13994 * symtab.c (dump_symtab_base): Print new flags.
13995 (verify_symtab_base): Verify new flags.
13996 (symtab_alias_ultimate_target): New function.
13997 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
13998 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
14000 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
14002 * i386.c (ix86_get_function_versions_dispatcher,
14003 ix86_generate_version_dispatcher_body): Update.
14004 (fold_builtin_cpu): Use varpool_add_new_variable.
14005 * varpool.c (varpool_remove_initializer): Break out from ...
14006 (varpool_remove_node): ... this one.
14007 (dump_varpool_node, varpool_node_for_asm,
14008 cgraph_variable_initializer_availability, varpool_analyze_node,
14009 varpool_assemble_decl, varpool_remove_unreferenced_decls,
14010 varpool_finalize_named_section_flags, varpool_create_variable_alias):
14013 2013-05-29 Jan Hubicka <jh@suse.cz>
14015 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
14017 2013-05-29 Easwaran Raman <eraman@google.com>
14019 PR tree-optimization/57442
14020 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
14021 when control exits the main loop.
14023 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
14025 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
14027 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
14028 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
14029 * rx/t-rx: Add rx100 under multi library matches option for nofpu
14032 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14034 PR tree-optimization/57441
14035 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
14036 Don't limit size of incr_vec to number of candidates.
14038 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
14040 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
14041 and mips16 directories.
14042 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
14043 (MULTILIB_DIRNAMES): Ditto.
14044 (MULTILIB_EXCEPTIONS): Add new exceptions.
14045 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
14046 (MULTILIB_DIRNAMES): Ditto.
14047 (MULTILIB_EXCEPTIONS): Add new exceptions.
14049 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14050 Marcus Shawcroft <marcus.shawcroft@arm.com>
14052 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
14053 SYMBOL_TINY_ABSOLUTE.
14054 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
14055 SYMBOL_TINY_ABSOLUTE.
14056 (aarch64_expand_mov_immediate): Likewise.
14057 (aarch64_classify_symbol): Likewise.
14058 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
14059 Permit SYMBOL_TINY_ABSOLUTE.
14060 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
14062 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14063 Marcus Shawcroft <marcus.shawcroft@arm.com>
14065 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
14066 Refactor if/switch. Replace gcc_assert with if.
14068 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14070 * config/i386/i386.c (initial_ix86_tune_features): Enable
14071 FP Reassociation for AMD bdver1 and bdver2.
14073 2013-05-29 Martin Jambor <mjambor@suse.cz>
14075 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
14076 and IMAGPART_EXPR do not occur within other handled_components.
14078 2013-05-29 Richard Biener <rguenther@suse.de>
14080 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
14081 access on whether the use is in the BB we currently try to
14083 (vect_bb_vectorization_profitable_p): Pass the BB we currently
14084 vectorize to vect_bb_slp_scalar_cost.
14086 2013-05-29 Richard Biener <rguenther@suse.de>
14088 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
14089 computing scalar cost offsetted by stmts that are kept live
14091 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
14092 for computation of scalar cost.
14094 2013-05-28 Steve Ellcey <sellcey@mips.com>
14096 * config/mips/mips-cpus.def (mips32r2): Change processor type.
14098 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
14100 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
14101 array notation built-in reduction functions.
14102 * doc/passes.texi (Passes): Added documentation about changes done
14104 * doc/invoke.texi (C Dialect Options): Added documentation about
14105 the -fcilkplus flag.
14106 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
14107 (BUILTINS_DEF): Depend on cilkplus.def.
14108 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
14109 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
14110 * cilkplus.def: New file.
14112 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
14114 PR rtl-optimization/57439
14115 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
14117 2013-05-28 Easwaran Raman <eraman@google.com>
14119 PR tree-optimization/57337
14120 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
14121 (find_insert_point): Correctly identify the insertion point
14122 when two statements with the same UID is compared.
14124 2013-05-28 Richard Biener <rguenther@suse.de>
14126 PR tree-optimization/56787
14127 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
14128 from the list of data references.
14129 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
14131 (vect_analyze_loop_operations): Likewise.
14132 (vect_transform_loop): Remove clobbers.
14134 2013-05-28 Martin Jambor <mjambor@suse.cz>
14136 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
14137 and REALPART_EXPRs have scalar type.
14139 2013-05-28 Richard Biener <rguenther@suse.de>
14141 PR tree-optimization/57411
14142 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
14144 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
14145 virtual operand propagation.
14147 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
14149 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
14150 destination register for bmasksi_vis.
14151 (vector_init_bshuffle): Likewise.
14152 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
14154 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
14156 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
14157 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
14158 mode if the instruction isn't available in the original mode.
14159 * config/sparc/sparc.opt (mfix-ut699): New option.
14160 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
14161 (divdf3): Turn into expander.
14162 (divdf3_nofix): New insn.
14163 (divdf3_fix): Likewise.
14164 (divsf3): Disable if -mfix-ut699.
14165 (sqrtdf2): Turn into expander.
14166 (sqrtdf2_nofix): New insn.
14167 (sqrtdf2_fix): Likewise.
14168 (sqrtsf2): Disable if -mfix-ut699.
14170 2013-05-27 Richard Biener <rguenther@suse.de>
14172 PR middle-end/57412
14173 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
14174 block for the new loop.
14176 2013-05-27 Richard Biener <rguenther@suse.de>
14178 PR tree-optimization/57343
14179 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
14180 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
14181 (number_of_iterations_cond): Do not build the folded tree.
14183 2013-05-27 Richard Biener <rguenther@suse.de>
14186 PR middle-end/57381
14187 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
14188 OEP_CONSTANT_ADDRESS_OF retained.
14190 PR tree-optimization/57417
14191 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
14192 for unchanged base.
14193 (set_ssa_val_to): Compare addresses using
14194 get_addr_base_and_unit_offset.
14196 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
14198 PR rtl-optimization/56833
14199 * postreload.c (move2add_record_mode): New function.
14200 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
14201 (move2add_use_add2_insn): Use move2add_record_sym_value.
14202 (move2add_use_add3_insn): Likewise.
14203 (reload_cse_move2add): Use move2add_valid_value_p and
14204 move2add_record_mode. Invalidate call-clobbered and REG_INC
14205 affected regs by setting reg_mode to VOIDmode.
14206 (move2add_note_store): Don't pretend the inside of a SUBREG is
14207 the actual destination. Invalidate single/leading registers by
14208 setting reg_mode to VOIDmode.
14209 Use move2add_record_sym_value, move2add_valid_value_p and
14210 move2add_record_mode.
14212 2013-05-27 Richard Biener <rguenther@suse.de>
14214 PR tree-optimization/57396
14215 * tree-affine.c (double_int_constant_multiple_p): Properly
14216 return false for val == 0 and div != 0.
14218 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14220 * config/mips/mips.h: Use #elif in preprocessor conditions.
14222 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14225 * config/mips/constraints.md (kl): New constraint.
14226 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
14227 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
14228 constraint for operand 0. Split after CSE for MIPS16. Emit a move
14229 from LO for MIPS16.
14230 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
14232 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14235 * config/mips/mips.c (mips_can_inline_p): New function.
14236 (TARGET_CAN_INLINE_P): Define.
14238 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
14240 * sched-int.h (ds_t, dw_t): Make unsigned int.
14241 Fix documentation that describes how all the ds_t bits are used.
14242 Reserve the last bit for delayed-branch scheduling.
14243 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
14244 (BITS_PER_DEP_WEAK): Fix definition and documentation.
14245 (gen_dep_weak_1): Remove prototype.
14246 * sched-deps.c (get_dep_weak_1): Make static.
14247 * target.def (speculate_insn, needs_block_p, gen_spec_check,
14248 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
14249 * doc/tm.texi: Regenerate.
14250 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
14252 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
14255 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
14257 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
14258 Sandra Loosemore <sandra@codesourcery.com>
14260 * config.gcc (powerpc-*): Allow native for with-cpu.
14262 2013-05-24 Jeff Law <law@redhat.com>
14264 PR tree-optimization/57124
14265 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
14266 conversion feeding a condition if the range has an overflow
14267 if -fstrict-overflow. Add warnings for when we do make the
14270 2013-05-24 Dehao Chen <dehao@google.com>
14272 * tree-cfg.c (locus_discrim_map): Fix the typo.
14273 (locus_discrim_hasher): Likewise.
14274 (locus_discrim_hasher::hash): Likewise.
14275 (locus_discrim_hasher::equal): Likewise.
14277 2013-05-24 Martin Jambor <mjambor@suse.cz>
14279 PR tree-optimization/57294
14280 * cgraph.h (ipa_record_stmt_references): Declare.
14281 * cgraphbuild.c (ipa_record_stmt_references): New function.
14282 (build_cgraph_edges): Use ipa_record_stmt_references.
14283 (rebuild_cgraph_edges): Likewise.
14284 (cgraph_rebuild_references): Likewise.
14285 * ipa-prop.c (ipa_modify_call_arguments): Discard references
14286 associated with the old statement and build references from the
14287 newly built statements.
14288 * ipa-ref.c (ipa_remove_stmt_references): New function.
14289 * ipa-ref.h (ipa_remove_stmt_references): Declare.
14291 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
14293 * lra-constraints.c (emit_spill_move): Use smaller mode for
14295 (check_and_process_move): Consider mem-reg moves for secondary
14297 (curr_insn_transform): Don't lose insns emitted before for
14298 secondary memory moves.
14299 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
14300 reg set up in the current insn.
14302 2013-05-24 Dehao Chen <dehao@google.com>
14304 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
14306 (locus_descrim_hasher::equal): Likewise.
14307 (build_gimple_cfg): New discriminator assignment algorithm.
14308 (make_edges): Likewise.
14309 (next_discriminator_for_locus): Likewise.
14310 (same_line_p): Likewise.
14311 (assign_discriminators): Likewise.
14312 (make_cond_expr_edges): Likewise.
14313 (make_gimple_switch_edges): Likewise.
14314 (make_goto_expr_edges): Likewise.
14315 (make_gimple_asm_edges): Likewise.
14317 2013-05-24 Ian Bolton <ian.bolton@arm.com>
14319 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
14320 X format specifier to only display bottom 16 bits.
14321 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
14322 immediate to match for operand 2, since it will be masked.
14324 2013-05-24 Richard Biener <rguenther@suse.de>
14326 PR tree-optimization/57287
14327 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
14328 all SSA names that occur in abnormal PHIs.
14330 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
14333 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
14334 that index is not negative.
14336 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
14338 PR rtl-optimization/55177
14339 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
14340 (simplify_byte_swapping_operation): New.
14341 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
14342 (simplify_relational_operation_1): Deal with BSWAP.
14344 2013-05-23 Richard Henderson <rth@redhat.com>
14347 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
14348 (ix86_reorg): Call it.
14350 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
14353 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
14354 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
14355 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
14357 2013-05-23 Christian Bruel <christian.bruel@st.com>
14360 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
14362 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14363 Marcus Shawcroft <marcus.shawcroft@arm.com>
14365 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
14366 * config/aarch64/constraints.md (Usa): Remove.
14367 * doc/md.texi (AArch64 Usa): Remove.
14369 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14370 Marcus Shawcroft <marcus.shawcroft@arm.com>
14372 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
14373 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
14374 * config/aarch64/predicates.md (aarch64_const_address): Remove.
14375 (aarch64_mov_operand): Use aarch64_mov_operand_p.
14377 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
14379 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
14380 instruction (AdvSIMD).
14381 * config/aarch64/aarch64-builtins.c
14382 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
14383 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
14385 2013-05-23 Martin Jambor <mjambor@suse.cz>
14387 PR middle-end/57347
14388 * tree.h (contains_bitfld_component_ref_p): Declare.
14389 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
14390 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
14392 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
14393 not access a bit-field. Assert all final offsets are byte-aligned.
14395 2013-05-23 Richard Biener <rguenther@suse.de>
14397 PR tree-optimization/57380
14398 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
14399 least one invariant or re-used load.
14400 * passes.c (init_optimization_passes): Move pass_phiprop before
14403 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
14405 * config/aarch64/aarch64-simd.md
14406 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
14408 2013-05-23 Richard Biener <rguenther@suse.de>
14410 PR middle-end/57381
14411 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
14412 OEP_CONSTANT_ADDRESS_OF retained.
14414 2013-05-23 Jakub Jelinek <jakub@redhat.com>
14416 PR middle-end/57344
14417 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
14418 don't lower unit. Handle unit not being always BITS_PER_WORD.
14420 2013-05-23 Richard Biener <rguenther@suse.de>
14422 PR rtl-optimization/57341
14423 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
14424 instead of true_dependence.
14426 2013-05-22 David Malcolm <dmalcolm@redhat.com>
14428 * bb-reorder.c (branch_threshold): Make const.
14429 (exec_threshold): Ditto.
14431 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14432 Pat Haugen <pthaugen@us.ibm.com>
14433 Peter Bergner <bergner@vnet.ibm.com>
14435 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
14436 documentation for the power8 crypto builtins.
14438 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
14440 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
14441 macros for defining power8 builtin functions.
14442 (BU_P8V_AV_2): Likewise.
14443 (BU_P8V_AV_P): Likewise.
14444 (BU_P8V_VSX_1): Likewise.
14445 (BU_P8V_OVERLOAD_1): Likewise.
14446 (BU_P8V_OVERLOAD_2): Likewise.
14447 (BU_CRYPTO_1): Likewise.
14448 (BU_CRYPTO_2): Likewise.
14449 (BU_CRYPTO_3): Likewise.
14450 (BU_CRYPTO_OVERLOAD_1): Likewise.
14451 (BU_CRYPTO_OVERLOAD_2): Likewise.
14452 (XSCVSPDP): Fix typo, point to the correct instruction.
14453 (VCIPHER): Add power8 crypto builtins.
14454 (VCIPHERLAST): Likewise.
14455 (VNCIPHER): Likewise.
14456 (VNCIPHERLAST): Likewise.
14457 (VPMSUMB): Likewise.
14458 (VPMSUMH): Likewise.
14459 (VPMSUMW): Likewise.
14460 (VPERMXOR_V2DI): Likewise.
14461 (VPERMXOR_V4SI: Likewise.
14462 (VPERMXOR_V8HI: Likewise.
14463 (VPERMXOR_V16QI: Likewise.
14464 (VSHASIGMAW): Likewise.
14465 (VSHASIGMAD): Likewise.
14466 (VPMSUM): Likewise.
14467 (VPERMXOR): Likewise.
14468 (VSHASIGMA): Likewise.
14470 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
14471 __CRYPTO__ if the crypto instructions are available.
14472 (altivec_overloaded_builtins): Add support for overloaded power8
14475 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
14476 support for power8 crypto builtins.
14477 (builtin_function_type): Likewise.
14478 (altivec_init_builtins): Add support for builtins that take vector
14479 long long (V2DI) arguments.
14481 * config/rs6000/crypto.md: New file, define power8 crypto
14484 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14485 Pat Haugen <pthaugen@us.ibm.com>
14486 Peter Bergner <bergner@vnet.ibm.com>
14488 * doc/invoke.texi (Option Summary): Add power8 options.
14489 (RS/6000 and PowerPC Options): Likewise.
14491 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
14492 constraints.md instead of rs6000.h. Reorder w* constraints. Add
14493 wm, wn, wr documentation.
14495 * config/rs6000/constraints.md (wm): New constraint for VSX
14496 registers if direct move instructions are enabled.
14497 (wn): New constraint for no registers.
14498 (wq): New constraint for quad word even GPR registers.
14499 (wr): New constraint if 64-bit instructions are enabled.
14500 (wv): New constraint if power8 vector instructions are enabled.
14501 (wQ): New constraint for quad word memory locations.
14503 * config/rs6000/predicates.md (const_0_to_15_operand): New
14504 constraint for 0..15 for crypto instructions.
14505 (gpc_reg_operand): If VSX allow registers in VSX registers as well
14506 as GPR and floating point registers.
14507 (int_reg_operand): New predicate to match only GPR registers.
14508 (base_reg_operand): New predicate to match base registers.
14509 (quad_int_reg_operand): New predicate to match even GPR registers
14510 for quad memory operations.
14511 (vsx_reg_or_cint_operand): New predicate to allow vector logical
14512 operations in both GPR and VSX registers.
14513 (quad_memory_operand): New predicate for quad memory operations.
14514 (reg_or_indexed_operand): New predicate for direct move support.
14516 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
14517 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
14518 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
14519 (POWERPC_MASKS): Add power8 options.
14520 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
14523 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
14524 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
14526 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
14527 (-mpower8-fusion): New power8 options.
14528 (-mpower8-fusion-sign): Likewise.
14529 (-mpower8-vector): Likewise.
14530 (-mcrypto): Likewise.
14531 (-mdirect-move): Likewise.
14532 (-mquad-memory): Likewise.
14534 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
14535 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
14537 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
14538 (rs6000_debug_vector_unit): Add p8_vector.
14539 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
14540 definitions. Also print fusion state.
14541 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
14542 (rs6000_builtin_mask_calculate): Add power8 builtin support.
14543 (rs6000_option_override_internal): Add support for power8.
14544 (rs6000_common_init_builtins): Add debugging for skipped builtins
14545 if -mdebug=builtin.
14546 (rs6000_adjust_cost): Add power8 support.
14547 (rs6000_issue_rate): Likewise.
14548 (insn_must_be_first_in_group): Likewise.
14549 (insn_must_be_last_in_group): Likewise.
14550 (force_new_group): Likewise.
14551 (rs6000_register_move_cost): Likewise.
14552 (rs6000_opt_masks): Likewise.
14554 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
14555 power8 capable assembler, default to power7 options.
14556 (TARGET_DIRECT_MOVE): Likewise.
14557 (TARGET_CRYPTO): Likewise.
14558 (TARGET_P8_VECTOR): Likewise.
14559 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
14560 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
14561 (VECTOR_MEM_P8_VECTOR_P): Likewise.
14562 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
14563 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
14564 (TARGET_XSCVDPSPN): Likewise.
14565 (TARGET_XSCVSPDPN): Likewsie.
14566 (TARGET_SYNC_HI_QI): Likewise.
14567 (TARGET_SYNC_TI): Likewise.
14568 (MASK_CRYPTO): Likewise.
14569 (MASK_DIRECT_MOVE): Likewise.
14570 (MASK_P8_FUSION): Likewise.
14571 (MASK_P8_VECTOR): Likewise.
14572 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
14573 temporary used by some of the direct move instructions to get two FP
14574 temporary registers does not force creation of a stack frame.
14575 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
14576 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
14577 that any VSX registers are tieable, even if they are also an
14578 Altivec vector mode.
14579 (r6000_reg_class_enum): Add wm, wr, wv constraints.
14580 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
14581 (RS6000_BTM_CRYPTO): Likewise.
14582 (RS6000_BTM_COMMON): Likewise.
14584 * config/rs6000/rs6000.md (cpu attribute): Add power8.
14585 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
14586 (enum rs6000_vector): Add power8 vector support.
14588 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14592 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
14593 (any_sibcall_could_use_r3): this and handle indirect calls.
14594 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
14596 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14598 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
14600 2013-05-22 Richard Biener <rguenther@suse.de>
14602 PR middle-end/57349
14603 * profile.c (branch_prob): Do not split blocks that are
14604 abnormally receiving from ECF_RETURNS_TWICE functions.
14606 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14608 * recog.c (offsettable_address_addr_space_p): Fix calculation of
14609 address mode. Move pointer mode initialization to the same place.
14611 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
14613 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
14614 while it has any effect.
14616 2013-05-21 Easwaran Raman <eraman@google.com>
14618 PR tree-optimization/57322
14619 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
14620 UID of the statement added to the BB to be 1.
14622 2013-05-21 Jakub Jelinek <jakub@redhat.com>
14624 PR tree-optimization/57331
14625 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
14626 of conversion from pointer type to integral type with integer.
14628 2013-05-21 Martin Jambor <mjambor@suse.cz>
14631 * ipa-prop.c (ipa_read_node_info): Process param_used and
14632 controlled_uses in the same order as when writing.
14634 2013-05-21 Magnus Granberg <baldrick@free.fr>
14637 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
14639 2013-05-21 Richard Biener <rguenther@suse.de>
14641 PR tree-optimization/57318
14642 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
14643 estimate stmts with side-effects as likely eliminated.
14645 2013-05-21 Richard Biener <rguenther@suse.de>
14647 PR tree-optimization/57330
14648 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
14649 preserve the call stmts fntype.
14651 2013-05-21 Richard Biener <rguenther@suse.de>
14653 PR tree-optimization/57303
14654 * tree-ssa-sink.c (statement_sink_location): Improve killing
14655 stmt detection and properly handle self-assignments.
14657 2013-05-21 Christian Bruel <christian.bruel@st.com>
14659 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
14660 spanning registers. LEAF_REG_REMAP is supported only for contiguous
14661 registers. Set register size out of the PARALLEL loop.
14663 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
14666 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
14667 (*fmasf4, *fmasf4_media): New insns.
14669 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14671 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
14672 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
14673 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
14674 (mips_idiv_insns): Update the comments to say that the returned
14675 instruction counts are in units of BASE_INSN_LENGTH.
14676 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
14677 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
14678 using 2 rather than 4 as the length of indirect MIPS16 and
14679 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
14680 length of a NOP. Don't divide MIPS16 lengths by 2.
14681 (mips16_split_long_branches): Assume a branch is long if the
14682 length is greater than 4 rather than 8.
14683 * config/mips/mips.md (length): Give MIPS16 lengths directly,
14684 rather than multiplying them by 2. Multiply instruction counts
14685 by BASE_INSN_LENGTH rather than 4.
14686 (*jump_mips16, tls_get_tp_mips16_<mode>)
14687 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
14689 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14691 * config/mips/mips.md (extended_mips16): Remove branch case.
14692 (length): Remove duplicated extended_mips16 test.
14694 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14696 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
14698 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
14700 * recog.h (Recog_data): Rename to...
14701 (recog_data_d): ...this.
14702 (recog_data): Update accordingly.
14703 * recog.c (recog_data): Likewise.
14704 * reload.c (save_recog_data): Likewise.
14705 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
14706 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
14708 2013-05-17 Julian Brown <julian@codesourcery.com>
14710 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
14711 found in a REG_EQUAL note, invalidate it.
14713 2013-05-17 Easwaran Raman <eraman@google.com>
14715 * tree-ssa-reassoc.c (find_insert_point): New function.
14716 (insert_stmt_after): Likewise.
14717 (get_def_stmt): Likewise.
14718 (ensure_ops_are_available): Likewise.
14719 (not_dominated_by): Likewise.
14720 (rewrite_expr_tree): Do not move statements beyond what is
14721 necessary. Remove call to swap_ops_for_binary_stmt...
14722 (reassociate_bb): ... and move it here.
14723 (build_and_add_sum): Assign UIDs for new statements.
14724 (linearize_expr): Likewise.
14725 (do_reassoc): Renumber gimple statement UIDs.
14727 2013-05-17 Jan Hubicka <jh@suse.cz>
14729 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
14731 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
14732 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
14734 (output_weakrefs): Update.
14736 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
14737 Martin Jambor <mjambor@suse.cz>
14739 PR middle-end/57276
14740 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
14741 value that corresponds to the given aggval is found in values vector.
14743 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
14745 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
14746 sse, sse2, sse3, ssse3 and sse4a flags to options.
14748 2013-05-17 David Malcolm <dmalcolm@redhat.com>
14750 * gengtype-state.c: (s_expr_writer): New class, to handle
14751 prettifying of output layout of s-expressions.
14752 (state_writer): New class, to write out gtype.state.
14753 (state_written_type_count): Move this variable into member data of
14755 (s_expr_writer::s_expr_writer): New code: constructor for new class
14756 (state_writer::state_writer(): ditto
14757 (s_expr_writer::write_new_line): New function
14758 (s_expr_writer::write_any_indent): ditto
14759 (s_expr_writer::begin_s_expr): ditto
14760 (s_expr_writer::end_s_expr): ditto
14761 (write_state_fileloc): convert to method of state_writer...
14762 (state_writer:: write_state_fileloc): ...and use methods of
14763 s_expr_writer to write indentation into the gtype.state output file
14764 to visually represent the hierarchical structure of the list
14766 (write_state_fields): ditto, renaming to...
14767 (state_writer::write_state_fields)
14768 (write_state_a_string): ditto, renaming to...
14769 (state_writer::write_state_a_string)
14770 (write_state_string_option): ditto, renaming to...
14771 (state_writer::write_state_string_option)
14772 (write_state_type_option): ditto, renaming to...
14773 (state_writer::write_state_type_option)
14774 (write_state_nested_option): ditto, renaming to...
14775 (state_writer::write_state_nested_option)
14776 (write_state_option): ditto, renaming to...
14777 (state_writer::write_state_option)
14778 (write_state_options): ditto, renaming to...
14779 (state_writer::write_state_options)
14780 (write_state_lang_bitmap): ditto, renaming to...
14781 (state_writer::write_state_lang_bitmap)
14782 (write_state_version): ditto, renaming to...
14783 (state_writer::write_state_version)
14784 (write_state_scalar_type): ditto, renaming to...
14785 (state_writer::write_state_scalar_type)
14786 (write_state_string_type): ditto, renaming to...
14787 (state_writer::write_state_string_type)
14788 (write_state_undefined_type): ditto, renaming to...
14789 (state_writer::write_state_undefined_type)
14790 (write_state_struct_union_type): ditto, renaming to...
14791 (state_writer::write_state_struct_union_type)
14792 (write_state_struct_type): ditto, renaming to...
14793 (state_writer::write_state_struct_type)
14794 (write_state_user_struct_type): ditto, renaming to...
14795 (state_writer::write_state_user_struct_type)
14796 (write_state_lang_struct_type): ditto, renaming to...
14797 (state_writer::write_state_lang_struct_type)
14798 (write_state_param_struct_type): ditto, renaming to...
14799 (state_writer::write_state_param_struct_type)
14800 (write_state_pointer_type): ditto, renaming to...
14801 (state_writer::write_state_pointer_type)
14802 (write_state_array_type): ditto, renaming to...
14803 (state_writer::write_state_array_type)
14804 (write_state_gc_used): ditto, renaming to...
14805 (state_writer::write_state_gc_used)
14806 (write_state_common_type_content): ditto, renaming to...
14807 (state_writer::write_state_common_type_content)
14808 (write_state_type): ditto, renaming to...
14809 (state_writer::write_state_type)
14810 (write_state_pair_list): ditto, renaming to...
14811 (state_writer::write_state_pair_list)
14812 (write_state_pair): ditto, renaming to...
14813 (state_writer::write_state_pair)
14814 (write_state_typedefs): ditto, renaming to...
14815 (state_writer::write_state_typedefs)
14816 (write_state_structures): ditto, renaming to...
14817 (state_writer::write_state_structures)
14818 (write_state_param_structs): ditto, renaming to...
14819 (state_writer::write_state_param_structs)
14820 (write_state_variables): ditto, renaming to...
14821 (state_writer::write_state_variables)
14822 (write_state_srcdir): ditto, renaming to...
14823 (state_writer::write_state_srcdir)
14824 (write_state_files_list): ditto, renaming to...
14825 (state_writer::write_state_files_list)
14826 (write_state_languages): ditto, renaming to...
14827 (state_writer::write_state_languages)
14828 (write_state): create a state_writer instance and use it when
14829 writing out the state file
14831 2013-05-17 Mike Stump <mikestump@comcast.net>
14833 PR rtl-optimization/57304
14834 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
14835 accessing DF_REF_REAL_LOC.
14837 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14839 PR rtl-optimization/57281
14840 PR rtl-optimization/57300
14841 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
14842 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
14843 what the other splitter did if the registers are dead.
14845 2013-05-17 Richard Biener <rguenther@suse.de>
14847 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
14850 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14852 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
14855 2013-05-17 Marek Polacek <polacek@redhat.com>
14857 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
14858 length when doing non-zero store of storing '\0' to '\0'.
14860 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14862 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
14863 vect_external_def oprnd1 with loop_vinfo, try to emit
14864 optional cast, negation and and stmts on the loop preheader
14865 edge instead of into the pattern def seq.
14867 PR tree-optimization/57051
14868 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
14869 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
14871 2013-05-16 Nick Clifton <nickc@redhat.com>
14873 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
14874 (rl78_is_naked_func): New function.
14875 (rl78_expand_prologue): Skip prologue generation for naked functions.
14876 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
14877 * doc/extend.texi (naked): Add RL78 to the list of processors
14878 that supports this attribute.
14880 2013-05-16 Jeff Law <law@redhat.com>
14882 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
14884 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
14886 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
14887 cache parameters using detect_caches_amd also for CYRIX,
14888 NSC and TM2 signatures.
14890 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
14891 Dzianis Kahanovich <mahatma@eu.by>
14895 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
14896 VIA/Centaur processors and determine their cache parameters
14897 using detect_caches_amd.
14899 2013-05-16 Teresa Johnson <tejohnson@google.com>
14901 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
14902 (rtl_verify_edges): New function.
14903 (rtl_verify_bb_insns): Ditto.
14904 (rtl_verify_bb_pointers): Ditto.
14905 (rtl_verify_bb_insn_chain): Ditto.
14906 (rtl_verify_fallthru): Ditto.
14907 (rtl_verify_bb_layout): Ditto.
14908 (rtl_verify_flow_info_1): Outline checks into new functions.
14909 (rtl_verify_flow_info): Ditto.
14911 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
14913 * cfghooks.c (copy_bbs): Add update_dominance argument.
14914 * cfghooks.h (copy_bbs): Update prototype.
14915 * tree-cfg.c (gimple_duplicate_sese_region):
14916 Add update_dominance argument.
14917 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
14918 * tree-ssa-loop-ch.c (copy_loop_headers): Update
14919 gimple_duplicate_sese_region call.
14920 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
14921 Update copy_bbs call.
14922 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
14923 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
14925 2013-05-16 Jakub Jelinek <jakub@redhat.com>
14927 * tree-vectorizer.h (NUM_PATTERNS): Increment.
14928 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
14929 vect_recog_rotate_pattern.
14930 (vect_recog_rotate_pattern): New function.
14932 2013-05-16 Jason Merrill <jason@redhat.com>
14934 * Makefile.in (LLINKER): New variable.
14935 (mostlyclean): Remove link mutex.
14936 * configure.ac: Handle --enable-link-mutex.
14937 * lock-and-run.sh: New script.
14939 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14942 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
14945 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14947 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
14949 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
14951 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
14952 * config/arm/arm.c (next_consecutive_mem): New function.
14953 (gen_movmem_ldrd_strd): Likewise.
14954 * config/arm/arm.md (movmemqi): Update condition and code.
14955 (unaligned_loaddi, unaligned_storedi): New patterns.
14957 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14959 * config.gcc: Obsolete *-*-solaris2.9*.
14960 * doc/install.texi (Specific, *-*-solaris2*): Document it.
14962 2013-05-16 Richard Biener <rguenther@suse.de>
14964 * passes.c (init_optimization_passes): Move pass_parallelize_loops
14965 earlier, after GRAPHITE transforms and IV canonicalization.
14967 2013-05-16 Jakub Jelinek <jakub@redhat.com>
14969 * omp-low.c (extract_omp_for_data): For collapsed loops,
14970 if at least one of the loops is known at compile time to
14971 iterate zero times, set count to 0.
14972 (expand_omp_regimplify_p): New function.
14973 (expand_omp_for_generic): For collapsed loops, if at least
14974 one of the loops isn't known to iterate at least once,
14975 add runtime check with setting count to 0.
14976 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
14977 For unsigned types if it isn't known at compile time that
14978 the loop will iterate at least once, add runtime check to bypass
14979 the whole loop if initial condition isn't true.
14981 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
14983 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
14985 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
14987 PR middle-end/57286
14988 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
14989 transformations to avoid an infinite loop.
14991 2013-05-16 Marek Polacek <polacek@redhat.com>
14993 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
14995 2013-05-15 Leif Ekblad <leif@rdos.net>
14997 * config/i386/i386.c (ix86_decompose_address): Use
14998 DEFAULT_TLS_SEG_REG to access TLS segment register.
14999 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
15000 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
15001 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
15003 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
15006 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
15007 sibling calls to functions that would normally be lazily bound,
15008 unless $gp is call-clobbered.
15010 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
15012 * config/i386/i386.c (ix86_option_override_internal): Update
15013 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
15014 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
15015 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
15016 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
15018 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
15020 2013-05-15 Andreas Schwab <schwab@suse.de>
15022 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
15023 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
15026 2013-05-15 Teresa Johnson <tejohnson@google.com>
15028 * loop-unroll.c (report_unroll_peel): Check decision before
15029 emitting unroll/peel message.
15031 2013-05-15 Teresa Johnson <tejohnson@google.com>
15033 * function.h (has_bb_partition): New rtl_data flag.
15034 (bb_reorder_complete): Ditto.
15035 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
15036 instead of flag_reorder_blocks_and_partition.
15037 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
15038 with some enhancements.
15039 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
15040 * bb-reorder.c (connect_traces): Check for has_bb_partition
15041 instead of flag_reorder_blocks_and_partition.
15042 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
15043 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
15044 verify_hot_cold_block_grouping.
15045 (partition_hot_cold_basic_blocks): Set has_bb_partition.
15047 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15050 * config/arm/predicates.md (call_insn_operand): New predicate.
15051 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
15052 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
15053 if insn is not a tail call.
15054 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
15056 * config/arm/arm.h (enum reg_class): New caller save register class.
15057 (REG_CLASS_NAMES): Likewise.
15058 (REG_CLASS_CONTENTS): Likewise.
15059 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
15062 2013-05-15 Richard Biener <rguenther@suse.de>
15064 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
15065 of MSG_OPTIMIZED_LOCATIONS.
15066 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
15067 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
15069 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
15070 of MSG_OPTIMIZED_LOCATIONS.
15071 (execute_vect_slp): Likewise.
15072 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
15073 (vect_create_cond_for_alias_checks): Likewise.
15074 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
15075 (vect_recog_widen_mult_pattern): Likewise.
15076 (vect_recog_widen_sum_pattern): Likewise.
15077 (vect_recog_over_widening_pattern): Likewise.
15078 (vect_recog_widen_shift_pattern): Likewise.
15079 (vect_recog_vector_vector_shift_pattern): Likewise.
15080 (vect_recog_divmod_pattern): Likewise.
15081 (vect_recog_mixed_size_cond_pattern): Likewise.
15082 (vect_recog_bool_pattern): Likewise.
15083 (vect_pattern_recog_1): Likewise.
15085 2013-05-15 Martin Jambor <mjambor@suse.cz>
15087 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
15088 non-functions to builtin_unreachable.
15089 * ipa-inline-transform.c (inline_call): Do not assert estimates were
15090 correct when new direct edges were discovered.
15092 2013-05-15 Martin Jambor <mjambor@suse.cz>
15094 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
15095 header, print symbol order instead of node uid, print more information
15096 about indirect edge targets.
15097 (ipa_make_edge_direct_to_target): Print symbol order instead of node
15099 (ipa_make_edge_direct_to_target): Likewise.
15100 (remove_described_reference): Likewise.
15101 (propagate_controlled_uses): Likewise.
15102 (ipa_print_node_params): Also print symbol order.
15103 (ipcp_transform_function): Print symbol order instead of node uids.
15104 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
15105 (cgraph_get_create_real_symbol_node): Likewise.
15106 * ipa-cp.c (print_lattice): Likewise.
15107 (print_all_lattices): Likewise.
15108 (determine_versionability): Likewise.
15109 (initialize_node_lattices): Likewise.
15110 (estimate_local_effects): Likewise.
15111 (update_profiling_info): Likewise.
15112 (create_specialized_node): Likewise.
15113 (perhaps_add_new_callers): Likewise.
15114 (decide_about_value): Likewise.
15115 (decide_whether_version_node): Likewise.
15116 (identify_dead_nodes): Likewise.
15117 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
15118 (dump_inline_summary): Likewise.
15119 (estimate_node_size_and_time): Likewise.
15120 (inline_analyze_function): Likewise.
15121 * ipa-inline.c (report_inline_failed_reason): Likewise.
15122 (want_early_inline_function_p): Likewise.
15123 (edge_badness): Likewise.
15124 (update_edge_key): Likewise.
15125 (inline_small_functions): Likewise. Add dumping of order to two other
15127 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
15128 instead of node uids.
15129 (propagate_pure_const): Likewise.
15130 (propagate_pure_const): Likewise.
15131 * ipa-utils.c (dump_cgraph_node_set): Likewise.
15132 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
15133 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
15135 * tree-pretty-print.c (dump_function_header): Likewise.
15136 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
15137 Print symbol order instead of node uids.
15139 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15141 * config/s390/s390.c (s390_register_move_cost): Don't impose the
15142 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
15144 2013-05-15 Richard Biener <rguenther@suse.de>
15146 PR tree-optimization/57275
15147 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
15148 return value for fail to do runtime alias checks for gather loads.
15150 2013-05-15 Jan Hubicka <jh@suse.cz>
15154 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
15155 weakrefs are not external.
15156 (lto_symtab_merge_decls): Fix thinko when dealing with
15157 non-lto_symtab decls.
15158 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
15159 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
15160 * varpool.c (dump_varpool_node): Dump more flags.
15162 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15164 * config/i386/i386.c (processor_alias_table): Add instruction
15165 FSGSBASE for AMD bdver3 architecture.
15167 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15169 * tree.c (warn_deprecated_use): Print file:line using locus color.
15170 * diagnostic.c (diagnostic_report_current_module): Print file:line
15171 and file:line:column using locus color.
15173 2013-05-14 Mike Stump <mikestump@comcast.net>
15175 * gdbinit.in: Add __null.
15177 2013-05-14 Mike Stump <mikestump@comcast.net>
15179 * recog.h: Rename struct recog_data to Recog_data.
15180 * recog.c: Likewise.
15181 * reload.c (can_reload_into): Likewise.
15182 * config/picochip/picochip.c: Likewise.
15184 2013-05-14 Mike Stump <mikestump@comcast.net>
15186 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
15188 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
15190 * resource.h (struct resources): Remove unch_memory member.
15191 (CLEAR_RESOURCE): Don't clear unch_memory.
15192 * resource.c (mark_referenced_resources): Don't set it.
15193 (mark_set_resources): Likewise.
15194 (mark_target_live_regs): Don't clear it.
15195 (init_resource_info): Likewise.
15196 * reorg.c (resource_conflicts_p): Don't compare it.
15197 (redundant_insn): Don't set it.
15199 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
15201 * emit-rtl.c (next_label): Remove unused function.
15202 (skip_consecutive_labels, link_cc0_insns): Move to ...
15203 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
15204 only place where these functions are used, and make them static.
15206 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
15208 * fold-const.c (fold_negate_expr): Handle vectors.
15209 (fold_truth_not_expr): Make it static.
15210 (fold_invert_truthvalue): New static function.
15211 (invert_truthvalue_loc): Handle vectors. Do not call
15212 fold_truth_not_expr directly.
15213 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
15214 <TRUTH_NOT_EXPR>: Do not cast to boolean.
15215 (fold_comparison): Handle vector constants.
15216 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
15217 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
15218 * tree.h (fold_truth_not_expr): Remove declaration.
15220 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
15222 * config/aarch64/aarch64-simd.md
15223 (aarch64_vcond_internal<mode>): Rename to...
15224 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
15225 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
15226 float modes. Clarify all iterator modes.
15227 (vcond<mode><mode>): Use new name for vcond expanders.
15228 (vcond<v_cmp_result><mode>): Likewise.
15229 (vcondu<mode><mode>: Likewise.
15230 * config/aarch64/iterators.md (VDQF_COND): New.
15232 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
15235 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
15236 variable for the shift amount. Check that we shift by non-negative
15239 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
15242 * config/arm/arm.h (EPILOGUE_USES): Only return true
15243 for LR_REGNUM after epilogue_completed.
15245 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
15247 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
15248 is error_mark_node.
15250 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15253 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
15254 and Solaris 11+/x86 with gld.
15255 * configure: Regenerate.
15257 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15259 * expmed.c (expand_shift_1): Canonicalize rotates by
15260 constant bitsize / 2 to bitsize - 1.
15261 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
15262 case ROTATERT>: Likewise.
15265 2013-05-10 Jakub Jelinek <jakub@redhat.com>
15267 * config/i386/i386.md (rotateinv): New code attr.
15268 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
15269 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
15270 roll $31, %eax, etc.
15272 2013-05-14 Richard Biener <rguenther@suse.de>
15274 PR middle-end/57235
15275 * tree-eh.c (sink_clobbers): Give up for successors with
15276 multiple predecessors and no virtual uses.
15278 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
15280 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
15281 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
15283 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15285 PR middle-end/57251
15286 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
15287 the case when both op0 and op1 have VOIDmode.
15289 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15291 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
15292 in multiply-accumulate mode.
15294 2013-05-13 Guozhi Wei <carrot@google.com>
15296 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
15298 2013-05-13 Kai Tietz <ktietz@redhat.com>
15301 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
15302 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
15303 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
15304 * config/i386/i386.c (ix86_option_override_internal): Likewise.
15305 (ix86_expand_prologue): Likewise.
15306 (ix86_expand_split_stack_prologue): Likewise.
15307 (legitimate_pic_address_disp_p): Likewise.
15308 (legitimize_pic_address): Likewise.
15309 (legitimize_tls_address): Likewise.
15310 (legitimize_pe_coff_symbol): Likewise.
15311 (output_pic_addr_const): Likewise.
15312 (construct_plt_address): Likewise.
15313 (ix86_expand_call): Likewise.
15314 (x86_output_mi_thunk): Likewise.
15315 (x86_function_profiler): Likewise.
15317 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
15319 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
15320 similar switch cases.
15321 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
15322 (aarch64_simd_mov_to_<mode>low): Delete.
15323 (aarch64_simd_mov_to_<mode>high): Delete.
15324 (move_lo_quad_<mode>): Add w<-r alternative.
15325 (aarch64_simd_move_hi_quad_<mode>): Likewise.
15326 (aarch64_simd_mov_from_*): Update type attribute.
15327 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
15330 2013-05-13 Jan Hubicka <jh@suse.cz>
15332 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
15333 * config/i386/i386.c (ix86_compute_frame_layout,
15334 ix86_expand_epilogue, emit_i387_cw_initialization,
15335 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
15336 ix86_local_alignment): Fix use of size/speed predicates.
15338 2013-05-13 Jakub Jelinek <jakub@redhat.com>
15340 PR tree-optimization/45216
15341 PR tree-optimization/57157
15342 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
15343 the (-Y) & (B - 1) variant if OP is |.
15344 * expmed.c (expand_shift_1): For rotations by const0_rtx just
15345 return shifted. Use (-op1) & (prec - 1) as other_amount
15346 instead of prec - op1.
15348 2013-05-13 Martin Jambor <mjambor@suse.cz>
15350 PR middle-end/42371
15351 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
15352 (ipa_constant_data): New type.
15353 (ipa_jump_func): Use ipa_constant_data to hold information about
15354 constant jump functions.
15355 (ipa_get_jf_constant): Adjust to jump function type changes.
15356 (ipa_get_jf_constant_rdesc): New function.
15357 (ipa_param_descriptor): New field controlled_uses.
15358 (ipa_get_controlled_uses): New function.
15359 (ipa_set_controlled_uses): Likewise.
15360 * ipa-ref.h (ipa_find_reference): Declare.
15361 * ipa-prop.c (ipa_cst_ref_desc): New type.
15362 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
15364 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
15365 New parameter cs. Adjust all callers.
15366 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
15367 (remove_described_reference): New function.
15368 (jfunc_rdesc_usable): Likewise.
15369 (try_make_edge_direct_simple_call): Decrement controlled use count,
15370 attempt to remove reference if it hits zero.
15371 (combine_controlled_uses_counters): New function.
15372 (propagate_controlled_uses): Likewise.
15373 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
15374 (ipa_edge_duplication_hook): Duplicate reference descriptions.
15375 (ipa_print_node_params): Print described use counter.
15376 (ipa_write_jump_function): Adjust to jump function type changes.
15377 (ipa_read_jump_function): New parameter CS, pass it to
15378 ipa_set_jf_constant. Adjust caller.
15379 (ipa_write_node_info): Stream controlled use count
15380 (ipa_read_node_info): Likewise.
15381 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
15383 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
15384 count. Remove cloning-added reference if it reaches zero.
15385 * ipa-ref.c (ipa_find_reference): New function.
15387 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15389 * config/i386/i386.c (processor_target_table): Modified default
15390 alignment values for AMD BD and BT architectures.
15392 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
15394 * tree-vect-generic.c (uniform_vector_p): Move ...
15395 * tree.c (uniform_vector_p): ... here.
15396 * tree.h (uniform_vector_p): Declare it.
15397 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
15400 2013-05-13 Jakub Jelinek <jakub@redhat.com>
15402 PR tree-optimization/57230
15403 * tree-ssa-strlen.c (handle_char_store): Record length for
15404 array store from STRING_CST.
15406 PR tree-optimization/57230
15407 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
15410 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
15412 * config/epiphany/epiphany.c (epiphany_init): Check size of
15413 NUM_MODES_FOR_MODE_SWITCHING.
15414 (epiphany_expand_prologue):
15415 Remove CONFIG_REGNUM initial value handling code.
15416 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
15417 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
15418 (emit_set_fp_mode, epiphany_mode_after): Likewise.
15419 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
15420 Don't return 1 for FP_MODE_NONE.
15421 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
15422 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
15423 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
15424 * config/epiphany/epiphany.md (save_config): New pattern.
15426 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
15428 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
15430 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
15432 * config/i386/i386.md (memory): Handle sseishft1.
15433 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
15434 (*vec_extractv2di_1): Ditto.
15436 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
15438 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
15441 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
15443 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
15444 Add mthumb/march=armv7-a multilib.
15445 Add mthumb/march=armv7-r multilib.
15446 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
15448 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
15450 * config/v850/t-rtems: Add more multilibs.
15452 2013-05-10 Richard Biener <rguenther@suse.de>
15454 PR tree-optimization/57214
15455 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
15456 not propagate from SSA names that occur in abnormal PHI nodes.
15458 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
15460 * stor-layout.c (element_precision): New function.
15461 * machmode.h (element_precision): Declare it.
15462 * tree.c (build_minus_one_cst): New function.
15463 (element_precision): Likewise.
15464 * tree.h (build_minus_one_cst): Declare new function.
15465 (element_precision): Likewise.
15466 * fold-const.c (operand_equal_p): Use element_precision.
15467 (fold_binary_loc): Handle vector types.
15468 * convert.c (convert_to_integer): Use element_precision.
15469 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
15472 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
15474 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
15475 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
15476 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
15477 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
15478 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
15479 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
15480 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
15481 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
15482 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
15483 (Uuw8): New constraints.
15484 (Usb4): Move into alphabetical order.
15485 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
15486 (sd8_operand, ub8_operand, uw8_operand): New predicates.
15487 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
15488 previously unnamed patterns.
15489 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
15490 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
15491 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
15492 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
15493 of set_attr_alternative/if_then_else. Use extended_mips16 instead
15494 of specific lengths.
15496 2013-05-10 Jakub Jelinek <jakub@redhat.com>
15498 * config/i386/i386.md (rotateinv): New code attr.
15499 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
15500 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
15501 roll $31, %eax, etc.
15503 PR tree-optimization/45216
15504 PR tree-optimization/57157
15505 * tree-ssa-forwprop.c (simplify_rotate): New function.
15506 (ssa_forward_propagate_and_combine): Call it.
15508 2013-05-10 Richard Biener <rguenther@suse.de>
15510 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
15511 disable peeling when we version for aliasing.
15512 (vector_alignment_reachable_p): Honor explicit user alignment.
15513 (vect_supportable_dr_alignment): Likewise.
15514 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
15515 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
15516 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
15517 then peeling to arrange for the cost-model check to come first.
15519 2013-05-10 Alan Modra <amodra@gmail.com>
15521 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
15522 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
15523 * configure: Regenerate.
15525 2013-05-10 Alan Modra <amodra@gmail.com>
15528 * varasm.c (default_elf_select_section): Move !DECL_P check..
15529 (get_named_section): ..to here before calling get_section_name.
15531 (default_section_type_flags): Add DECL_P check.
15532 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
15533 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
15535 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
15537 * config/epiphany/epiphany.c (epiphany_expand_prologue):
15538 When using gen_stack_adjust_str with a register offset, add a
15539 REG_FRAME_RELATED_EXPR note.
15541 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
15543 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
15544 (*vec_extractv4si_zext_mem): Ditto.
15545 (*vec_extractv2di): Add 0->x and x->x alternatives.
15546 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
15547 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
15549 2013-05-09 Jason Merrill <jason@redhat.com>
15551 N3639 C++1y VLA support
15552 * gimplify.c (gimplify_vla_decl): Don't touch an existing
15555 * tree.c (build_constructor_va): New.
15556 * tree.h: Declare it.
15558 2013-05-09 Martin Jambor <mjambor@suse.cz>
15561 * gimple-fold.c (canonicalize_constructor_val): Call
15562 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
15564 2013-05-09 Jan Hubicka <jh@suse.cz>
15565 Richard Biener <rguenther@suse.de>
15568 * symtab.c (symtab_make_decl_local): Do not add private names.
15570 2013-05-09 Jan Hubicka <jh@suse.cz>
15573 * symtab.c (insert_to_assembler_name_hash): Handle clones.
15574 (unlink_from_assembler_name_hash): Likewise.
15575 (symtab_prevail_in_asm_name_hash, symtab_register_node,
15576 symtab_unregister_node, symtab_initialize_asm_name_hash,
15577 change_decl_assembler_name): Update.
15579 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
15581 * config/aarch64/aarch64.md: New movtf split.
15582 (*movtf_aarch64): Update.
15583 (aarch64_movdi_tilow): Handle TF modes and rename to
15584 aarch64_movdi_<mode>low.
15585 (aarch64_movdi_tihigh): Handle TF modes and rename to
15586 aarch64_movdi_<mode>high
15587 (aarch64_movtihigh_di): Handle TF modes and rename to
15588 aarch64_mov<mode>high_di
15589 (aarch64_movtilow_di): Handle TF modes and rename to
15590 aarch64_mov<mode>low_di
15591 (aarch64_movtilow_tilow): Remove spurious whitespace.
15592 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
15594 (aarch64_print_operand): Update.
15596 2013-05-09 Alan Modra <amodra@gmail.com>
15598 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
15600 * configure: Regenerate.
15602 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
15604 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
15605 splitter preparation statements.
15606 * config/i386/sse.md (*vec_extract* splitters): Ditto.
15607 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
15610 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15612 * gimple-ssa-strength-reduction.c (count_candidates): Change
15613 return value to int.
15614 (analyze_candidates_and_replace): Change type of length to int.
15616 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
15618 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
15619 (*vec_extract<mode>): Use VI12_128 mode iterator.
15620 (*vec_extract<mode>_mem): Ditto.
15621 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
15624 2013-05-08 Diego Novillo <dnovillo@google.com>
15629 2012-08-17 Diego Novillo <dnovillo@google.com>
15632 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
15633 * config.in: Regenerate.
15634 * configure: Regenerate.
15635 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
15637 2013-05-08 Jan Hubicka <jh@suse.cz>
15640 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
15641 * cgraph.h (symtab_node_base): Add unique_name.
15642 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15643 input_overwrite_node, input_varpool_node): Stream unique_name.
15644 * cgraphclones.c (cgraph_create_virtual_clone,
15645 cgraph_function_versioning): Set unique_name.
15646 * ipa.c (function_and_variable_visibility): Set unique_name.
15648 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15650 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
15651 (alloc_cand_and_find_basis): Restrict conditional candidate
15652 processing to CAND_MULTs.
15654 2013-05-08 Jan Hubicka <jh@suse.cz>
15657 lto-symtab.c (lto_symtab_symbol_p): New function.
15658 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
15659 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
15660 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
15661 Skip static symbols.
15663 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
15665 PR tree-optimization/57200
15666 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
15667 Only call inform if the preceding warning_at returns true.
15669 2013-05-07 Han Shen <shenhan@google.com>
15671 * cfgexpand.c (record_or_union_type_has_array_p): New function.
15672 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
15673 * common.opt (fstack-protector-strong): New option.
15674 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
15675 * doc/invoke.texi (Optimization Options): Document
15676 "-fstack-protector-strong".
15677 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
15679 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
15681 * config/mips/mips.c (mips_machine_reorg2): Return 0.
15683 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
15685 * ira.c (update_equiv_regs): Add insn having equiv memory even if
15686 it is not lhs of the insn.
15687 (setup_reg_equiv): Remove insn having equiv memory which it is not
15689 * lra-constraints.c (process_address): Try to improve generation
15690 code for address base + disp.
15691 (lra_constraints): Make correct the code for checking insn setting
15692 up backward equivalence. Remove insn only if it is in the init
15694 * lra-eliminations.c (update_reg_eliminate): Change return value.
15695 (lra_eliminate): Use the result.
15697 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
15699 * config/i386/sse.md (ssescalarnummask): New mode attribute.
15700 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
15701 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
15702 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
15703 register target operands.
15704 (*vec_extractv8hi_sse2): New pattern.
15705 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
15706 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
15707 (*vec_extract<mode>_mem): New insn and split pattern.
15709 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
15711 * config/arm/arm.c (arm_asan_shadow_offset): New function.
15712 (TARGET_ASAN_SHADOW_OFFSET): Define.
15713 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
15714 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
15716 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15718 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
15719 (incr_vec_index): Return -1 if increment not found.
15720 (create_add_on_incoming_edge): Assert if increment not found.
15721 (record_increment): Limit number of increments recorded.
15722 (all_phi_incrs_profitable): Return false if an increment not found.
15723 (replace_profitable_candidates): Don't process increments that were
15725 (analyze_candidates_and_replace): Limit size of incr_vec.
15727 2013-05-07 Richard Biener <rguenther@suse.de>
15729 * calls.c (special_function_p): setjmp-like functions are leaf.
15730 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
15731 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
15733 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
15735 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
15736 (aarch64_simd_mov<mode>): New expander.
15737 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
15738 (aarch64_simd_mov_to_<mode>high): Likewise.
15739 (aarch64_simd_mov_from_<mode>low): Likewise.
15740 (aarch64_simd_mov_from_<mode>high): Likewise.
15741 (aarch64_dup_lane<mode>): Update.
15742 (aarch64_dup_lanedi): New instruction pattern.
15743 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
15744 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
15746 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15748 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
15749 (replace_mult_candidate): Remove unnecessary argument; remove
15750 unnecessary parameter from call to introduce_cast_before_cand.
15751 (replace_unconditional_candidate): Remove unnecessary parameter
15752 from call to replace_mult_candidate.
15753 (replace_conditional_candidate): Likewise.
15754 (insert_initializers): Use make_temp_ssa_name.
15755 (introduce_cast_before_cand): Remove unnecessary argument; use
15756 make_temp_ssa_name.
15757 (replace_one_candidate): Remove unnecessary argument; remove
15758 unnecessary parameter from calls to introduce_cast_before_cand.
15759 (replace_profitable_candidates): Remove unnecessary parameters
15760 from calls to replace_one_candidate.
15762 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15764 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
15765 phi def as possibly hiding a basis for a CAND_ADD whose operands
15766 have been commuted in the analysis.
15767 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
15769 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15771 * config/aarch64/aarch64.md
15772 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
15773 shift value between 0-4.
15775 2013-05-07 Richard Biener <rguenther@suse.de>
15777 * double-int.h (rshift): New overload.
15778 * double-int.c (rshift): New function.
15779 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
15780 (create_reference_ops_from_ref): Remove.
15781 (vn_reference_insert): Use shared ops for constructing the
15782 reference and copy it.
15784 2013-05-07 Richard Biener <rguenther@suse.de>
15786 PR middle-end/57190
15787 * tree-eh.c (sink_clobbers): Properly propagate
15788 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
15790 2013-05-07 Jakub Jelinek <jakub@redhat.com>
15792 PR tree-optimization/57149
15793 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
15794 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
15795 collect_phi_def_edges, execute_late_warn_uninitialized): Use
15796 uninit_undefined_value_p instead of ssa_undefined_value_p.
15799 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
15800 for modifier == EXPAND_INITIALIZER.
15802 2013-05-07 Anton Blanchard <anton@samba.org>
15804 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
15805 for powerpc64 little endian.
15806 * configure: Regenerate.
15808 2013-05-06 Graham Stott <grahams@btinternet.com>
15810 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
15811 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
15812 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
15813 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
15815 2013-05-06 Graham Stott <grahams@btinternet.com>
15817 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
15818 codes which allow non-lvalues.
15820 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
15822 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
15823 components are all 1s.
15824 (integer_minus_onep): New function.
15825 * tree.h (integer_minus_onep): Declare it.
15826 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
15827 integer_minus_onep instead of integer_all_onesp.
15829 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15832 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
15833 variations of these patterns.
15835 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
15837 * config/i386/i386.md (isa): Add x64_sse4 member.
15838 (enabled): Handle x64_sse4.
15839 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
15840 instruction for 64bit SSE4_1 targets. Update insn attributes.
15841 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
15842 instruction for SSE4_1 targets. Update insn attributes.
15843 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
15844 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
15845 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
15847 (*vec_extractv4si): Rename from *sse4_1_pextrd.
15848 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
15849 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
15851 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15854 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
15856 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
15858 * final.c (do_assembler_dialects): Don't handle curly braces and
15859 vertical bar escaped by % as dialect delimiters.
15860 (output_asm_insn): Print curly braces and vertical bar if escaped
15861 by % and ASSEMBLER_DIALECT defined.
15862 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
15863 * doc/tm.texi: Regenerated.
15865 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
15867 * config/mips/mips.c: Include tree-pass.h.
15868 (mips_reorg): Split in pre- and post-dbr_schedule parts.
15869 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
15870 (pass_mips_machine_reorg2): New machine specific pass.
15871 (insert_pass_mips_machine_reorg2): New pass plugin definition.
15872 (mips_option_override): Register the new pass.
15873 * rtl.h (cleanup_barriers): Remove prototype.
15874 (dbr_schedule): Likewise.
15875 * jump.c (cleanup_barriers): Make static.
15876 * reorg.c (dbr_schedule): Likewise.
15878 2013-05-06 Richard Biener <rguenther@suse.de>
15880 PR tree-optimization/57185
15881 * tree-parloops.c (add_field_for_reduction): Handle anonymous
15882 SSA names properly.
15884 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
15887 * config/i386/i386.c (add_parameter_dependencies): Add dependence
15888 between "first_arg" and "insn", not "last" and "insn".
15890 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15892 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
15893 (find_candidates_in_block): Re-enable slsr_process_phi.
15894 (create_phi_basis): Fix double counting of candidate adjustment.
15896 2013-05-06 Richard Biener <rguenther@suse.de>
15898 PR middle-end/57147
15899 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
15900 the edge is also fallthru, preserve it and just clear the
15902 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
15903 also complex, preserve that and just clear the fallthru flag.
15904 * tree-inline.c (update_ssa_across_abnormal_edges): Also
15905 update virtual operands.
15907 2013-05-06 Alan Modra <amodra@gmail.com>
15909 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
15910 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
15911 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
15912 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
15913 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
15914 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
15916 2013-05-06 Alan Modra <amodra@gmail.com>
15918 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
15919 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
15920 (DEFAULT_ASM_ENDIAN): Define.
15921 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
15922 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
15923 Update -K PIC clause from sysv4.h.
15924 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
15925 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
15927 2013-05-06 Alan Modra <amodra@gmail.com>
15929 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
15930 twice for little-endian.
15931 (ashrdi3_no_power, ashrdi3): Support little-endian.
15933 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15936 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
15937 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
15939 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
15940 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
15941 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
15944 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
15945 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15947 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
15948 * config/pa/pa.opt: Make mbig-switch a no-op.
15949 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
15950 (CASE_VECTOR_MODE): Always return SImode.
15951 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
15952 for the !TARGET_BIG_SWITCH case.
15953 * config/pa/pa-linux.h: Likewise.
15954 * config/pa/pa-openbsd.h: Likewise.
15955 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
15956 * config/pa/pa.md (short_jump): Remove define_insn.
15957 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
15958 (casesi0): Remove define_insn.
15959 (type): Remove btable_branch.
15960 (pa_combine_type): Likewise.
15961 (in_nullified_branch_delay): Likewise.
15962 (in_call_delay): Likewise.
15963 (define_delay): Likewise.
15964 (define_insn_reservation "Z3"): Likewise.
15965 (define_insn_reservation "Z4"): Likewise.
15966 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
15967 (pa_adjust_insn_length): Remove adjustment for btable branches.
15968 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
15971 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
15973 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
15974 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
15975 Add m->r,x alternatives.
15976 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
15977 splitters using SWI48x mode iterator.
15978 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
15979 TARGET_64BIT. Add m->x alternative.
15980 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
15981 Add o->x alternative. Enable for TARGET_SSE.
15982 (sse_storeq): Remove expander.
15983 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
15984 with memory input operand.
15985 (*vec_extractv2di_1 splitter): New.
15986 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
15987 * config/i386/i386.md (ssevecmodelower): New mode attribute.
15989 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
15991 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
15992 (INT_LOWPART): Delete.
15993 (extract_MB): Adjust.
15994 (extract_ME): Adjust.
15995 (print_operand): Adjust.
15997 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
15999 * config/rs6000/predicates.md (reg_or_add_cint_operand,
16000 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
16001 (reg_or_logical_cint_operand, easy_fp_constant,
16002 logical_const_operand): Delete "CONST_DOUBLE" case.
16003 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
16004 "HOST_BITS_PER_WIDE_INT == 64" test.
16005 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
16006 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
16007 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
16008 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
16010 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
16011 CONST_DOUBLE DImode/VOIDmode case.
16012 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
16013 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
16014 CONST_DOUBLE VOIDmode case.
16015 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
16016 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
16017 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
16018 Delete CONST_DOUBLE case.
16019 (splitters for mov FMOVE64 const_double): Delete
16020 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
16021 "HOST_BITS_PER_WIDE_INT >= 64" test.
16022 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
16024 (mov DI const_double): Delete.
16026 2013-05-04 Jakub Jelinek <jakub@redhat.com>
16028 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
16029 on op shows all bits zero in mode of a lowpart subreg, return zero.
16031 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
16034 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
16035 to save TFmode registers and DImode to save TImode registers for
16036 caller save operations.
16037 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
16038 mark being partially clobbered since they only use the first
16041 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
16042 and TDmode only use the upper 64-bits of each VSX register.
16044 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16046 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
16047 (find_candidates_in_block): Disable slsr_process_phi.
16049 2013-05-03 Guozhi Wei <carrot@google.com>
16051 * coverage.c (coverage_obj_init): Move the construction of gcov
16053 (build_init_ctor): ... here.
16055 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16057 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
16058 (slsr_cand_d): Redefine def_phi.
16059 (stride_status, phi_adjust_status, count_phis_status): New enums.
16060 (find_phi_def): New.
16061 (find_basis_for_base_expr): New.
16062 (find_basis_for_candidate): Handle hidden bases.
16063 (alloc_cand_and_find_basis): Handle phi candidates.
16064 (slsr_process_phi): New.
16065 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
16066 (create_mul_imm_cand): Likewise.
16067 (create_add_ssa_cand): Exclude phi base candidates.
16068 (create_add_imm_cand): Likewise.
16069 (slsr_process_cast): Likewise.
16070 (slsr_process_copy): Likewise.
16071 (find_candidates_in_block): Handle phi candidates.
16072 (dump_candidate): Likewise.
16073 (unconditional_cands): Delete.
16074 (unconditional_cands_with_known_stride_p): Delete.
16075 (phi_dependent_cand_p): New.
16076 (cand_increment): Handle phi-dependent candidates.
16077 (replace_dependent): Delete.
16078 (replace_mult_candidate): New.
16079 (replace_unconditional_candidate): New.
16080 (incr_vec_index): Move to avoid forward reference.
16081 (create_add_on_incoming_edge): New.
16082 (create_phi_basis): New.
16083 (replace_dependents): Delete.
16084 (replace_conditional_candidate): New.
16085 (phi_add_costs): New.
16086 (replace_uncond_cands_and_profitable_phis): New.
16087 (record_increment): Handle phi adjustments.
16088 (record_phi_increments): New.
16089 (record_increments): Handle phi adjustments.
16090 (phi_incr_cost): New.
16091 (lowest_cost_path): Handle phis.
16092 (total_savings): Likewise.
16093 (analyze_increments): Likewise.
16094 (ncd_with_phi): New.
16095 (ncd_of_cand_and_phis): New.
16096 (nearest_common_dominator_for_cands): Handle phi increments.
16097 (all_phi_incrs_profitable): New.
16098 (replace_profitable_candidates): Handle phi-dependent candidates.
16099 (analyze_candidates_and_replace): Likewise.
16101 2013-05-03 Teresa Johnson <tejohnson@google.com>
16104 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
16105 do not exceed REG_BR_PROB_BASE.
16107 2013-05-03 Jeff Law <law@redhat.com>
16109 PR tree-optimization/57144
16110 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
16111 operand of the condition will bit into the new type when eliminating
16112 a cast feeding a condition.
16114 2013-05-03 Jakub Jelinek <jakub@redhat.com>
16116 PR rtl-optimization/57130
16117 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
16118 of COMPARE as in_code to the recursive call if needed.
16120 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
16122 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
16123 (enabled): Handle new members.
16124 * config/i386/sse.md (*vec_concatv2si): Merge from
16125 *vec_concatv2si_sse2 and vec_concatv2si_sse.
16126 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
16128 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
16130 PR tree-optimization/57027
16131 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
16132 for fnms opportunity, check we got the prerequisite kind
16133 of tree / gimple before using accessor functions.
16135 2013-05-03 Richard Biener <rguenther@suse.de>
16137 * double-int.h (lshift): New overload without precision
16138 and arith argument.
16139 (operator *=, operator +=, operator -=): Move ...
16140 * double-int.c (operator *=, operator +=, operator -=): ... here
16141 and implement more efficiently.
16142 (mul_double_with_sign): Remove.
16143 (lshift_double): Adjust to take unsinged shift argument, push
16144 dispatching code to callers.
16145 (mul_double_wide_with_sign): Add early out for callers that
16146 are not interested in high parts or overflow.
16147 (lshift): New function.
16148 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
16149 dispatch code here.
16150 (lrotate, rrotate): Use logical shifts.
16151 * expr.c (get_inner_reference): Use lshift.
16152 * fixed-value.c (do_fixed_divide): Likewise.
16153 * tree-dfa.c (get_ref_base_and_extent): Likewise.
16154 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
16155 (indirect_refs_may_alias_p): Likewise.
16156 (stmt_kills_ref_p_1): Likewise.
16158 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
16160 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
16162 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
16164 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
16165 scalar form of FABD instruction.
16167 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
16169 * lra-constraints.c (process_alt_operands): Add checking alt
16170 number to choose the best alternative.
16172 2013-05-02 Richard Biener <rguenther@suse.de>
16174 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
16175 bitmap and its handling.
16176 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
16178 2013-05-02 Richard Biener <rguenther@suse.de>
16180 PR middle-end/57140
16181 * tree-inline.c (copy_loops): Properly handle removed loops.
16182 (copy_cfg_body): Mark destination loops for fixup if source
16183 loops needed fixup.
16185 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
16188 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
16189 generating simple_return for naked functions.
16191 2013-05-02 Martin Jambor <mjambor@suse.cz>
16193 PR middle-end/56988
16194 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
16195 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
16197 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
16198 ipa_agg_replacement_value structures.
16199 (known_aggs_to_agg_replacement_list): Likewise.
16200 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
16201 (read_agg_replacement_chain): Likewise.
16202 (ipcp_transform_function): Also check that by_ref flags match.
16204 2013-05-02 Richard Biener <rguenther@suse.de>
16206 * graphds.h (struct graph): Add obstack member.
16207 * graphds.c (new_graph): Initialize obstack and allocate
16209 (add_edge): Allocate edge from the obstack.
16210 (free_graph): Free the obstack instead of all edges and vertices.
16212 2013-05-02 Teresa Johnson <tejohnson@google.com>
16214 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
16216 * cfg.c (update_bb_profile_for_threading): Ditto.
16217 * tree-inline.c (copy_bb): Ditto.
16218 (copy_edges_for_bb): Ditto.
16219 (initialize_cfun): Ditto.
16220 (copy_cfg_body): Ditto.
16221 (expand_call_inline): Ditto.
16222 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
16223 (estimate_node_size_and_time): Ditto.
16224 (inline_merge_summary): Ditto.
16225 * cgraphclones.c (cgraph_clone_edge): Ditto.
16226 (cgraph_clone_node): Ditto.
16227 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16228 (compute_trg_info): Ditto.
16230 2013-05-02 Ian Bolton <ian.bolton@arm.com>
16232 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
16233 S reg when fp attribute set.
16234 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
16236 2013-05-02 Ian Bolton <ian.bolton@arm.com>
16238 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
16240 (*and_one_cmplsi3_compare0_uxtw): Likewise.
16241 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
16242 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
16244 2013-05-02 Richard Biener <rguenther@suse.de>
16246 * tree-scalar-evolution.c (scev_info_hasher): Remove.
16247 (struct instantiate_cache_entry): New type.
16248 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
16249 (struct instantiate_cache_type): New type.
16250 (set_instantiated_value, get_instantiated_value): Remove.
16251 (get_instantiated_value_entry): New function.
16252 (instantiate_scev_name): Use the new cache and adjust.
16253 (instantiate_scev_poly): Adjust.
16254 (instantiate_scev_binary): Likewise.
16255 (instantiate_array_ref): Likewise.
16256 (instantiate_scev_convert): Likewise.
16257 (instantiate_scev_not): Likewise.
16258 (instantiate_scev_3): Likewise.
16259 (instantiate_scev_2): Likewise.
16260 (instantiate_scev_r): Likewise.
16261 (instantiate_scev): Likewise.
16262 (resolve_mixers): Likewise.
16264 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
16267 * lra-constraints.c (best_small_class_operands_num): Remove.
16268 (process_alt_operands): Remove small_class_operands_num. Take
16269 small classes operands into losers and only if the operand is not
16270 matched. Modify debugging output.
16271 (curr_insn_transform): Remove best_small_class_operands_num.
16274 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16276 * config/aarch64/aarch64-builtins.c
16277 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
16278 * config/aarch64/aarch64-simd-builtins.def
16279 (reduc_splus_): Add new modes.
16280 (reduc_uplus_): New.
16281 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
16282 (reduc_uplus_v4sf): Likewise.
16283 (reduc_splus_v4sf): Likewise.
16284 (aarch64_addv<mode>): Likewise.
16285 (reduc_uplus_<mode>): Likewise.
16286 (reduc_splus_<mode>): Likewise.
16287 (aarch64_addvv2di): Likewise.
16288 (reduc_uplus_v2di): Likewise.
16289 (reduc_splus_v2di): Likewise.
16290 (aarch64_addvv2si): Likewise.
16291 (reduc_uplus_v2si): Likewise.
16292 (reduc_splus_v2si): Likewise.
16293 (reduc_<sur>plus_<mode>): New.
16294 (reduc_<sur>plus_v2di): Likewise.
16295 (reduc_<sur>plus_v2si): Likewise.
16296 (reduc_<sur>plus_v4sf): Likewise.
16297 (aarch64_addpv4sf): Likewise.
16298 * config/aarch64/arm_neon.h
16299 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
16300 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
16301 add UNSPEC_SADDV, UNSPEC_UADDV.
16303 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
16305 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16307 * config/aarch64/arm_neon.h
16308 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
16310 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16312 * config/aarch64/aarch64-builtins
16313 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
16315 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16317 * config/aarch64/aarch64-simd-builtins.def
16318 (reduc_smax_): New.
16319 (reduc_smin_): Likewise.
16320 (reduc_umax_): Likewise.
16321 (reduc_umin_): Likewise.
16322 (reduc_smax_nan_): Likewise.
16323 (reduc_smin_nan_): Likewise.
16326 (smax): Update for V2SF, V4SF and V2DF modes.
16329 (smin_nan): Likewise.
16330 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
16331 (<su><maxmin><mode>3): ...This, refactor.
16332 (s<maxmin><mode>3): New.
16333 (<maxmin_uns><mode>3): Likewise.
16334 (reduc_<maxmin_uns>_<mode>): Refactor.
16335 (reduc_<maxmin_uns>_v4sf): Likewise.
16336 (reduc_<maxmin_uns>_v2si): Likewise.
16337 (aarch64_<fmaxmin><mode>: Remove.
16338 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
16340 (vmin<q>_f<32,64>): Likewise.
16341 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
16343 (su): Add mappings for smax, smin, umax, umin.
16345 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
16346 (FMAXMIN): Rename as...
16347 (FMAXMIN_UNS): ...This.
16349 (fmaxminv): Likewise.
16350 (fmaxmin): Likewise.
16352 (maxmin_uns_op): Likewise.
16354 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16356 * config/aarch64/arm_neon.h
16357 (vac<ge, gt><sd>_f<32, 64>): Rename to...
16358 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
16359 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
16361 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16363 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
16364 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
16366 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16368 * config/aarch64/aarch64-simd.md
16369 (vcond<mode>_internal): Handle special cases for constant masks.
16370 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
16371 (vcondu<mode><mode>): Likewise.
16372 (vcond<v_cmp_result><mode>): New.
16374 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16376 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
16377 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
16378 * config/aarch64/aarch64-simd-builtins.def
16379 (cmeq): Update to BUILTIN_VALLDI.
16384 * config/aarch64/arm_neon.h
16385 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
16386 to builtins or C as appropriate.
16388 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16390 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
16392 (cmhi): Rename to...
16394 * config/aarch64/aarch64-simd.md
16395 (simd_mode): Add SF.
16396 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
16397 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
16398 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
16399 (cstore<mode>_neg): ...This.
16400 * config/aarch64/iterators.md
16402 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
16403 (COMPARISONS): New.
16404 (UCOMPARISONS): Likewise.
16405 (optab): Add missing comparisons.
16411 (VCMP_S): Likewise.
16412 (VCMP_U): Likewise.
16413 (V_cmp_result): Add DF, SF modes.
16414 (v_cmp_result): Likewise.
16416 (vmtype): Likewise.
16417 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
16419 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
16421 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
16422 define_insn to define_insn_and_split.
16423 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
16424 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
16425 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
16426 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
16427 (thumb2_negscc): Likewise.
16429 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16431 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
16433 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16435 * config/arm/thumb2.md: Remove trailing whitespaces.
16437 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16439 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
16440 Use gen_int_mode rather than GEN_INT.
16442 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
16444 * value-prof.c (stream_in_histogram_value): Remove the strayed
16447 2013-04-30 Richard Biener <rguenther@suse.de>
16449 PR middle-end/57122
16450 * cfghooks.c (split_edge): Properly check for the loop latch edge.
16452 2013-04-30 Richard Biener <rguenther@suse.de>
16454 PR middle-end/57107
16455 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
16457 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
16459 PR rtl-optimization/56957
16460 PR rtl-optimization/57105
16461 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
16462 variable. Use just INSN_UID for determining whether an insn
16463 should be only disconnected from the insn stream.
16464 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
16466 2013-04-30 Jakub Jelinek <jakub@redhat.com>
16468 PR tree-optimization/57104
16469 * tsan.c (instrument_expr): Don't instrument accesses to
16470 DECL_HARD_REGISTER VAR_DECLs.
16472 2013-04-30 Richard Biener <rguenther@suse.de>
16474 * function.h (loops_for_fn): New inline function.
16475 (set_loops_for_fn): Likewise.
16476 * cfgloop.h (place_new_loop): Add struct function parameter.
16477 (get_loop): Likewise.
16478 (get_loops): Likewise.
16479 (number_of_loops): Likewise.
16480 (fel_next): Adjust.
16481 (fel_init): Likewise.
16482 * cfg.c (get_loop_copy): Adjust.
16483 * cfgloop.c (flow_loops_dump): Likewise.
16484 (record_loop_exits): Likewise.
16485 (verify_loop_structure): Likewise.
16486 * cfgloopanal.c (mark_irreducible_loops): Likewise.
16487 (estimate_reg_pressure_cost): Likewise.
16488 (mark_loop_exit_edges): Likewise.
16489 * cfgloopmanip.c (place_new_loop): Likewise.
16490 (add_loop): Likewise.
16491 (duplicate_loop): Likewise.
16492 * graph.c (draw_cfg_nodes): Likewise.
16493 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
16494 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
16495 (extract_affine_chrec): Likewise.
16496 (build_scop_iteration_domain): Likewise.
16497 * graphite.c (graphite_initialize): Likewise.
16498 * ira-build.c (create_loop_tree_nodes): Likewise.
16499 (more_one_region_p): Likewise.
16500 (rebuild_regno_allocno_maps): Likewise.
16501 (mark_loops_for_removal): Likewise.
16502 (mark_all_loops_for_removal): Likewise.
16503 (remove_unnecessary_regions): Likewise.
16504 (ira_build): Likewise.
16505 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16506 * loop-init.c (fix_loop_structure): Likewise.
16507 (gate_rtl_move_loop_invariants): Likewise.
16508 (gate_rtl_unswitch): Likewise.
16509 (gate_rtl_unroll_and_peel_loops): Likewise.
16510 (rtl_doloop): Likewise.
16511 * lto-streamer-in.c (input_cfg): Likewise.
16512 * lto-streamer-out.c (output_cfg): Likewise.
16513 * modulo-sched.c (sms_schedule): Likewise.
16514 * predict.c (tree_estimate_probability): Likewise.
16515 (tree_estimate_probability_driver): Likewise.
16516 (estimate_loops): Likewise.
16517 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
16518 (move_sese_region_to_fn): Likewise.
16519 (debug_loop_num): Likewise.
16520 * tree-chrec.c (chrec_evaluate): Likewise.
16521 (hide_evolution_in_other_loops_than_loop): Likewise.
16522 (chrec_component_in_loop_num): Likewise.
16523 (reset_evolution_in_loop): Likewise.
16524 (evolution_function_is_invariant_rec_p): Likewise.
16525 * tree-if-conv.c (main_tree_if_conversion): Likewise.
16526 * tree-inline.c (copy_loops): Likewise.
16527 (copy_cfg_body): Likewise.
16528 (tree_function_versioning): Likewise.
16529 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
16530 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
16532 (add_to_evolution_1): Likewise.
16533 (scev_const_prop): Likewise.
16534 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
16535 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
16536 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
16537 (tree_ssa_lim_initialize): Likewise.
16538 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
16539 (verify_loop_closed_ssa): Likewise.
16540 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
16541 (tree_ssa_loop_im): Likewise.
16542 (tree_ssa_loop_unswitch): Likewise.
16543 (tree_vectorize): Likewise.
16544 (check_data_deps): Likewise.
16545 (tree_ssa_loop_ivcanon): Likewise.
16546 (tree_ssa_loop_bounds): Likewise.
16547 (tree_complete_unroll): Likewise.
16548 (tree_complete_unroll_inner): Likewise.
16549 (tree_parallelize_loops): Likewise.
16550 (tree_ssa_loop_prefetch): Likewise.
16551 (tree_ssa_loop_ivopts): Likewise.
16552 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
16553 * tree-vectorizer.c (vectorize_loops): Likewise.
16555 2013-04-29 Mike Frysinger <vapier@gentoo.org>
16557 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
16558 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
16559 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
16560 with EABI_LINK_SPEC.
16562 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16565 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
16568 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
16571 * lra-constraints.c (process_alt_operands): Discourage a bit more
16572 using memory for pseudos. Print cost dump for alternatives.
16573 Modify cost values for conflicts with early clobbers.
16574 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
16576 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16579 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
16581 2013-04-29 Ian Bolton <ian.bolton@arm.com>
16583 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
16584 from/to S register.
16585 (movdi_aarch64): Support LDR/STR from/to D register.
16587 2013-04-29 Ian Bolton <ian.bolton@arm.com>
16589 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
16590 or higher by default.
16592 2013-04-29 Richard Biener <rguenther@suse.de>
16594 PR middle-end/57075
16595 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
16596 even if not adding abnormal edges for calls that can make
16599 2013-04-29 Richard Biener <rguenther@suse.de>
16601 PR middle-end/57103
16602 * tree-cfg.c (move_stmt_op): Fix condition under which to update
16604 (move_stmt_r): Remove redundant checking.
16606 2013-04-29 Teresa Johnson <tejohnson@google.com>
16609 * basic-block.h (apply_scale): New function.
16610 (apply_probability): Use apply_scale.
16611 * gimple-streamer-in.c (input_bb): Ditto.
16612 * lto-streamer-in.c (input_cfg): Ditto.
16613 * lto-cgraph.c (merge_profile_summaries): Ditto.
16614 * tree-optimize.c (execute_fixup_cfg): Ditto.
16615 * tree-inline.c (copy_bb): Update comment to use apply_scale.
16616 (copy_edges_for_bb): Ditto.
16617 (copy_cfg_body): Ditto.
16619 2013-04-29 Tom de Vries <tom@codesourcery.com>
16621 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
16622 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
16623 (tail_merge_optimize): Handle current_loops == NULL.
16625 2013-04-26 Jeff Law <law@redhat.com>
16627 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
16628 (simplify_cond_using_ranges): Generalize code to simplify
16629 COND_EXPRs where one argument is a constant and the other
16630 is an SSA_NAME created by an integral type conversion.
16632 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16634 * config/arm/arm.md (store_minmaxsi): Use only when
16635 optimize_insn_for_size_p.
16637 2013-04-29 Christian Bruel <christian.bruel@st.com>
16640 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
16642 2013-04-29 Richard Biener <rguenther@suse.de>
16644 PR middle-end/57089
16645 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
16646 loop tree make sure to schedule a fixup for the child as well.
16647 (expand_omp_for_generic): Properly add loops.
16648 (expand_omp_for_static_nochunk): Likewise.
16649 (expand_omp_for_static_chunk): Likewise.
16650 (expand_omp_for): For the degenerate case fixup loops.
16651 (expand_omp_sections): Fix default bb placement in loops.
16652 (expand_omp_atomic_pipeline): Properly add loops.
16654 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16656 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
16658 2013-04-29 Tom de Vries <tom@codesourcery.com>
16660 * tree-ssa-tail-merge.c: Update header comment.
16662 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16664 * config/aarch64/arm_neon.h
16665 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
16666 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
16667 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
16668 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
16669 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
16670 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
16671 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
16672 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
16674 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16676 * config/aarch64/aarch64-simd.md
16677 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
16678 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
16679 fix_trunc, fixuns_trunc.
16680 (ftrunc<VDQF:mode>2): New.
16681 * config/aarch64/iterators.md (optab): Add fix, fixuns.
16682 (fix_trunc_optab): New.
16684 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16686 * config/aarch64/aarch64-builtins.c
16687 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
16688 iceilf, lround, iroundf.
16690 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16693 * config/i386/i386.h (enum ix86_tune_indices)
16694 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
16695 New, split from X86_TUNE_INTER_UNIT_MOVES.
16696 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
16697 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
16698 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
16699 (TARGET_INTER_UNIT_MOVES): Remove.
16700 * config/i386/i386.c (initial_ix86_tune_features): Update.
16701 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
16702 (ix86_expand_convert_uns_didf_sse): Use
16703 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
16704 (ix86_expand_vector_init_one_nonzero): Ditto.
16705 (ix86_expand_vector_init_interleave): Ditto.
16706 (inline_secondary_memory_needed): Return true for moves from SSE class
16707 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
16708 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
16709 * config/i386/constraints.md (Yi, Ym): Depend on
16710 TARGET_INTER_UNIT_MOVES_TO_VEC.
16711 (Yj, Yn): New constraints.
16712 * config/i386/i386.md (*movdi_internal): Change constraints of
16713 operand 1 from Yi to Yj and from Ym to Yn.
16714 (*movsi_internal): Ditto.
16715 (*movdf_internal): Ditto.
16716 (*movsf_internal): Ditto.
16717 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
16718 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
16719 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
16720 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
16721 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
16722 * config/i386/sse.md (movdi_to_sse): Ditto.
16723 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
16724 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
16725 TARGET_INTER_UNIT_MOVES.
16726 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
16727 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
16728 instead of TARGET_INTER_UNIT_MOVES.
16729 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
16730 operand 1 from Yi to Yj and from Ym to Yn.
16732 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16734 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
16735 (float_truncate_hi_): Likewise.
16736 (float_extend_lo_): Likewise.
16737 (float_truncate_lo_): Likewise.
16738 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
16739 (aarch64_float_extend_lo_v2df): Likewise.
16740 (vec_unpacks_hi_v4sf): Likewise.
16741 (aarch64_float_truncate_lo_v2sf): Likewise.
16742 (aarch64_float_truncate_hi_v4sf): Likewise.
16743 (vec_pack_trunc_v2df): Likewise.
16744 (vec_pack_trunc_df): Likewise.
16746 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16748 * config/aarch64/aarch64-builtins.c
16749 (aarch64_fold_builtin): Fold float conversions.
16750 * config/aarch64/aarch64-simd-builtins.def
16751 (floatv2si, floatv4si, floatv2di): New.
16752 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
16753 * config/aarch64/aarch64-simd.md
16754 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
16755 * config/aarch64/iterators.md (FLOATUORS): New.
16756 (optab): Add float, floatuns.
16757 (su_optab): Likewise.
16759 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16761 * config/aarch64/aarch64-builtins.c
16762 (aarch64_builtin_vectorized_function): Use new names for
16764 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
16765 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
16766 (fcvtzu): Split as...
16767 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
16768 (fcvtas): Split as...
16769 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
16770 (fcvtau): Split as...
16771 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
16772 (fcvtps): Split as...
16773 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
16774 (fcvtpu): Split as...
16775 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
16776 (fcvtms): Split as...
16777 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
16778 (fcvtmu): Split as...
16779 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
16780 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
16781 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
16782 (lfrintnusf, lfrintnudf): Likewise.
16783 * config/aarch64/aarch64-simd.md
16784 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
16786 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
16787 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
16788 (fcvt_pattern): Likewise.
16790 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16792 * config/aarch64/aarch64-simd.md
16793 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
16794 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
16796 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16798 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
16799 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
16800 (vrnd<a,m,n,p>_f32): Implement using builtins.
16801 (vrnd<i,x><q>_f<32, 64>): New.
16803 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16805 * config/aarch64/aarch64-builtins.c
16806 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
16807 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
16808 (frintz): Rename to...
16810 (frintp): Rename to...
16812 (frintm): Rename to...
16814 (frinti): Rename to...
16815 (nearbyint): ...this.
16816 (frintx): Rename to...
16818 (frinta): Rename to...
16820 * config/aarch64/aarch64-simd.md
16821 (aarch64_frint<frint_suffix><mode>): Delete.
16822 (<frint_pattern><mode>2): Convert to insn.
16823 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
16824 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
16825 (frint_pattern): Likewise.
16826 (frint_suffix): Likewise.
16828 2013-04-29 Richard Biener <rguenther@suse.de>
16830 PR tree-optimization/57081
16831 * loop-init.c: Include tree-flow.h.
16832 (loop_optimizer_finalize): Free number of iteration estimates.
16833 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
16835 2013-04-29 Jakub Jelinek <jakub@redhat.com>
16837 PR tree-optimization/57083
16838 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
16839 non-singleton shift count range, zero extend low_bound for uns case.
16841 * config/i386/predicates.md (general_vector_operand): New predicate.
16842 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
16843 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
16844 if they aren't nonimmediate operands. If their original values
16845 satisfy const_vector_equal_evenodd_p, don't shift them.
16846 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
16847 predicates. For the SSE4.1 case force operands[{1,2}] into registers
16848 if not nonimmediate_operand.
16849 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
16850 instead of register_operand.
16851 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
16853 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
16855 * stor-layout.c (finalize_size_functions): Allocate a structure and
16856 reset cfun before dumping the functions.
16858 2013-04-27 Jakub Jelinek <jakub@redhat.com>
16860 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
16863 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
16864 use xop_pmacsdqh if uns_p.
16865 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
16866 the immediate rotate count.
16868 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
16870 * rtl.h (struct rtx_def): Add comment for field jump.
16871 (LRA_SUBREG_P): New macro.
16872 * recog.c (register_operand): Check LRA_SUBREG_P.
16873 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16875 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16876 (lra_final_code_change): Skip subreg change for operators.
16877 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16878 if there are no operand changes.
16879 * lra-constraints.c (curr_insn_set): New.
16880 (match_reload): Set LRA_SUBREG_P.
16881 (emit_spill_move): Ditto.
16882 (check_and_process_move): Use curr_insn_set. Process only single
16883 set insns. Don't initialize sec_mem_p and change_p.
16884 (simplify_operand_subreg): Use LRA_SUBREG_P.
16885 (reg_in_class_p): New function.
16886 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16887 of #ifdef. Add code to remove cycling.
16888 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16889 non-null disp. Reload inner instead of disp when base and index
16890 are null. Try to put lo_sum into register.
16891 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16892 (check_and_process_move): Move code for move cost check to
16893 simple_move_p. Remove equiv_substitution.
16894 (simple_move_p): New function.
16895 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16896 curr_insn_set. Call check_and_process_move only for single set
16897 insns. Use the new function. Move call of check_and_process_move
16898 after operand equiv substitution and address process.
16900 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16903 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
16904 with nonlocal goto receivers or returns twice calls, ignore
16905 unininitialized values from abnormal edges to nl goto receiver
16906 or returns twice call.
16908 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16910 PR tree-optimization/57051
16911 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
16912 and VEC_RSHIFT_EXPR if shift count is a multiple of element
16915 2013-04-26 Richard Biener <rguenther@suse.de>
16917 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
16918 (expand_omp_taskreg): Likewise. Mark loops for fixup.
16919 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
16920 (fixup_loop_arrays_after_move): New function.
16921 (move_sese_region_to_fn): Properly outline the loop tree parts
16922 of the SESE region.
16924 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
16926 * config/i386/i386.md (type, unit): Fix long lines.
16928 2013-04-26 Richard Biener <rguenther@suse.de>
16930 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
16931 (lto-streamer-out.o): Likewise.
16932 * cfgloop.c (init_loops_structure): Export, add struct function
16933 argument and adjust.
16934 (flow_loops_find): Adjust.
16935 * cfgloop.h (enum loop_estimation): Add EST_LAST.
16936 (init_loops_structure): Declare.
16937 * lto-streamer-in.c: Include cfgloop.h.
16938 (input_cfg): Input the loop tree.
16939 * lto-streamer-out.c: Include cfgloop.h.
16940 (output_cfg): Output the loop tree.
16941 (output_struct_function_base): Do not drop PROP_loops.
16943 2013-03-26 Richard Biener <rguenther@suse.de>
16945 * tree-cfg.c (execute_build_cfg): Build the loop tree.
16946 (pass_build_cfg): Provide PROP_loops.
16947 (move_sese_region_to_fn): Remove loops that are outlined into fn
16949 * tree-inline.c: Include cfgloop.h.
16950 (initialize_cfun): Do not drop PROP_loops.
16951 (copy_loops): New function.
16952 (copy_cfg_body): Copy loop structure.
16953 (tree_function_versioning): Initialize destination loop tree.
16954 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
16955 (pass_parallelize_loops): Do IL verification.
16956 * loop-init.c (loop_optimizer_init): Fixup loops if required.
16957 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
16958 the CFG make sure we fixup loops as well.
16959 * tree-ssa-tail-merge.c: Include cfgloop.h.
16960 (replace_block_by): When merging loop latches mark loops for fixup.
16961 * lto-streamer-out.c (output_struct_function_base): Drop
16962 PROP_loops for now.
16963 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
16964 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
16965 * ipa-split.c: Include cfgloop.h.
16966 (split_function): Add the new return block to the loop tree root.
16967 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
16968 whether we have removed the forwarder block.
16969 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
16970 * cfgloop.h (place_new_loop): Declare.
16971 * cfgloopmanip.c (place_new_loop): Export.
16972 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
16973 (tree-switch-conversion.o): Likewise.
16974 (tree-complex.o): Likewise.
16975 (tree-inline.o): Likewise.
16976 (tree-ssa-tailmerge.o): Likewise.
16977 (ipa-split.o): Likewise.
16978 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
16979 (tree-ssa-copy.o): Likewise.
16980 * tree-switch-conversion.c: Include cfgloop.h
16981 (process_switch): If we emit a bit-test cascade, schedule loops
16983 * tree-complex.c: Include cfgloop.h.
16984 (expand_complex_div_wide): Properly add new basic-blocks to loops.
16985 * asan.c: Include cfgloop.h.
16986 (create_cond_insert_point): Properly add new basic-blocks to
16987 loops, schedule loop fixup.
16988 * cfgloop.c (verify_loop_structure): Check that looks are not
16990 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
16992 (expand_omp_for_generic): Likewise.
16993 (expand_omp_sections): Likewise.
16994 (expand_omp_atomic_pipeline): Schedule loops for fixup.
16995 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
16996 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
16997 is initialized, not when loops are present.
16998 * tree-parloops.c (parallelize_loops): Remove checking here.
16999 * passes.c (init_optimization_passes): Schedule a copy-propagation
17000 pass before complete unrolling of inner loops.
17002 2013-04-26 Jakub Jelinek <jakub@redhat.com>
17004 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
17005 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
17006 (colorize_init): Add argument to _WIN32 version.
17007 * toplev.c: Include diagnostic-color.h.
17008 (process_options): Default to -fdiagnostics-color=auto if
17009 GCC_COLORS env var is in the environment.
17010 * common.opt (fdiagnostics-color=): Add Var and Init.
17011 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
17012 env var is in the environment, the default is auto rather than never.
17014 * diagnostic.h (file_name_as_prefix): Add context argument.
17015 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
17016 the string as locus.
17017 * langhooks.c (lhd_print_error_function): Adjust caller.
17019 2013-04-25 Lawrence Crowl <crowl@google.com>
17021 * var-tracking.c (shared_hash_def::htab):
17022 Change type to hash_table. Update dependent calls and types.
17024 2013-04-25 Lawrence Crowl <crowl@google.com>
17026 * Makefile.in: Update as needed below.
17028 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
17029 Move declaration to after the type's method definitons.
17031 * attribs.c (htab_t scoped_attributes::attribute_hash):
17032 Change type to hash_table. Update dependent calls and types.
17034 * bitmap.c (htab_t bitmap_desc_hash):
17035 Change type to hash_table. Update dependent calls and types.
17037 * cselib.c (htab_t cselib_hash_table):
17038 Change type to hash_table. Update dependent calls and types.
17040 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
17041 (hash_string_slot_node): Move implementation into lto-streamer.h
17042 struct string_slot_hasher.
17043 (eq_string_slot_node): Likewise.
17045 * data-streamer-out.c: Update output_block::string_hash_table
17046 dependent calls and types.
17048 * dwarf2cfi.c (htab_t trace_index):
17049 Change type to hash_table. Update dependent calls and types.
17051 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
17052 Change type to hash_table. Update dependent calls and types.
17053 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
17054 (htab_t optimize_external_refs::map): Likewise.
17055 (htab_t output_comp_unit::extern_map): Likewise.
17056 (htab_t output_comdat_type_unit::extern_map): Likewise.
17057 (htab_t output_macinfo::macinfo_htab): Likewise.
17058 (htab_t optimize_location_lists::htab): Likewise.
17059 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
17061 * except.c (htab_t ehspec_hash_type):
17062 Change type to hash_table. Update dependent calls and types.
17063 (assign_filter_values::ttypes): Likewise.
17064 (assign_filter_values::ehspec): Likewise.
17065 (sjlj_assign_call_site_values::ar_hash): Likewise.
17066 (convert_to_eh_region_ranges::ar_hash): Likewise.
17068 * gcse.c (htab_t pre_ldst_table):
17069 Change type to hash_table. Update dependent calls and types.
17071 * ggc-common.c (htab_t saving_htab):
17072 Change type to hash_table. Update dependent calls and types.
17073 (htab_t loc_hash): Likewise.
17074 (htab_t ptr_hash): Likewise.
17075 (call_count): Rename ggc_call_count.
17076 (call_alloc): Rename ggc_call_alloc.
17077 (loc_descriptor): Rename make_loc_descriptor.
17078 (add_statistics): Rename ggc_add_statistics.
17080 * ggc-common.c (saving_htab):
17081 Change type to hash_table. Update dependent calls and types.
17083 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
17084 (push_gimplify_context): Likewise.
17085 (pop_gimplify_context): Likewise.
17086 (struct gimple_temp_hash_elt): Added.
17087 (struct gimplify_hasher): Likewise.
17088 (struct gimplify_ctx.temp_htab):
17089 Change type to hash_table. Update dependent calls and types.
17091 * gimple-fold.c: Include gimplify-ctx.h.
17093 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
17094 Change type to hash_table. Update dependent calls and types.
17095 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
17096 avoid potential global name collision.
17098 * gimplify.c: Include gimplify-ctx.h.
17099 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
17100 (htab_t gimplify_ctx::temp_htab):
17101 Update dependent calls and types for new type hash_table.
17102 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
17103 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
17105 * gimplify-ctx.h: New.
17106 (struct gimple_temp_hash_elt): Move from gimplify.c.
17107 (class gimplify_hasher): New.
17108 (struct gimplify_ctx): Move from gimple.h.
17109 (htab_t gimplify_ctx::temp_htab):
17110 Change type to hash_table. Update dependent calls and types.
17112 * graphite-clast-to-gimple.c: Include graphite-htab.h.
17113 (htab_t ivs_params::newivs_index):
17114 Change type to hash_table. Update dependent calls and types.
17115 (htab_t ivs_params::params_index): Likewise.
17116 (htab_t print_generated_program::params_index): Likewise.
17117 (htab_t gloog::newivs_index): Likewise.
17118 (htab_t gloog::params_index): Likewise.
17120 * graphite.c: Include graphite-htab.h.
17121 4htab_t graphite_transform_loops::bb_pbb_mapping):
17122 Change type to hash_table. Update dependent calls and types.
17124 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
17125 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
17126 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
17128 * graphite-dependences.c: Include graphite-htab.h.
17129 (loop_is_parallel_p): Change hash table type of parameter.
17131 * graphite-htab.h: New.
17132 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
17133 (extern find_pbb_via_hash): Move from graphite-poly.h.
17134 (extern loop_is_parallel_p): Move from graphite-poly.h.
17135 (extern get_loop_body_pbbs): Move from graphite-poly.h.
17137 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
17138 (extern loop_is_parallel_p): Move to graphite-htab.h.
17139 (extern get_loop_body_pbbs): Move to graphite-htab.h.
17141 * haifa-sched.c (htab_t delay_htab):
17142 Change type to hash_table. Update dependent calls and types.
17143 (htab_t delay_htab_i2): Likewise.
17145 * ira-color.c (htab_t allocno_hard_regs_htab):
17146 Change type to hash_table. Update dependent calls and types.
17148 * ira-costs.c (htab_t cost_classes_htab):
17149 Change type to hash_table. Update dependent calls and types.
17151 * loop-invariant.c (htab_t merge_identical_invariants::eq):
17152 Change type to hash_table. Update dependent calls and types.
17154 * loop-iv.c (htab_t bivs):
17155 Change type to hash_table. Update dependent calls and types.
17157 * loop-unroll.c (htab_t opt_info::insns_to_split):
17158 Change type to hash_table. Update dependent calls and types.
17159 (htab_t opt_info::insns_with_var_to_expand): Likewise.
17161 * lto-streamer.h (struct string_slot): Move from data-streamer.h
17162 (struct string_slot_hasher): New.
17163 (htab_t output_block::string_hash_table):
17164 Change type to hash_table. Update dependent calls and types.
17166 * lto-streamer-in.c (freeing_string_slot_hasher): New.
17167 (htab_t file_name_hash_table):
17168 Change type to hash_table. Update dependent calls and types.
17170 * lto-streamer-out.c: Update output_block::string_hash_table dependent
17173 * lto-streamer.c (htab_t tree_htab):
17174 Change type to hash_table. Update dependent calls and types.
17176 * omp-low.c: Include gimplify-ctx.h.
17178 * passes.c (htab_t name_to_pass_map):
17179 Change type to hash_table. Update dependent calls and types.
17180 (pass_traverse): Rename to passes_pass_traverse.
17182 * plugin.c (htab_t event_tab):
17183 Change type to hash_table. Update dependent calls and types.
17185 * postreload-gcse.c (htab_t expr_table):
17186 Change type to hash_table. Update dependent calls and types.
17187 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
17189 * sese.c (debug_rename_map_1): Make extern.
17190 (htab_t copy_bb_and_scalar_dependences::rename_map):
17191 Change type to hash_table. Update dependent calls and types.
17193 * sese.h (extern debug_rename_map): Move to .c file.
17195 * store-motion.c (htab_t store_motion_mems_table):
17196 Change type to hash_table. Update dependent calls and types.
17198 * trans-mem.c (htab_t tm_new_mem_hash):
17199 Change type to hash_table. Update dependent calls and types.
17201 * tree-browser.c (htab_t TB_up_ht):
17202 Change type to hash_table. Update dependent calls and types.
17204 * tree-cfg.c (htab_t discriminator_per_locus):
17205 Change type to hash_table. Update dependent calls and types.
17207 * tree-complex.c: Include tree-hasher.h
17208 (htab_t complex_variable_components):
17209 Change type to hash_table. Update dependent calls and types.
17211 * tree-eh.c (htab_t finally_tree):
17212 Change type to hash_table. Update dependent calls and types.
17214 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
17215 struct int_tree_hasher.
17216 (extern int_tree_map_eq): Likewise.
17217 (uid_decl_map_hash): Removed.
17218 (extern decl_tree_map_eq): Likewise.
17220 * tree-hasher.h: New.
17221 (struct int_tree_hasher): New.
17222 (typedef int_tree_htab_type): New.
17224 * tree-inline.c: Include gimplify-ctx.h.
17226 * tree-mudflap.c: Include gimplify-ctx.h.
17228 * tree-parloops.c: Include tree-hasher.h.
17229 (htab_t eliminate_local_variables_stmt::decl_address):
17230 Change type to hash_table. Update dependent calls and types.
17231 (htab_t separate_decls_in_region::decl_copies): Likewise.
17233 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
17234 Change type to hash_table. Update dependent calls and types.
17236 * tree-sra.c (candidates):
17237 Change type to hash_table. Update dependent calls and types.
17239 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
17241 (int_tree_map_hash): Likewise.
17243 * tree-ssa-dom.c (htab_t avail_exprs):
17244 Change type to hash_table. Update dependent calls and types.
17246 * tree-ssa-live.c (var_map_base_init::tree_to_index):
17247 Change type to hash_table. Update dependent calls and types.
17249 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
17250 Change type to hash_table. Update dependent calls and types.
17252 * tree-ssa-phiopt.c (seen_ssa_names):
17253 Change type to hash_table. Update dependent calls and types.
17255 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
17256 Change type to hash_table. Update dependent calls and types.
17258 * tree-ssa-uncprop.c (equiv):
17259 Change type to hash_table. Update dependent calls and types.
17261 2013-04-25 Jakub Jelinek <jakub@redhat.com>
17263 PR rtl-optimization/57003
17264 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
17265 call note_stores with kill_clobbered_value callback again after
17266 killing regs_invalidated_by_call.
17268 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17270 * config/aarch64/aarch64-simd.md
17271 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
17272 (aarch64_simd_bsl<mode>): Likewise.
17273 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
17275 2013-04-25 Marek Polacek <polacek@redhat.com>
17277 PR tree-optimization/57066
17278 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
17280 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
17282 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
17284 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17286 * config/aarch64/aarch64-builtins.c
17287 (aarch64_fold_builtin): New.
17288 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
17289 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
17290 * config/aarch64/aarch64-simd-builtins.def (abs): New.
17291 * config/aarch64/arm_neon.h
17292 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
17294 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17295 Tejas Belagod <tejas.belagod@arm.com>
17297 * config/aarch64/aarch64-builtins.c
17298 (aarch64_gimple_fold_builtin): New.
17299 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
17300 * config/aarch64/aarch64-simd-builtins.def (addv): New.
17301 * config/aarch64/aarch64-simd.md (addpv4sf): New.
17302 (addvv4sf): Update.
17303 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
17305 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17307 * config/aarch64/aarch64.md
17308 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
17310 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17312 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
17313 (*ngcsi_uxtw): New pattern.
17315 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17316 Julian Brown <julian@codesourcery.com>
17318 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
17319 (TB_DREG): Add T_V4HF.
17320 (v4hf_UP): New macro.
17321 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17322 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17323 Handle initialisation of V4HF. Adjust initialisation of reinterpret
17325 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17326 (arm_vector_mode_supported_p): Handle V4HF.
17327 (arm_mangle_map): Handle V4HFmode.
17328 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
17329 * config/arm/arm_neon_builtins.def: Add entries for
17330 vcvtv4hfv4sf, vcvtv4sfv4hf.
17331 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
17332 (neon_vcvtv4hfv4sf): Likewise.
17333 * config/arm/neon-gen.ml: Handle half-precision floating point
17335 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
17336 * config/arm/arm_neon.h: Regenerate.
17337 * config/arm/neon.ml (type elts): Add F16.
17338 (type vectype): Add T_float16x4, T_floatHF.
17339 (type vecmode): Add V4HF.
17340 (type features): Add Requires_FP_bit feature.
17341 (elt_width): Handle F16.
17342 (elt_class): Likewise.
17343 (elt_of_class_width): Likewise.
17344 (mode_of_elt): Refactor.
17345 (type_for_elt): Handle F16, fix error messages.
17346 (vectype_size): Handle T_float16x4.
17347 (vcvt_sh): New function.
17348 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
17349 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
17350 (string_of_mode): Handle V4HF.
17351 * doc/arm-neon-intrinsics.texi: Regenerate.
17353 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17355 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
17356 format specifier in 'X' case.
17358 2013-04-25 Alan Modra <amodra@gmail.com>
17361 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
17362 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
17363 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
17364 Repeat for many other rotate/shift and mask patterns using subregs.
17365 Name lshiftrt insns.
17366 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
17367 on WORDS_BIG_ENDIAN.
17369 2013-04-25 Alan Modra <amodra@gmail.com>
17371 * config.gcc: Support little-endian powerpc-linux targets.
17372 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
17373 (LINK_OS_LINUX_SPEC): Define.
17374 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
17375 Preserve MASK_LITTLE_ENDIAN.
17376 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
17377 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
17378 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
17379 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
17380 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
17381 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
17382 Correct fp word order for little-endian. Don't shift toc entries
17383 smaller than a word for little-endian.
17384 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
17385 (bswapdi2 splits): Correct low-part subreg for little-endian.
17386 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
17387 low/high where such is correct only for be.
17388 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
17389 little-endian for -mcall-aixdesc.
17391 2013-04-25 Alan Modra <amodra@gmail.com>
17393 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
17394 replace_equiv_address_nv.
17396 2013-04-25 Alan Modra <amodra@gmail.com>
17398 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
17400 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17403 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17404 * rtl.h (struct rtx_def): ...
17406 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17408 PR rtl-optimizations/57046
17409 * lra-constraints (split_reg): Set up lra_risky_transformations_p
17410 for multi-reg splits.
17412 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
17414 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
17416 2013-04-24 Sterling Augustine <saugustine@google.com>
17418 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
17419 (comp_dir_string, debug_str_dwo_section): New.
17420 (DEBUG_STR_DWO_SECTION): Rename to ...
17421 (DEBUG_DWO_STR_SECTION): ... this.
17422 (DEBUG_NORM_STR_SECTION): Delete.
17423 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
17424 (DEBUG_STR_DWO_SECTION_FLAGS): New.
17425 (find_AT_string): Move most logic to ...
17426 (find_AT_string_in_table): ... here. New.
17427 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
17428 add_skeleton_AT_string. Delete logic.
17429 (output_skeleton_debug_sections): Remove call to
17430 add_top_level_skeleton_die_attrs.
17431 (add_comp_dir_attribute): Move logic to comp_dir_string.
17432 (dwarf2out_init): Initialize debug_str_dwo_section.
17433 (output_indirect_string): Call find_string_form.
17434 (output_indirect_strings): Rewrite.
17435 (prune_unused_types): Empty skeleton_debug_str_hash.
17436 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
17437 (dwarf2out_finish): Call output_indirect_strings.
17439 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17441 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
17443 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17445 * rtl.h (struct rtx_def): Add comment for field jump.
17446 (LRA_SUBREG_P): New macro.
17447 * recog.c (register_operand): Check LRA_SUBREG_P.
17448 * lra.c (lra): Add note at the end of RTL code. Align non-empty
17450 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
17451 (lra_final_code_change): Skip subreg change for operators.
17452 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
17453 if there are no operand changes.
17454 * lra-constraints.c (curr_insn_set): New.
17455 (match_reload): Set LRA_SUBREG_P.
17456 (emit_spill_move): Ditto.
17457 (check_and_process_move): Use curr_insn_set. Process only single
17458 set insns. Don't initialize sec_mem_p and change_p.
17459 (simplify_operand_subreg): Use LRA_SUBREG_P.
17460 (reg_in_class_p): New function.
17461 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
17462 of #ifdef. Add code to remove cycling.
17463 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
17464 non-null disp. Reload inner instead of disp when base and index
17465 are null. Try to put lo_sum into register.
17466 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
17467 (check_and_process_move): Move code for move cost check to
17468 simple_move_p. Remove equiv_substitution.
17469 (simple_move_p): New function.
17470 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
17471 curr_insn_set. Call check_and_process_move only for single set
17472 insns. Use the new function. Move call of check_and_process_move
17473 after operand equiv substitution and address process.
17475 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
17477 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
17478 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
17479 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
17481 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17483 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
17485 2013-04-24 Marek Polacek <polacek@redhat.com>
17487 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
17488 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
17489 (select_loops_exit_conditions): Likewise.
17490 (number_of_iterations_for_all_loops): Likewise.
17491 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
17492 (scev_analysis): Likewise.
17494 2013-04-02 Catherine Moore <clm@codesourcery.com>
17495 Chao-ying Fu <fu@mips.com>
17497 * config/mips/micromips.md (jraddiusp): New pattern.
17498 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
17499 instruction if possible.
17501 2013-04-24 Alan Modra <amodra@gmail.com>
17503 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
17505 2013-04-24 Julian Brown <julian@codesourcery.com>
17506 Chung-Lin Tang <cltang@codesourcery.com>
17508 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
17509 dependency behavior in enumeration type DIE generation. Add TODO note
17510 to comments about future DW_FORM_sdata/udata re-work of related code.
17512 2013-04-23 Lawrence Crowl <crowl@google.com>
17514 * Makefile.in: Update as needed below.
17516 * hash-table.h (class hash_table):
17517 Correct many methods with parameter types compare_type to the correct
17518 value_type. (Correct code was unlikely to notice the change.)
17519 (hash_table::elements_with_deleted) New.
17520 (class hashtable::iterator): New.
17521 (hashtable::begin()): New.
17522 (hashtable::end()): New.
17523 (FOR_EACH_HASH_TABLE_ELEMENT): New.
17525 * statistics.c (statistics_hashes):
17526 Change type to hash_table. Update dependent calls and types.
17528 * tree-into-ssa.c (var_infos):
17529 Change type to hash_table. Update dependent calls and types.
17531 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
17532 Change type to hash_table. Update dependent calls and types.
17534 * tree-ssa-loop-im.c (struct mem_ref.refs):
17535 Change type to hash_table. Update dependent calls and types.
17537 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
17538 Change type to hash_table. Update dependent calls and types.
17540 * tree-ssa-sccvn.c (vn_tables_s::nary):
17541 Change type to hash_table. Update dependent calls and types.
17542 (vn_tables_s::phis): Likewise.
17543 (vn_tables_s::references): Likewise.
17545 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
17546 (vn_reference_eq): Update parameter and return types.
17548 * tree-ssa-structalias.c (pointer_equiv_class_table):
17549 Change type to hash_table. Update dependent calls and types.
17550 (location_equiv_class_table): Likewise.
17552 * tree-vect-data-refs.c: Consequential changes for making
17553 peeling a hash_table.
17555 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
17556 (destroy_loop_vec_info): Dependent hash_table update.
17558 * tree-vectorizer.h (peeling_htab):
17559 Change type to hash_table. Update dependent calls and types.
17561 2013-04-23 Shiva Chen <shiva0217@gmail.com>
17563 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
17564 to check the register content is equal or not.
17565 * lra-constraints.c (match_reload): Use lra_assign_reg_val
17566 to assign register content record.
17567 * lra-eliminations.c (update_reg_eliminate): Use
17568 lra_update_reg_val_offset to update register content offset.
17569 * lra-int.h (struct lra_reg): Add offset member.
17570 (lra_reg_val_equal_p): New static inline function.
17571 (lra_update_reg_val_offset): New static inline function.
17572 (lra_assign_reg_val): New static inline function.
17573 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
17574 to assign register content record.
17575 (initialize_lra_reg_info_element): Initial offset to zero.
17577 2013-04-23 Catherine Moore <clm@codesourcery.com>
17579 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
17580 operands. Record compression.
17582 2013-04-23 Xinliang David Li <davidxl@google.com>
17584 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
17586 2013-04-23 Richard Biener <rguenther@suse.de>
17588 PR middle-end/57036
17589 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
17590 parameter, only add abnormal goto edges from the copied body
17591 if the call could perform abnormal gotos.
17592 (copy_cfg_body): Adjust.
17594 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
17596 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
17598 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
17600 * coretypes.h (gimple_stmt_iterator): Add struct to make
17603 2013-04-23 Richard Biener <rguenther@suse.de>
17605 PR tree-optimization/57026
17606 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
17607 from SSA names occuring in abnormal PHI nodes.
17609 2013-04-22 Andi Kleen <ak@linux.intel.com>
17611 * lto/lto.c (print_lto_report_1): Fix LTO report names.
17613 2013-04-22 Andi Kleen <ak@linux.intel.com>
17615 * lto/lto.c (print_lto_report_1): Declare early.
17616 (read_cgraph_and_symbols): Call print_lto_report_1 early.
17618 2013-04-22 Andi Kleen <ak@linux.intel.com>
17620 * common.opt (-flto-report-wpa): Add.
17621 * doc/invoke.texi (-flto-report-wpa): Add.
17622 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
17625 2013-04-22 Xinliang David Li <davidxl@google.com>
17627 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
17628 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
17629 * Makefile.in: New dependency
17631 David Daney <ddaney.cavm@gmail.com>
17633 * configure.ac (gcc_cv_as_micromips_support): Use the
17634 --fatal-warnings option.
17635 * configure: Regenerate.
17637 2013-04-22 Marek Polacek <polacek@redhat.com>
17640 * tsan.c (instrument_expr): Don't instrument expression
17641 in case its size is zero.
17643 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
17647 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17649 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17651 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17653 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
17654 (gimple_stmt_iterator): New typedef.
17655 * gimple.h (gimple_stmt_iterator): Rename to...
17656 (gimple_stmt_iterator_d): ... This.
17657 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
17658 trees be valid for GIMPLE and GENERIC.
17659 (TARGET_GIMPLE_FOLD_BUILTIN): New.
17660 * gimple-fold.c (gimple_fold_call): Call target hook
17661 gimple_fold_builtin.
17662 * hooks.c (hook_bool_gsiptr_false): New.
17663 * hooks.h (hook_bool_gsiptr_false): New.
17664 * target.def (fold_stmt): New.
17665 * doc/tm.texi: Regenerate.
17667 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
17670 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
17671 a set sp if no stack realignment.
17673 2013-04-22 Nick Clifton <nickc@redhat.com>
17675 * config.gcc (tilegx-linux): Extend extra_objs rather than
17677 (tilepro-linux): Likewise.
17679 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17681 * config/aarch64/aarch64-builtins.c
17683 (CF0, CF1, CF2, CF3, CF4, CF10): New.
17684 (VAR<1-12>): Add MAP parameter.
17685 (BUILTIN_*): Likewise.
17686 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
17687 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
17688 (aarch64_ushl_n<mode>): Likewise.
17689 (aarch64_sshr_n<mode>): Likewise.
17690 (aarch64_ushr_n<mode>): Likewise.
17691 (aarch64_<maxmin><mode>): Likewise.
17692 (aarch64_sqrt<mode>): Likewise.
17693 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
17694 (vshr<q>_n_*): Likewise.
17696 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17698 * config/aarch64/aarch64-builtins.c
17699 (aarch64_simd_builtin_type_mode): Handle SF types.
17701 (BUILTIN_GPF): Define.
17702 (aarch64_init_simd_builtins): Handle SF types.
17703 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
17704 (frecps): Likewise.
17705 (frecpx): Likewise.
17706 * config/aarch64/aarch64-simd.md
17707 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
17708 (aarch64_frecpe<mode>): New.
17709 (aarch64_frecps<mode>): Likewise.
17710 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
17711 (v8type): Add frecp<esx>.
17712 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
17713 (aarch64_frecps<mode>): Likewise.
17714 * config/aarch64/iterators.md (FRECP): New.
17715 (frecp_suffix): Likewise.
17716 * config/aarch64/arm_neon.h
17717 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
17719 2013-04-22 Christian Bruel <christian.bruel@st.com>
17722 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
17723 (REG_CLASS_NAMES): Idem.
17724 (REG_CLASS_CONTENTS): Idem.
17725 (REGCLASS_HAS_FP_REG): Idem.
17726 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
17727 (sh_conditional_register_usage): Idem.
17729 2013-04-21 Jeff Law <law@redhat.com>
17731 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
17732 (ssa_forward_propagate_and_combine): Use it.
17734 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
17736 * lra.c: Update the flow chart diagram.
17738 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
17740 PR rtl-optimization/56847
17741 * lra-constraints.c (process_alt_operands): Discourage alternative
17742 with non-matche doffsettable memory constraint fro memory with
17745 2013-04-19 Richard Biener <rguenther@suse.de>
17747 PR tree-optimization/56982
17748 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
17750 * gimplify.c (gimplify_call_expr): Notice special calls.
17751 (gimplify_modify_expr): Likewise.
17752 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
17753 abnormal control flow receivers.
17754 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
17755 in the same way as cfun->has_nonlocal_labels.
17756 (gimple_purge_dead_abnormal_call_edges): Likewise.
17757 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
17758 receivers start a basic-block.
17760 2013-04-19 Richard Biener <rguenther@suse.de>
17762 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
17764 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
17765 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
17766 (SLP_TREE_LOAD_PERMUTATION): Add.
17767 (vect_transform_slp_perm_load): Adjust prototype.
17768 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
17769 (vect_free_slp_instance): Likewise.
17770 (vect_create_new_slp_node): Likewise.
17771 (vect_supported_slp_permutation_p): Remove.
17772 (vect_slp_rearrange_stmts): Adjust.
17773 (vect_supported_load_permutation_p): Likewise. Inline
17774 vect_supported_slp_permutation_p here.
17775 (vect_analyze_slp_instance): Compute load permutations per
17776 slp node instead of per instance.
17777 (vect_get_slp_defs): Adjust.
17778 (vect_transform_slp_perm_load): Likewise.
17779 (vect_schedule_slp_instance): Remove redundant code.
17780 (vect_schedule_slp): Remove hack for PR56270, add it ...
17781 * tree-vect-stmts.c (vectorizable_load): ... here, do not
17782 CSE loads for SLP. Adjust.
17784 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
17786 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
17787 spelling in two comments.
17789 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
17792 * config/arm/arm.c (load_multiple_sequence): Require SP
17793 as base register for loads if SP is in the register list.
17795 2013-04-19 Martin Jambor <mjambor@suse.cz>
17797 PR tree-optimization/56718
17798 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
17799 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
17800 and made public. Adjusted all callers.
17801 (ipa_intraprocedural_devirtualization): New function.
17802 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
17803 (ipa_intraprocedural_devirtualization): Likewise.
17804 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
17806 2013-04-19 Richard Biener <rguenther@suse.de>
17808 PR tree-optimization/57000
17809 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
17811 2013-04-19 Terry Guo <terry.guo@arm.com>
17813 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
17815 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
17816 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
17817 (cortex_m4_fmacs): Use new reservations.
17818 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
17820 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
17822 PR rtl-optimization/56999
17823 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
17825 (lra_coalesce): Remove split_origin_bitmap and related code.
17826 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
17827 ranges if necessary.
17829 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
17831 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
17833 (ix86_expand_call): Remove clobbered_registers array and use
17834 x86_64_ms_sysv_extra_clobbered_registers instead.
17835 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
17837 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
17839 * config/i386/i386.md (*call_rex64_ms_sysv): Use
17840 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
17841 (*call_value_rex64_ms_sysv): Ditto.
17843 2013-04-18 Cary Coutant <ccoutant@google.com>
17845 * dwarf2out.c (output_pubnames): Check die_perennial_p of
17846 parent instead of die_mark.
17848 2013-04-18 Diego Novillo <dnovillo@google.com>
17850 * gimple.c (create_gimple_tmp): New.
17851 (get_expr_type): New.
17852 (build_assign): New.
17853 (build_type_cast): New.
17854 * gimple.h (enum ssa_mode): Define.
17855 (gimple_seq_set_location): New.
17856 * asan.c (build_check_stmt): Change some gimple_build_* calls
17857 to use build_assign and build_type_cast.
17859 2013-04-18 Richard Biener <rguenther@suse.de>
17861 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
17862 handle negative step. Remove redundant checks.
17863 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
17864 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
17865 for negative step and grouped loads fail to vectorize.
17867 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
17869 * emit-rtl.c (reset_insn_used_flags): New function.
17870 (reset_all_used_flags): Use it.
17871 (verify_insn_sharing): New function.
17872 (verify_rtl_sharing): Fix verification for SEQUENCEs.
17874 2013-04-18 Jakub Jelinek <jakub@redhat.com>
17876 PR tree-optimization/56984
17877 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
17878 and (x >> M) >= N don't register any assertion if N << M is the
17881 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
17883 * lower-subreg.c (resolve_simple_move): If called self-recursive,
17884 do not delete_insn insns that have not yet been emitted, only
17885 unlink them with remove_insn.
17886 * df-scan.c (df_insn_delete): Revert r197492.
17888 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
17890 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
17891 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
17893 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
17895 * config/arm/arm.md (movsicc_insn): Convert define_insn into
17896 define_insn_and_split.
17897 (and_scc,ior_scc,negscc): Likewise.
17898 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
17900 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
17902 * config/arm/arm.c (use_return_insn): Return 0 for targets that
17903 can benefit from using a sequence of LDRD instructions in epilogue
17904 instead of a single LDM instruction.
17906 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
17909 * doc/extend.texi: Fix typo.
17911 2013-04-17 Richard Biener <rguenther@suse.de>
17913 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
17914 (vect_build_slp_tree): ... here.
17915 (vect_build_slp_tree_1): Compute which stmts of the SLP group
17916 match. Remove special-casing of mismatched complex loads.
17917 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
17918 re-try the match with swapped commutative operands.
17919 (vect_supported_load_permutation_p): Remove special-casing of
17920 mismatched complex loads.
17921 (vect_analyze_slp_instance): Adjust.
17923 2013-04-17 Richard Biener <rguenther@suse.de>
17925 PR rtl-optimization/56921
17926 * cfgloop.h (struct loop): Add simple_loop_desc member.
17927 (struct niter_desc): Mark with GTY(()).
17928 (simple_loop_desc): Do not use aux field but simple_loop_desc.
17929 * loop-iv.c (get_simple_loop_desc): Likewise.
17930 (free_simple_loop_desc): Likewise.
17933 2013-04-16 Richard Biener <rguenther@suse.de>
17935 PR rtl-optimization/56921
17936 * loop-init.c (pass_rtl_move_loop_invariants): Add
17937 TODO_do_not_ggc_collect to todo_flags_finish.
17938 (pass_rtl_unswitch): Same.
17939 (pass_rtl_unroll_and_peel_loops): Same.
17940 (pass_rtl_doloop): Same.
17942 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
17944 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
17945 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
17946 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
17948 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
17949 * tree-streamer.c (record_common_node): Adjust reference in comment.
17951 2013-04-17 Terry Guo <terry.guo@arm.com>
17953 * config/arm/cortex-m4.md: Add a new bypass.
17955 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17957 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
17959 (*subs_<optab><mode>_multp2): New pattern.
17960 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
17961 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
17963 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17965 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
17966 (*subs_mul_imm_<mode>): New pattern.
17968 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
17971 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
17972 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
17973 (vsx_movti_32bit): Change j->wa to O->wa.
17975 2013-04-16 Richard Biener <rguenther@suse.de>
17977 PR rtl-optimization/56921
17978 * loop-init.c (pass_rtl_move_loop_invariants): Add
17979 TODO_do_not_ggc_collect to todo_flags_finish.
17980 (pass_rtl_unswitch): Same.
17981 (pass_rtl_unroll_and_peel_loops): Same.
17982 (pass_rtl_doloop): Same.
17984 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
17986 * config/arm/arm.c (emit_multi_reg_push): New declaration
17987 for an existing function.
17988 (arm_emit_strd_push): New function.
17989 (arm_expand_prologue): Used here.
17990 (arm_emit_ldrd_pop): New function.
17991 (arm_expand_epilogue): Used here.
17992 (arm_get_frame_offsets): Update condition.
17993 (arm_emit_multi_reg_pop): Add a special case for load of a single
17994 register with writeback.
17996 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
17998 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
18001 2013-04-16 Richard Biener <rguenther@suse.de>
18003 PR tree-optimization/56756
18004 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
18005 (first_mem_ref_loc): New.
18006 (execute_sm): Place the load temporarily before a previous
18007 access instead of in the latch edge to ensure its SSA dependencies
18008 are defined at points dominating the load.
18010 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
18012 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
18013 correct fix by moving header and footer insn to the footer of
18014 the merged basic block. Clear BB_END of the merged-away block.
18016 PR middle-end/43631
18017 * emit-rtl.c (make_note_raw): New function.
18018 (link_insn_into_chain): New static inline function.
18019 (add_insn): Use it.
18020 (add_insn_before, add_insn_after): Factor insn chain linking code...
18021 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
18022 using link_insn_into_chain.
18023 (note_outside_basic_block_p): New helper function for emit_note_after
18024 and emit_note_before.
18025 (emit_note_after): Use nobb variant of add_insn_after if the note
18026 should not be contained in a basic block.
18027 (emit_note_before): Use nobb variant of add_insn_before if the note
18028 should not be contained in a basic block.
18029 (emit_note_copy): Use make_note_raw.
18030 (emit_note): Likewise.
18031 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
18032 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
18033 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
18034 the moved barrier the tail of the basic block it follows.
18035 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
18037 2013-04-15 Jakub Jelinek <jakub@redhat.com>
18039 PR tree-optimization/56962
18040 * gimple-ssa-strength-reduction.c (record_increment): Only set
18041 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
18042 either rhs1 or rhs2 is equal to c->base_expr.
18044 2013-04-15 Richard Biener <rguenther@suse.de>
18046 PR tree-optimization/56933
18047 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
18049 (GROUP_READ_WRITE_DEPENDENCE): Remove.
18050 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
18051 * tree-vect-data-refs.c (vect_analyze_group_access): Move
18052 dependence check ...
18053 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
18055 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
18056 GROUP_READ_WRITE_DEPENDENCE.
18058 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18060 * emit-rtl.c (reset_all_used_flags): New function.
18061 (verify_rtl_sharing): Call reset_all_used_flags before and after
18062 performing the checks.
18064 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18066 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
18067 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
18068 * config/arm/constraints.md (De): New constraint.
18069 * config/arm/neon.md (anddi3_neon): Delete.
18070 (neon_vand<mode>): Expand to standard anddi3 pattern.
18071 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
18072 Move earlier in the file.
18073 (neon_inv_logic_op2): Likewise.
18074 (arm_anddi_operand_neon): New predicate.
18076 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18078 * configure.ac (gcc_cv_ld_as_needed): Set
18079 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
18080 Use -z ignore, -z record on *-*-solaris2*.
18081 (HAVE_LD_AS_NEEDED): Update comment.
18082 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
18083 * configure: Regenerate.
18084 * config.in: Regenerate.
18085 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
18086 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
18087 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
18088 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
18089 equivalents. Fix markup.
18090 * doc/tm.texi: Regenerate.
18092 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
18094 * config/i386/i386.opt: New option mstack-protector-guard=.
18095 * config/i386/i386-opts.h: Add enum stack_protector_guard.
18096 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
18097 TARGET_SSP_TLS_GUARD.
18098 * config/i386/i386.c (ix86_option_override_internal): Set
18099 ix86_stack_protector_guard.
18100 * config/i386/i386.md (stack_protect_set): Enable for
18101 TARGET_SSP_TLS_GUARD only.
18102 (stack_protect_set_<mode>): Ditto.
18103 (stack_protect_test): Ditto.
18104 (stack_protect_test_<mode>): Ditto.
18105 * doc/invoke.texi (i386 Option): Document.
18107 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
18110 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
18111 (S_MODES): Set H_MODE bit.
18112 (SF_MODES): Set only S_MODE and SF_MODE bits.
18113 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
18114 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
18115 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
18116 <MODE_FLOAT>: Likewise.
18118 2013-04-15 Joey Ye <joey.ye@arm.com>
18120 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
18122 2013-04-15 Joey Ye <joey.ye@arm.com>
18124 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
18126 (thumb_far_jump_used_p): Count instruction size and set
18129 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
18131 * reorg.c (fill_simple_delay_slots): Reindent block of code.
18132 * resource.c (mark_target_live_regs): Reformat conditional block.
18134 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
18136 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
18137 notes, they are emitted only just before final.
18138 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
18140 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
18142 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
18143 * cfgrtl.c (delete_insn): Call it here instead.
18144 * lra-spills.c (lra_final_code_change): Use delete_insn.
18145 * haifa-sched.c (sched_remove_insn): Likewise.
18146 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
18147 returning to the nop pool.
18148 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
18149 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
18151 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
18153 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
18154 * doc/tm.texi: Regenerated.
18156 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
18158 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
18161 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
18163 * df-core.c (df_find_def): Compare register numbers.
18164 (df_find_use): Likewise.
18166 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
18169 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
18170 lra_in_progress for return.
18172 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
18174 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
18175 define_insn into define_insn_and_split and emit movsicc patterns.
18177 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
18179 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
18181 2013-04-12 Richard Biener <rguenther@suse.de>
18183 * tree-pass.h (TODO_do_not_ggc_collect): New.
18184 * passes.c (execute_one_ipa_transform_pass): Honor
18185 TODO_do_not_ggc_collect.
18186 (execute_one_pass): Likewise.
18189 2013-04-10 Richard Biener <rguenther@suse.de>
18191 * passes.c (init_optimization_passes): Remove reload pass.
18192 * ira.c (do_reload): Merge into ...
18194 (rest_of_handle_reload): Remove.
18195 (pass_reload): Likewise.
18196 * config/i386/i386.c (ix86_option_override): Refer to ira instead
18197 of reload for vzeroupper pass placement.
18199 2013-04-12 Jakub Jelinek <jakub@redhat.com>
18201 PR tree-optimization/56918
18202 PR tree-optimization/56920
18203 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
18204 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
18205 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
18206 use wide_mul_with_sign method.
18208 2013-04-12 Richard Biener <rguenther@suse.de>
18210 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
18211 not be considered a gimple constant.
18213 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
18215 * fold-const.c (const_binop): Handle vector shifts by a scalar.
18216 (fold_binary_loc): Call const_binop also for mixed vector-scalar
18219 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
18220 Jakub Jelinek <jakub@redhat.com>
18222 * opts.c: Include diagnostic-color.h.
18223 (common_handle_option): Handle OPT_fdiagnostics_color_.
18224 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
18225 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
18226 (diagnostic-color.o): New.
18227 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
18228 (diagnostic_color_rule): New enum.
18229 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
18230 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
18231 the location string.
18232 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
18233 either NULL, or color kind.
18234 * diagnostic-color.c: New file.
18235 * diagnostic-color.h: New file.
18236 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
18238 * doc/invoke.texi (-fdiagnostics-color): Document.
18239 * pretty-print.h (pp_show_color): Define.
18240 (struct pretty_print_info): Add show_color field.
18241 * diagnostic.c: Include diagnostic-color.h.
18242 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
18243 macros. Colorize error:, warning: etc. strings and also the location
18245 (diagnostic_show_locus): Colorize the caret line.
18246 * pretty-print.c: Include diagnostic-color.h.
18247 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
18248 inside of %< %> quotes or quoted through q format modifier.
18250 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18252 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
18254 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18256 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
18257 code in CC_NZ mode.
18258 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
18261 2013-04-11 Marek Polacek <polacek@redhat.com>
18263 PR tree-optimization/48184
18264 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
18266 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
18268 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
18269 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
18270 (skip_simple_arithmetic): Tidy up.
18271 * tree.h (skip_simple_constant_arithmetic): Declare.
18273 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18275 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
18277 2013-04-11 Richard Biener <rguenther@suse.de>
18279 * tree-vect-loop.c (get_initial_def_for_induction): Properly
18280 generate vector constants.
18282 2013-04-11 Richard Biener <rguenther@suse.de>
18284 PR tree-optimization/56878
18285 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
18286 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
18288 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
18289 Prefer to align the DR with the most invariant base address.
18291 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
18293 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
18296 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
18298 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
18299 floating-point vector comparisons against 0.
18301 2013-04-11 Jakub Jelinek <jakub@redhat.com>
18303 PR tree-optimization/56899
18304 * fold-const.c (extract_muldiv_1): Apply distributive law
18305 only if TYPE_OVERFLOW_WRAPS (ctype).
18307 2013-04-11 Bin Cheng <bin.cheng@arm.com>
18310 * ira-costs.c (scan_one_insn): Check whether the source rtx of
18311 loading has side effect.
18313 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
18315 * config/sparc/sparc.c: Include tree-pass.h.
18316 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
18317 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
18318 head of file. Change return type. Split off gate function.
18319 (sparc_gate_work_around_errata): New function.
18320 (pass_work_around_errata): New pass definition.
18321 (insert_pass_work_around_errata) New pass insert definition to
18322 insert pass_work_around_errata just after delayed-branch scheduling.
18323 (sparc_option_override): Insert the pass.
18324 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
18326 2013-04-10 David S. Miller <davem@davemloft.net>
18328 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
18329 or -mcpu=hypersparc.
18331 * target.def (cstore_mode): New hook.
18332 * target.h: Include insn-codes.h
18333 * targhooks.c: Likewise.
18334 (default_cstore_mode): New function.
18335 * targhooks.h: Declare it.
18336 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
18337 * doc/tm.texi: Rebuild.
18338 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
18339 target hook, rather than inspecting the insn_data.
18340 * config/sparc/sparc.c (sparc_cstore_mode): New function.
18341 (TARGET_CSTORE_MODE): Redefine.
18342 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
18344 * config/sparc/predicates.md (cstore_result_operand): New special
18346 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
18347 Use it for operand 0.
18348 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
18349 (*snesi_special): Likewise.
18350 (*snesi_zero): Likewise.
18351 (*seqsi_zero): Likewise.
18352 (*sltu_insn): Likewise.
18353 (*sgeu_insn): Likewise.
18354 (*seqdi_special): Make operand 0 and comparison operation be of
18356 (*snedi_special): Likewise.
18357 (*snedi_special_vis3): Likewise.
18358 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
18359 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
18360 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
18361 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
18362 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
18363 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
18364 (*sltu_extend_sp64): Likewise.
18365 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
18366 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
18367 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
18368 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
18369 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
18371 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
18373 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
18374 (aarch64_start_file): Use the new function.
18376 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
18377 Jason Merrill <jason@redhat.com>
18379 * common.opt: Add -gdwarf.
18380 * opts.c (common_handle_option): Handle it.
18381 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
18383 2013-04-10 Richard Biener <rguenther@suse.de>
18385 * passes.c (execute_todo): Do not call ggc_collect conditional here.
18386 (execute_one_ipa_transform_pass): But unconditionally here.
18387 (execute_one_pass): And here.
18388 (init_optimization_passes): Remove reload pass.
18389 * tree-pass.h (TODO_ggc_collect): Remove.
18390 (pass_reload): Likewise.
18391 * ira.c (do_reload): Merge into ...
18393 (rest_of_handle_reload): Remove.
18394 (pass_reload): Likewise.
18395 * config/i386/i386.c (ix86_option_override): Refer to ira instead
18396 of reload for vzeroupper pass placement.
18397 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
18398 and todo_flags_finish of all passes.
18400 2013-04-10 Richard Biener <rguenther@suse.de>
18402 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
18403 first_const_oprnd field, rename first_def_type to first_op_type.
18404 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
18405 (vect_get_and_check_slp_defs): Always use the type of the
18406 operand. Allow mixed vect_external_def, vect_constant_def types.
18407 (vect_get_constant_vectors): Handle mixed vect_external_def,
18408 vect_constant_def types.
18410 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
18412 PR tree-optimization/55524
18413 * tree-ssa-math-opts.c
18414 (convert_mult_to_fma): Don't use an fms construct
18415 when we don't have an fms operation, but fnma, and it looks
18416 likely that we'll be able to use the latter.
18418 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
18420 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
18422 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
18423 inline fail caused by overwritable functions.
18425 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
18427 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
18428 unnecessary bits in the constant power of two case.
18430 2013-04-10 Richard Biener <rguenther@suse.de>
18432 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
18433 broken code swapping operands.
18434 (vect_build_slp_tree): Do not compute load permutations here.
18435 (vect_analyze_slp_instance): Compute load permutations here,
18436 after building the SLP tree.
18438 2013-04-09 Christian Bruel <christian.bruel@st.com>
18440 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
18441 of next/prev_real_insn.
18443 2013-04-09 Jan Hubicka <jh@suse.cz>
18445 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18446 Drop aliased parameter.
18447 (function_and_variable_visibility): Do not handle alias pairs.
18448 * cgraph.c (varpool_externally_visible_p): Update prototype.
18449 * varpool.c (varpool_add_new_variable): Update.
18451 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18453 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
18455 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
18457 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
18459 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
18461 2013-04-09 Marek Polacek <polacek@redhat.com>
18463 PR tree-optimization/48762
18464 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
18466 2013-04-09 Richard Biener <rguenther@suse.de>
18468 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
18470 (vect_build_slp_tree): Likewise.
18471 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
18472 calculating the cost of a SLP instance.
18473 (vect_analyze_slp_instance): Use it from here, after building
18476 2013-04-09 Jakub Jelinek <jakub@redhat.com>
18478 PR middle-end/56883
18479 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
18480 expand_omp_for_static_chunk): Use simple_p = true in
18481 force_gimple_operand_gsi calls when assigning to addressable decls.
18483 2013-04-09 Jeff Law <law@redhat.com>
18485 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
18486 when the boolean was created by converting a wider object which
18487 had a boolean range.
18489 2013-04-09 Richard Biener <rguenther@suse.de>
18491 * tree-vectorizer.h (slp_void_p): Remove.
18492 (slp_tree): Typedef before _slp_tree declaration.
18493 (struct _slp_tree): Use a vector of slp_tree as children.
18494 (vect_get_place_in_interleaving_chain): Remove.
18495 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
18497 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
18499 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
18500 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
18501 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
18502 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
18503 Use slp_node instead of slp_void_p and adjust.
18505 2013-04-09 Richard Biener <rguenther@suse.de>
18507 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
18508 work that is not necessary.
18510 2013-04-09 Jakub Jelinek <jakub@redhat.com>
18512 PR tree-optimization/56854
18513 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
18514 forward into clobber stmts if it would change MEM_REF lhs into
18517 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
18519 * tree.c (type_hash_lookup, type_hash_add): Make static.
18520 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
18522 2013-04-09 Richard Biener <rguenther@suse.de>
18524 * tree.h (unsave_expr_now): Remove.
18525 * tree-inline.c (mark_local_for_remap_r): Remove.
18526 (unsave_expr_1): Likewise.
18527 (unsave_r): Likewise.
18528 (unsave_expr_now): Likewise.
18529 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
18530 (propagate_tree_value): Likewise.
18532 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
18534 * doc/rtl.texi (sequence): Rewrite documentation to match the
18535 current use of SEQUENCE rtl objects.
18536 * rtl.def (SEQUENCE): Likewise.
18538 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
18539 Update documentation.
18540 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
18541 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
18543 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
18545 2013-04-08 Teresa Johnson <tejohnson@google.com>
18547 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
18548 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
18550 (estimate_edge_size_and_time): Add comment to suggest using rounding
18552 (estimate_node_size_and_time): Ditto.
18553 (remap_edge_change_prob): Use helper rounding divide methods.
18554 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
18555 (gimple_mod_pow2_value_transform): Ditto.
18556 (gimple_mod_subtract_transform): Ditto.
18557 (gimple_ic_transform): Ditto.
18558 (gimple_stringops_transform): Ditto.
18559 * stmt.c (conditional_probability): Ditto.
18560 (emit_case_dispatch_table): Ditto.
18561 * lto-cgraph.c (merge_profile_summaries): Ditto.
18562 * tree-optimize.c (execute_fixup_cfg): Ditto.
18563 * cfgcleanup.c (try_forward_edges): Ditto.
18564 * cfgloopmanip.c (scale_loop_profile): Ditto.
18566 (duplicate_loop_to_header_edge): Ditto.
18567 (lv_adjust_loop_entry_edge): Ditto.
18568 * tree-vect-loop.c (vect_transform_loop): Ditto.
18569 * profile.c (compute_branch_probabilities): Ditto.
18570 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
18571 * lto-streamer-in.c (input_cfg): Ditto.
18572 * gimple-streamer-in.c (input_bb): Ditto.
18573 * ipa-cp.c (update_profiling_info): Ditto.
18574 (update_specialized_profile): Ditto.
18575 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
18576 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
18578 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18579 (compute_trg_info): Ditto.
18580 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
18581 (purge_dead_edges): Ditto.
18582 * loop-unswitch.c (unswitch_loop): Ditto.
18583 * cgraphclones.c (cgraph_clone_edge): Ditto.
18584 (cgraph_clone_node): Ditto.
18585 * tree-inline.c (copy_bb): Ditto.
18586 (copy_edges_for_bb): Ditto.
18587 (initialize_cfun): Ditto.
18588 (copy_cfg_body): Ditto.
18589 (expand_call_inline): Ditto.
18591 2013-04-08 Kai Tietz <ktietz@redhat.com>
18593 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
18594 TARGET_CYGWIN64 by TARGET_64BIT.
18596 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
18598 * config/epiphany/epiphany.md (GPR_1): New constant.
18599 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
18600 * config/epiphany/epiphany.c (gen_compare_reg):
18601 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
18602 is already in place.
18603 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
18604 Don't require being called during rtl expansion; If y operlaps r0,
18606 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
18607 (epiphany_expand_epilogue): Likewise.
18609 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
18610 Don't use CC_FPmode for ORDERED / UNORDERED.
18611 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
18613 * config/epiphany/constraints.md (CnL): New constraint.
18614 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
18615 * config/epiphany/predicates.md (add_operand): Allow 1024.
18617 * config/epiphany/epiphany.md (logical_op): New code iterator.
18618 (op_mnc): New code attribute.
18619 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
18620 (mov_f+1, mov_f+2): New peephole2 patterns.
18622 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
18623 (cstoresi4): Also allow re-use of zero result when doing a NE
18624 comparison to a non-zero operand.
18625 Use (clobber (scratch)) for first insn if the gpr output is not needed.
18627 * config/epiphany/epiphany.md (<insn_opname>v2si3):
18628 Use gen_addsi3_i / gen_subsi3_i.
18630 2013-04-08 Jakub Jelinek <jakub@redhat.com>
18634 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
18635 contain anything but clobbers, at most one __builtin_stack_restore,
18636 optionally debug stmts and final resx, and if it has at least one
18637 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
18638 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
18639 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
18640 which isn't defaut definition, remove them.
18641 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
18642 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
18643 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
18644 with MEM_REF LHS with SSA_NAME address.
18646 2013-04-08 Jeff Law <law@redhat.com>
18648 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
18650 2013-04-08 Richard Biener <rguenther@suse.de>
18652 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
18654 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
18655 determined vector type.
18656 (vect_analyze_data_refs): Likewise.
18657 (vect_get_new_vect_var): Adjust.
18658 (vect_create_destination_var): Preserve SSA name versions.
18659 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
18660 not dump anything here.
18662 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
18664 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
18665 Add member lr_slot_known.
18666 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
18668 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
18669 Remove code that sets lr_slot_offset according to what a previous
18670 version of epiphany_emit_save_restore used to do.
18671 (epiphany_emit_save_restore): When doing an lr save or restore,
18672 set/verify lr_slot_known and lr_slot_offset.
18674 2013-04-08 Xinyu Qi <xyqi@marvell.com>
18677 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
18680 2013-04-08 Richard Biener <rguenther@suse.de>
18682 * alias.c (find_base_term): Fix thinko in previous change.
18684 2013-04-08 Jakub Jelinek <jakub@redhat.com>
18686 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
18687 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
18688 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
18689 if possible to compute val.
18690 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
18691 For QImode integers don't require anything about precision. Use
18692 const_with_all_bytes_same to find out if the constant doesn't have
18693 repeated bytes in it.
18695 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18697 * config/s390/s390.c (s390_expand_insv): Only accept insertions
18700 2013-04-08 Marek Polacek <polacek@redhat.com>
18702 PR rtl-optimization/48182
18703 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
18706 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18709 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
18710 nuses, make sure we have a label.
18712 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18715 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
18716 (rs6000_emit_swdiv_low_precision): Remove.
18717 (rs6000_emit_swdiv): Rewrite to handle between one and four
18718 iterations of Newton-Raphson generally; modify required number of
18719 iterations for some cases.
18720 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
18722 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
18724 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
18725 set-but-unused variable.
18727 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
18728 basic blocks of released function bodies garbage-collectable.
18730 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
18731 (struct rtl_opt_pass): Add TODO_df_finish.
18733 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
18735 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18737 * config/arm/constraints.md (q): New constraint.
18738 * config/arm/ldrdstrd.md: New file.
18739 * config/arm/arm.md (ldrdstrd.md) New include.
18740 (arm_movdi): Use "q" instead of "r" constraint
18741 for double-word memory access.
18742 (movdf_soft_insn): Likewise.
18743 * config/arm/vfp.md (movdi_vfp): Likewise.
18744 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
18745 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
18746 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
18747 (mem_ok_for_ldrd_strd): Likewise.
18748 (output_move_double): Update assertion.
18750 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18752 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
18754 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18756 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
18757 define_insn_and_split.
18758 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
18760 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18762 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
18763 define_insn_and_split.
18764 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
18765 (shiftsi3_compare): New pattern.
18766 (rrx): New pattern.
18767 * config/arm/unspecs.md (UNSPEC_RRX): New.
18769 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18771 * config/arm/arm.md (negdi_extendsidi): New pattern.
18772 (negdi_zero_extendsidi): Likewise.
18774 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18776 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
18777 define_insn_and_split.
18778 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
18779 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
18781 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18783 * config/arm/arm.md (arm_subdi3): Convert define_insn into
18784 define_insn_and_split.
18785 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
18786 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
18788 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18790 * config/arm/arm.md (subsi3_carryin): New pattern.
18791 (subsi3_carryin_const): Likewise.
18792 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
18793 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
18795 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18797 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
18799 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18801 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
18802 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
18804 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18806 * config/arm/arm.c (arm_expand_builtin): Change fcode
18807 type to unsigned int.
18809 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18811 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
18813 2013-04-04 Ian Lance Taylor <iant@google.com>
18815 * doc/standards.texi (Standards): The Go frontend supports the Go 1
18818 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
18820 PR middle-end/56729
18821 * df-scan.c (df_insn_delete): Disable failing assert.
18823 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18825 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
18826 New function prototype.
18827 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
18828 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
18829 (arm_builtin_vectorized_function): New function.
18831 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18833 * config/arm/arm_neon_builtins.def: New file.
18834 * config/arm/arm.c (neon_builtin_data): Move contents to
18835 arm_neon_builtins.def.
18836 (enum arm_builtins): Include neon builtin definitions.
18837 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
18838 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
18840 2013-04-04 Marek Polacek <polacek@redhat.com>
18842 PR tree-optimization/48186
18843 * predict.c (maybe_hot_frequency_p): Return false if
18844 HOT_BB_FREQUENCY_FRACTION is 0.
18845 (cgraph_maybe_hot_edge_p): Likewise.
18847 2013-04-04 Richard Biener <rguenther@suse.de>
18849 PR tree-optimization/56826
18850 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
18853 2013-04-04 Richard Biener <rguenther@suse.de>
18855 PR tree-optimization/56213
18856 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
18857 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
18859 2013-04-04 Richard Biener <rguenther@suse.de>
18861 PR tree-optimization/56837
18862 * tree-loop-distribution.c (classify_partition): For non-zero
18863 values require that the value has the same precision as its
18864 mode to be useful as memset value.
18866 2013-04-03 Nick Clifton <nickc@redhat.com>
18868 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
18869 (fmssf4): Use fmsf.s on E3V5 architectures.
18870 (fnmasf4): Use fnmaf.s on E3V5 architectures.
18871 (fnmssf4): Use fnmsf.s on E3V5 architectures.
18873 2013-04-03 Jeff Law <law@redhat.com>
18875 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
18876 (lra-eliminations.o): Likewise.
18878 2013-04-03 Teresa Johnson <tejohnson@google.com>
18880 * gcov-io.c (compute_working_sets): Moved most of body of old
18881 compute_working_sets here from profile.c.
18882 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
18883 (gcov_working_set_t): Moved typedef here from basic-block.h
18884 (compute_working_set): Declare.
18885 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
18886 (get_working_sets): Renamed from compute_working_set,
18887 replace most of body with call to new compute_working_sets.
18888 (get_exec_counts): Replace call to compute_working_sets
18889 to get_working_sets.
18890 * profile.h (get_working_sets): Renamed from compute_working_set.
18891 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
18892 to get_working_sets.
18893 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
18894 * gcov-dump.c (dump_working_sets): New function.
18896 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
18898 * hwint.c (sext_hwi, zext_hwi): New functions.
18899 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
18900 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
18901 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
18902 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
18903 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
18904 (sext_hwi, zext_hwi): New functions.
18906 2013-04-03 Jeff Law <law@redhat.com>
18908 PR tree-optimization/56799
18909 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
18910 back test for widening conversion erroneously dropped in prior change.
18912 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18915 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
18916 instead of next_real_insn.
18918 2013-04-03 Marek Polacek <polacek@redhat.com>
18921 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
18923 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18926 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
18928 (thumb1_output_casesi): Likewise.
18929 (thumb2_output_casesi): Likewise.
18931 2013-04-03 Richard Biener <rguenther@suse.de>
18933 PR tree-optimization/56817
18934 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
18936 (tree_unroll_loops_completely_1): ... new function to manually
18937 walk the loop tree, properly defering outer loops of unrolled
18938 loops to later iterations.
18940 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18942 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
18943 (vectorizable_load): Likewise.
18944 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
18945 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
18947 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18949 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
18952 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
18954 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
18956 2013-04-03 Bin Cheng <bin.cheng@arm.com>
18958 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
18960 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18962 PR tree-optimization/56790
18963 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
18966 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18968 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
18970 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
18971 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
18974 2013-04-03 Jakub Jelinek <jakub@redhat.com>
18977 * tree.h (force_folding_builtin_constant_p): New decl.
18978 * builtins.c (force_folding_builtin_constant_p): New variable.
18979 (fold_builtin_constant_p): Fold immediately also if
18980 force_folding_builtin_constant_p.
18982 2013-04-03 Richard Biener <rguenther@suse.de>
18984 PR tree-optimization/56812
18985 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
18986 DRs of the same interleaving chain are independent.
18988 2013-04-02 Jason Merrill <jason@redhat.com>
18990 * gdbinit.in (pbb): Use debug fn.
18992 2013-04-02 Lawrence Crowl <crowl@google.com>
18994 * sese.h (struct ivtype_map_elt_s): Remove unused.
18995 (extern debug_ivtype_map): Remove unused.
18996 (extern eq_ivtype_map_elts): Remove unused.
18997 * sese.c (debug_ivtype_map): Removed unused.
18998 (debug_ivtype_map_1): Removed unused.
18999 (debug_ivtype_elt): Remove unused.
19000 (eq_ivtype_map_elts): Remove unused.
19003 2013-04-02 Kai Tietz <ktietz@redhat.com>
19006 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
19007 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
19008 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
19010 (legitimize_pe_coff_symbol): Likewise.
19011 (is_imported_p): New helper-function.
19012 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
19013 for Windows x64 targets.
19014 (ix86_expand_prologue): Optimize for pe-coff targets.
19015 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
19016 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
19017 medium/large code-model.
19018 (legitimize_pic_address): Likewise.
19019 (legitimize_tls_address): Likewise.
19020 (ix86_expand_call): Likewise.
19021 (x86_output_mi_thunk): Likewise.
19022 (get_dllimport_decl): Add new beimport argument.
19023 (construct_plt_address): Don't assert for x64 pe-coff targets.
19024 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
19026 (SYMBOL_FLAG_STUBVAR): New macro.
19027 (SYMBOL_REF_STUBVAR_P): Likewise.
19028 * config/i386/winnt.c (stub_list): New structure.
19029 (stub_head): New local variable.
19030 (i386_pe_record_stub): New function.
19031 (i386_pe_file_end): Emit refptr-stubs.
19033 2013-04-02 Jakub Jelinek <jakub@redhat.com>
19035 PR rtl-optimization/56745
19036 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
19037 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
19040 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
19041 and both of them are MEM_REFs, just compare first argument for
19042 equality and attempt to deal even with differing offsets.
19045 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
19046 of gimple_clobber_p to be MEM_REF.
19047 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
19048 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
19049 after gimplification.
19050 * asan.c (get_mem_ref_of_assignment): Don't instrument
19051 gimple_clobber_p stmts.
19052 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
19053 gimple_clobber_p stmt if they have MEM_REF lhs and
19054 are dead because of another gimple_clobber_p stmt.
19055 * tree-ssa-live.c (clear_unused_block_pointer): Treat
19056 gimple_clobber_p stmts like debug stmts.
19057 (remove_unused_locals): Remove clobbers with MEM_REF lhs
19058 that refer to unused VAR_DECLs or uninitialized values.
19059 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
19060 gimple_clobber_p stmts if they refer to removed parameters.
19061 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
19064 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
19066 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
19067 using SWI48 mode attribute.
19069 2013-04-02 Wei Mi <wmi@google.com>
19071 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
19072 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
19073 *<rotate_insn><mode>3_mask in i386.md.
19075 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
19077 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
19079 2013-04-02 Richard Biener <rguenther@suse.de>
19081 PR tree-optimization/56778
19082 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
19083 Runtime alias tests are not supported for gather loads.
19084 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
19085 stmts referenced from SSA operands before updating SSA form.
19087 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
19088 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19090 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
19091 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
19092 * config/arm/cortex-a53.md: New file.
19093 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
19094 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
19095 * config/arm/arm.c (arm_issue_rate): Likewise.
19096 * config/arm/arm-tune.md: Regenerate
19097 * config/arm/arm-tables.opt: Regenerate.
19098 * config/arm/arm-cores.def: Add cortex-a53.
19100 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
19102 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
19105 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
19107 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
19108 scalar load/store operations using B/H registers.
19109 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
19111 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
19113 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
19115 * config/aarch64/aarch64.c
19116 (aarch64_simd_scalar_immediate_valid_for_move): New.
19117 * config/aarch64/aarch64-protos.h
19118 (aarch64_simd_scalar_immediate_valid_for_move): New.
19119 * config/aarch64/constraints.md (Dh, Dq): New.
19120 * config/aarch64/iterators.md (hq): New.
19122 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
19124 * reorg.c (get_branch_condition): Deal with conditional returns.
19125 (fill_simple_delay_slots): Remove dead code dealing with jumps.
19127 2013-04-01 Wei Mi <wmi@google.com>
19129 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
19130 Truncate operand 2 using %b asm operand modifier.
19131 (*<shift_insn><mode>3_mask): Ditto.
19132 (*<rotate_insn><mode>3_mask): Ditto.
19134 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
19136 PR middle-end/56798
19137 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
19139 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
19141 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
19143 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
19145 2013-03-30 Lawrence Crowl <crowl@google.com>
19147 * dse.c (clear_alias_sets): Remove never set.
19148 (disqualified_clear_alias_sets): Remove never set.
19149 (clear_alias_mode_pool): Remove never set.
19150 (dse_step0): Remove condition that is never true.
19151 (canon_address): Remove condition that is never true.
19152 (dse_step7): Remove condition that is never true.
19153 (rest_of_handle_dse): Remove condition that is never true.
19154 (rest_of_handle_dse::did_global): Remove never read from above.
19155 (dse_step2_spill): Remove never called from above.
19156 (dse_step5_spill): Remove never called from above.
19158 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
19160 * doc/md.texi (Standard Names) <casesi>: Update documentation for
19161 JUMP_TABLE_DATA changes.
19162 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
19163 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
19164 (Insns) <jump_table_data>: New entry.
19165 * doc/tm.texi: Regenerate.
19167 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
19169 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
19170 for table jump at the end of a basic block using tablejump_p.
19171 * targhooks.c (default_invalid_within_doloop): Likewise.
19172 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
19173 target hook implementation that is identical to the default hook.
19174 (rs6000_invalid_within_doloop): Remove.
19176 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
19177 unused variable from tablejump_p call.
19179 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
19180 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
19181 (INSN_DELETED_P): Likewise.
19182 (emit_jump_table_data): New prototype.
19183 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
19184 after 4th as unused.
19185 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
19186 * sched-vis.c (print_insn): Likewise.
19187 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
19188 insn for compatibility with back ends that use next_active_insn to
19189 identify jump table data.
19190 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
19191 (remove_insn): Likewise.
19192 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
19194 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
19195 (emit_jump_table_data): New function.
19197 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
19198 basic block, a JUMP_TABLE_DATA never is.
19199 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
19200 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
19201 off from code handling real insns.
19202 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
19203 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
19204 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
19205 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
19206 is not a NONDEBUG_INSN_P.
19207 * ira-costs.c (scan_one_insn): Likewise.
19208 * jump.c (mark_all_labels): Likewise.
19209 (mark_jump_label_1): Likewise.
19210 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
19211 * lra.c (get_insn_freq): Expect all insns reaching here to be in
19213 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
19214 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
19215 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
19216 JUMP_TABLE_DATA_P insns.
19217 (calculate_elim_costs_all_insns): Likewise.
19218 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
19219 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
19220 (delete_output_reload): Code style fixups.
19221 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
19222 insn flags on this non-insn.
19223 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
19224 as scheduling barriers, for pre-change compatibility.
19225 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
19226 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
19228 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
19229 redundant JUMP_TABLE_DATA_P test.
19230 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
19231 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
19232 (frv_for_each_packet): Likewise.
19233 * config/i386/i386.c (min_insn_size): Likewise.
19234 (ix86_avoid_jump_mispredicts): Likewise.
19235 * config/m32r/m32r.c (m32r_is_insn): Likewise.
19236 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
19237 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
19238 (mips16_insn_length): Robustify.
19239 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
19240 (mips16_split_long_branches): Likewise.
19241 * config/pa/pa.c (pa_combine_instructions): Likewise.
19242 * config/rs6000/rs6000.c (get_next_active_insn): Treat
19243 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
19244 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
19245 as contributing to pool range lengths.
19246 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
19247 Remove redundant JUMP_TABLE_DATA_P test.
19248 (sh_loop_align): Likewise.
19249 (split_branches): Likewise.
19250 (sh_insn_length_adjustment): Likewise.
19251 * config/spu/spu.c (get_branch_target): Likewise.
19253 2013-03-29 Jan Hubicka <jh@suse.cz>
19255 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
19256 gcov streaming; stream hot bb threshold to ltrans.
19257 * predict.c (get_hot_bb_threshold): Break out from ....
19258 (maybe_hot_count_p): ... here.
19259 (set_hot_bb_threshold): New function.
19260 * lto-section-in.c (lto_section_name): Add profile.
19261 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
19262 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
19263 and data-streamer.h
19264 (histogram_entry): New structure.
19265 (histogram, histogram_pool): New global vars.
19266 (histogram_hash): New structure.
19267 (histogram_hash::hash): New method.
19268 (histogram_hash::equal): Likewise.
19269 (account_time_size): New function.
19270 (cmp_counts): New function.
19271 (dump_histogram): New function.
19272 (ipa_profile_generate_summary): New function.
19273 (ipa_profile_write_summary): New function.
19274 (ipa_profile_read_summary): New function.
19275 (ipa_profile): Decide on threshold.
19276 (pass_ipa_profile): Add ipa_profile_write_summary and
19277 ipa_profile_read_summary.
19278 * Makefile.in (ipa.o): Update dependencies.
19279 * lto-streamer.h (LTO_section_ipa_profile): New section.
19281 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
19283 * tree.h (VAR_P): New.
19285 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
19288 * doc/invoke.texi ([-fwhole-program]): Fix typo.
19290 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
19292 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
19293 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
19294 (control_flow_insn_p): Likewise.
19295 * cfgrtl.c (duplicate_insn_chain): Likewise.
19296 * final.c (get_attr_length_1): Likewise.
19297 (shorten_branches): Likewise.
19298 (final_scan_insn): Likewise.
19299 * function.c (instantiate_virtual_regs): Likewise.
19300 * gcse.c (insert_insn_end_basic_block): Likewise.
19301 * ira-costs.c (scan_one_insn): Likewise.
19302 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
19303 * lra.c (check_rtl): Likewise.
19304 * reload1.c (elimination_costs_in_insn): Likewise.
19305 * reorg.c (follow_jumps): Likewise.
19307 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
19308 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
19309 (thumb_far_jump_used_p): Likewise.
19310 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
19311 (workaround_speculation): Likewise.
19312 (add_sched_insns_for_speculation): Likewise.
19313 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
19314 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
19315 (frv_for_each_packet): Likewise.
19316 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
19317 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
19318 (final_emit_insn_group_barriers): Likewise.
19319 * config/m32r/m32r.c (m32r_is_insn): Likewise.
19320 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
19321 (mips16_insn_length): Likewise.
19322 * config/pa/pa.c (pa_reorg): Likewise.
19323 (pa_combine_instructions): Likewise.
19324 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
19325 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
19326 (sh_reorg): Likewise.
19327 (split_branches): Likewise.
19328 * config/spu/spu.c (get_branch_target): Likewise.
19330 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
19333 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
19335 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
19336 Fix declaration name.
19338 2013-03-28 Lawrence Crowl <crowl@google.com>
19340 * graphds.h (struct graph.indicies): Remove unused.
19341 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
19342 (SCOP_ORIGINAL_PDDRS): Remove unused.
19343 * sese.h (extern insert_loop_close_phis): Removed unused.
19344 (extern insert_guard_phis): Removed unused.
19345 (extern ivtype_map_elt_info): Removed unused.
19346 (new_ivtype_map_elt): Removed unused.
19347 * sese.c (ivtype_map_elt_info): Removed unused.
19349 2013-03-28 Lawrence Crowl <crowl@google.com>
19351 * Makefile.in: Add several missing include dependences.
19353 (test-dump.o): New. This object is not added to any executable,
19354 but is present for ad-hoc testing.
19356 (debug (const bitmap_head_def &)): New.
19357 (debug (const bitmap_head_def *)): New.
19359 (extern debug (const bitmap_head_def &)): New.
19360 (extern debug (const bitmap_head_def *)): New.
19362 (debug (edge_def &)): New.
19363 (debug (edge_def *)): New.
19365 (debug (basic_block_def &)): New.
19366 (debug (basic_block_def *)): New.
19368 (dump_node (const_tree, int, FILE *)): Correct source file.
19370 (debug (die_struct &)): New.
19371 (debug (die_struct *)): New.
19373 (extern debug (die_struct &)): New.
19374 (extern debug (die_struct *)): New.
19375 * gimple-pretty-print.c
19376 (debug (gimple_statement_d &)): New.
19377 (debug (gimple_statement_d *)): New.
19378 * gimple-pretty-print.h
19379 (extern debug (gimple_statement_d &)): New.
19380 (extern debug (gimple_statement_d *)): New.
19382 (debug (ira_allocno_copy &)): New.
19383 (debug (ira_allocno_copy *)): New.
19384 (debug (ira_allocno &)): New.
19385 (debug (ira_allocno *)): New.
19387 (extern debug (ira_allocno_copy &)): New.
19388 (extern debug (ira_allocno_copy *)): New.
19389 (extern debug (ira_allocno &)): New.
19390 (extern debug (ira_allocno *)): New.
19392 (debug (live_range &)): New.
19393 (debug (live_range *)): New.
19395 (debug (lra_live_range &)): New.
19396 (debug (lra_live_range *)): New.
19398 (debug (lra_live_range &)): New.
19399 (debug (lra_live_range *)): New.
19401 (debug (omega_pb_d &)): New.
19402 (debug (omega_pb_d *)): New.
19404 (extern debug (omega_pb_d &)): New.
19405 (extern debug (omega_pb_d *)): New.
19407 (debug (const rtx_def &)): New.
19408 (debug (const rtx_def *)): New.
19410 (debug_tree (tree): Move within file.
19411 (debug_raw (const tree_node &)): New.
19412 (debug_raw (const tree_node *)): New.
19413 (dump_tree_via_hooks (const tree_node *, int)): New.
19414 (debug (const tree_node &)): New.
19415 (debug (const tree_node *)): New.
19416 (debug_verbose (const tree_node &)): New.
19417 (debug_verbose (const tree_node *)): New.
19418 (debug_head (const tree_node &)): New.
19419 (debug_head (const tree_node *)): New.
19420 (debug_body (const tree_node &)): New.
19421 (debug_body (const tree_node *)): New.
19422 (debug_vec_tree (tree): Move and reimplement in terms of dump.
19423 (debug (vec<tree, va_gc> &)): New.
19424 (debug (vec<tree, va_gc> *)): New.
19426 (extern debug (const rtx_def &)): New.
19427 (extern debug (const rtx_def *)): New.
19429 (debug_raw (simple_bitmap_def &)): New.
19430 (debug_raw (simple_bitmap_def *)): New.
19431 (debug (simple_bitmap_def &)): New.
19432 (debug (simple_bitmap_def *)): New.
19434 (extern debug (simple_bitmap_def &)): New.
19435 (extern debug (simple_bitmap_def *)): New.
19436 (extern debug_raw (simple_bitmap_def &)): New.
19437 (extern debug_raw (simple_bitmap_def *)): New.
19439 (debug (vinsn_def &)): New.
19440 (debug (vinsn_def *)): New.
19441 (debug_verbose (vinsn_def &)): New.
19442 (debug_verbose (vinsn_def *)): New.
19443 (debug (expr_def &)): New.
19444 (debug (expr_def *)): New.
19445 (debug_verbose (expr_def &)): New.
19446 (debug_verbose (expr_def *)): New.
19447 (debug (vec<rtx> &)): New.
19448 (debug (vec<rtx> *)): New.
19450 (extern debug (vinsn_def &)): New.
19451 (extern debug (vinsn_def *)): New.
19452 (extern debug_verbose (vinsn_def &)): New.
19453 (extern debug_verbose (vinsn_def *)): New.
19454 (extern debug (expr_def &)): New.
19455 (extern debug (expr_def *)): New.
19456 (extern debug_verbose (expr_def &)): New.
19457 (extern debug_verbose (expr_def *)): New.
19458 (extern debug (vec<rtx> &)): New.
19459 (extern debug (vec<rtx> *)): New.
19461 (_list_iter_cond_expr): Make inline instead of static.
19463 (debug (sreal &)): New.
19464 (debug (sreal *)): New.
19466 (extern debug (sreal &)): New.
19467 (extern debug (sreal *)): New.
19469 (extern debug_raw (const tree_node &)): New.
19470 (extern debug_raw (const tree_node *)): New.
19471 (extern debug (const tree_node &)): New.
19472 (extern debug (const tree_node *)): New.
19473 (extern debug_verbose (const tree_node &)): New.
19474 (extern debug_verbose (const tree_node *)): New.
19475 (extern debug_head (const tree_node &)): New.
19476 (extern debug_head (const tree_node *)): New.
19477 (extern debug_body (const tree_node &)): New.
19478 (extern debug_body (const tree_node *)): New.
19479 (extern debug (vec<tree, va_gc> &)): New.
19480 (extern debug (vec<tree, va_gc> *)): New.
19482 (debug (struct loop &)): New.
19483 (debug (struct loop *)): New.
19484 (debug_verbose (struct loop &)): New.
19485 (debug_verbose (struct loop *)): New.
19486 * tree-dump.c: Add header dependence.
19488 (extern debug (struct loop &)): New.
19489 (extern debug (struct loop *)): New.
19490 (extern debug_verbose (struct loop &)): New.
19491 (extern debug_verbose (struct loop *)): New.
19493 (debug (data_reference &)): New.
19494 (debug (data_reference *)): New.
19495 (debug (vec<data_reference_p> &)): New.
19496 (debug (vec<data_reference_p> *)): New.
19497 (debug (vec<ddr_p> &)): New.
19498 (debug (vec<ddr_p> *)): New.
19500 (extern debug (data_reference &)): New.
19501 (extern debug (data_reference *)): New.
19502 (extern debug (vec<data_reference_p> &)): New.
19503 (extern debug (vec<data_reference_p> *)): New.
19504 (extern debug (vec<ddr_p> &)): New.
19505 (extern debug (vec<ddr_p> *)): New.
19507 (debug (pt_solution &)): New.
19508 (debug (pt_solution *)): New.
19510 (extern debug (pt_solution &)): New.
19511 (extern debug (pt_solution *)): New.
19513 (debug (_var_map &)): New.
19514 (debug (_var_map *)): New.
19515 (debug (tree_live_info_d &)): New.
19516 (debug (tree_live_info_d *)): New.
19518 (extern debug (_var_map &)): New.
19519 (extern debug (_var_map *)): New.
19520 (extern debug (tree_live_info_d &)): New.
19521 (extern debug (tree_live_info_d *)): New.
19523 2013-03-28 Jan Hubicka <jh@suse.cz>
19525 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
19527 2013-03-28 Ian Bolton <ian.bolton@arm.com>
19529 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
19530 record only when desired or required.
19532 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
19534 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
19535 *vec_extractv2di_1_rex64. Use x64 isa attribute.
19537 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19539 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
19540 (*andsi3_compare0_uxtw): New pattern.
19541 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
19542 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
19544 2013-03-28 Jan Hubicka <jh@suse.cz>
19546 * data-streamer-in.c (streamer_read_gcov_count): New function.
19547 * gimple-streamer-out.c: Include value-prof.h.
19548 (output_gimple_stmt): Output histogram.
19549 (output_bb): Use streamer_write_gcov_count.
19550 * value-prof.c: Include data-streamer.h
19551 (dump_histogram_value): Add HIST_TYPE_MAX.
19552 (stream_out_histogram_value): New function.
19553 (stream_in_histogram_value): New function.
19554 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
19555 (stream_out_histogram_value, stream_in_histogram_value): Declare.
19556 * data-streamer-out.c (streamer_write_gcov_count): New function.
19557 (streamer_write_gcov_count_stream): New function.
19558 * lto-cgraph.c (lto_output_edge): Update counter streaming.
19559 (lto_output_node): Likewise.
19560 (input_node, input_edge): Likewise.
19561 * lto-streamer-out.c (output_cfg): Update streaming.
19562 * lto-streamer-in.c (input_cfg): Likewise.
19563 * data-streamer.h (streamer_write_gcov_count,
19564 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
19565 * gimple-streamer-in.c: Include value-prof.h
19566 (input_gimple_stmt): Input histograms.
19567 (input_bb): Update profile streaming.
19569 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
19571 * genmodes.c (emit_max_int): New function.
19572 (emit_insn_modes_h): Added call to emit_max_function.
19573 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
19575 * machmode.def: Fixed comment.
19577 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
19579 * combine.c (try_combine): Removed useless assert.
19580 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
19582 2013-03-28 Marek Polacek <polacek@redhat.com>
19583 Richard Biener <rguenther@suse.de>
19585 PR tree-optimization/56695
19586 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
19587 build signed result of a vector comparison.
19588 * tree-cfg.c (verify_gimple_comparison): Check that a result
19589 of a vector comparison has signed type.
19591 2013-03-28 Richard Biener <rguenther@suse.de>
19593 PR tree-optimization/37021
19594 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
19595 do not restrict gaps between groups.
19596 * tree-vect-stmts.c (vectorizable_load): Properly account for
19597 a gap between groups.
19599 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
19601 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
19602 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
19605 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
19607 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
19608 * doc/extend.texi (Named Address Spaces): Ditto.
19609 (Variable Attributes): Ditto.
19611 2013-03-27 Kai Tietz <ktietz@redhat.com>
19613 * config.build: Add support for cygwin x64 target.
19614 * config.gcc: Likewise.
19615 * config.host: Likewise.
19616 * configure.ac: Likewise
19617 * configure: Regenerated.
19619 2013-03-27 Kai Tietz <ktietz@redhat.com>
19621 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
19622 * config/i386/t-cygwin-w64: New file.
19623 * config/i386/cygwin-w64.h: New file.
19624 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
19625 and add support for x64-cygwin target.
19626 (CPP_SPEC): Likewise.
19627 (CXX_WRAP_SPEC_LIST): Undefine before define.
19628 (LIBGCJ_SONAME): Use 15 as version.
19630 2013-03-27 Richard Biener <rguenther@suse.de>
19632 PR tree-optimization/56716
19633 * tree-ssa-structalias.c (perform_var_substitution): Adjust
19634 dumping for ref nodes.
19636 2013-03-27 Martin Jambor <mjambor@suse.cz>
19638 PR tree-optimization/55334
19639 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
19640 restricted pointers to arrays.
19642 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
19644 * Makefile.in (.SUFFIXES): Add .cc.
19645 (.c.o): Apply same recipe for implicit rule .cc.o.
19647 2013-03-27 Richard Biener <rguenther@suse.de>
19649 PR tree-optimization/37021
19650 * tree-vect-data-refs.c (vect_check_strided_load): Allow
19651 REALPART/IMAGPART_EXPRs around the supported refs.
19652 * tree-ssa-structalias.c (find_func_aliases): Assume that
19653 floating-point values are not used to transfer pointers.
19655 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
19657 * target.def (TARGET_HAS_IFUNC_P): New target hook.
19658 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
19659 * doc/tm.texi: Regenerate.
19660 * targhooks.h (default_has_ifunc_p): New.
19661 * targhooks.c (default_has_ifunc_p): Ditto.
19662 * config/linux-protos.h: New file.
19663 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
19664 hook for linux which disables support of indirect functions in android.
19665 * config/linux-android.c: New file.
19666 * config/t-linux-android.c: Ditto.
19667 * config.gcc: Added new object file linux-android.o.
19668 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
19669 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
19670 * varasm.c (do_assemble_alias): Likewise.
19671 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
19672 doesn't support indirect functions.
19673 * configure: Regenerate.
19675 2013-03-27 Bin Cheng <bin.cheng@arm.com>
19678 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
19679 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
19682 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19684 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
19686 2013-03-27 Terry Guo <terry.guo@arm.com>
19688 * config/arm/arm-cores.def: Added core cortex-r7.
19689 * config/arm/arm-tune.md: Regenerated.
19690 * config/arm/arm-tables.opt: Regenerated.
19691 * doc/invoke.texi: Added entry for core cortex-r7.
19693 2013-03-27 Walter Lee <walt@tilera.com>
19695 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
19696 double-decrement of next_scratch_regno.
19698 2013-03-27 Walter Lee <walt@tilera.com>
19700 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
19702 (insn_v1mulus): Ditto.
19703 (insn_v2muls): Ditto.
19705 2013-03-27 Walter Lee <walt@tilera.com>
19707 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
19708 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
19710 2013-03-27 Walter Lee <walt@tilera.com>
19712 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
19713 (*sibcall_value): Ditto.
19715 2013-03-27 Walter Lee <walt@tilera.com>
19717 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
19718 (insn_mnz_v8qi): ... this ...
19719 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
19721 (insn_v<n>mnz): Replaced by ...
19722 (insn_v1mnz): ... this ...
19723 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
19725 (insn_mz_<mode>): Replaced by ...
19726 (insn_mz_v8qi): ... this ...
19727 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
19729 (insn_v<n>mz): Replaced by ...
19730 (insn_v1mz): ... this ...
19731 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
19734 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
19736 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
19738 2013-03-26 Roland McGrath <mcgrathr@google.com>
19740 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
19741 than fprintf with a non-constant, non-format string.
19743 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
19745 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
19746 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
19747 operand 0 predicate.
19748 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
19749 attribute. Use general_x64nomem_operand as operand 1 predicate.
19750 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
19751 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
19752 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
19753 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
19754 (mov<mode>_insv_1): Remove expander. Merge insn with
19755 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
19756 Use general_x64nomem_operand as operand 1 predicate.
19757 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
19758 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
19759 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
19760 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
19761 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
19762 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
19763 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
19764 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
19765 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
19766 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
19767 (general_x64nomem_operand): Ditto.
19769 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
19771 * config/rtems.opt: Add -pthread option.
19773 2013-03-26 Richard Biener <rguenther@suse.de>
19775 * alias.c (find_base_term): Avoid redundant and not used recursion.
19776 (base_alias_check): Get the initial base term from the caller.
19777 (true_dependence_1): Compute and pass base terms to base_alias_check.
19778 (write_dependence_p): Likewise.
19779 (may_alias_p): Likewise.
19781 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
19783 * config/aarch64/aarch64.c (aarch64_classify_address): Support
19784 PC-relative load in SI modes and above only.
19786 2013-03-26 Xinyu Qi <xyqi@marvell.com>
19788 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
19789 * config/arm/iwmmxt.md (WCGR0): Update.
19790 (WCGR1, WCGR2, WCGR3): Likewise.
19792 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
19794 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
19795 Use x64 and nox64 isa attributes.
19797 2013-03-26 Richard Biener <rguenther@suse.de>
19799 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
19800 alignment computations and rely on get_object_alignment_1
19801 for the !TYPE_P case.
19802 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
19804 2013-03-26 Walter Lee <walt@tilera.com>
19806 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
19807 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
19809 2013-03-25 Jeff Law <law@redhat.com>
19811 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
19812 check for INTEGRAL_TYPE_P that was missing due to checking in
19813 wrong version of prior patch.
19815 2013-03-25 Walter Lee <walt@tilera.com>
19817 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
19818 TILEGX_INSN_SHUFFLEBYTES1.
19819 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
19821 (tilegx_builtins): Ditto.
19822 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
19824 2013-03-25 Walter Lee <walt@tilera.com>
19826 * config/tilegx/tilegx.md (floatsisf2): New pattern.
19827 (floatunssisf2): New pattern.
19828 (floatsidf2): New pattern.
19829 (floatunssidf2): New pattern.
19831 2013-03-25 Walter Lee <walt@tilera.com>
19833 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
19834 tests for constraint J, K, N, P.
19836 2013-03-25 Walter Lee <walt@tilera.com>
19838 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
19839 Use indirect/pcrel encoding.
19840 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
19843 2013-03-25 Steve Ellcey <sellcey@mips.com>
19845 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
19846 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
19847 * config/mips/mips.c (mips_option_override): Set IMADD default.
19848 * config/mips/mips.h (PTF_AVOID_IMADD): New.
19849 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
19850 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
19851 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
19852 * doc/invoke.texi (-mimadd/-mno-imadd): New.
19854 2013-03-25 Jeff Law <law@redhat.com>
19856 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
19857 slightly to avoid creating and folding useless trees. Simplify
19858 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
19860 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
19862 * config/i386/i386.md (*zero_extendsidi2): Merge with
19863 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
19864 * config/i386/predicates.md (x86_64_zext_operand): Rename from
19865 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
19866 targets. Clarify comment.
19868 2013-03-25 Martin Jambor <mjambor@suse.cz>
19870 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
19871 pass-through jump functions differently.
19872 (ipa_read_jump_function): Likewise. Also use setter functions to set
19875 2013-03-25 Martin Jambor <mjambor@suse.cz>
19877 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
19878 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
19880 (ipa_get_indirect_edge_target): New function.
19881 (devirtualization_time_bonus): New parameter known_aggs, pass it to
19882 ipa_get_indirect_edge_target. Update all callers.
19883 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
19884 ipa_get_indirect_edge_target_1 instead of calling
19885 ipa_get_indirect_edge_target.
19886 (create_specialized_node): Pass aggvlas to
19887 ipcp_discover_new_direct_edges.
19889 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19891 * config/arm/arm.md (f_sels, f_seld): New types.
19892 (*cmov<mode>): New pattern.
19893 * config/arm/predicates.md (arm_vsel_comparison_operator): New
19896 2013-03-25 Kai Tietz <ktietz@redhat.com>
19898 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
19899 POSIX-printf for mingw-hosted builds.
19901 2013-03-25 Richard Biener <rguenther@suse.de>
19903 PR middle-end/56694
19904 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
19905 must-not-throw stmt location.
19907 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19909 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
19910 Emit load-acquire versions when acq is true.
19911 (arm_emit_store_exclusive): Add rel parameter.
19912 Emit store-release versions when rel is true.
19913 (arm_split_compare_and_swap): Use acquire-release instructions
19915 of barriers when appropriate.
19916 (arm_split_atomic_op): Likewise.
19917 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
19918 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
19919 (VUNSPEC_SLX): Likewise.
19920 (VUNSPEC_LDA): Likewise.
19921 (VUNSPEC_STL): Likewise.
19922 * config/arm/sync.md (atomic_load<mode>): New pattern.
19923 (atomic_store<mode>): Likewise.
19924 (arm_load_acquire_exclusive<mode>): Likewise.
19925 (arm_load_acquire_exclusivesi): Likewise.
19926 (arm_load_acquire_exclusivedi): Likewise.
19927 (arm_store_release_exclusive<mode>): Likewise.
19929 2013-03-25 Catherine Moore <clm@codesourcery.com>
19931 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
19932 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
19933 * config/mip/predicates.md (lwsp_swsp_operand,
19934 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
19935 sb16_operand, db4_operand, db7_operand, ib3_operand,
19936 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
19937 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
19938 andi16_operand): New predicates.
19939 * config/mips/mips.md (compression): New attribute.
19940 (enabled): New attribute.
19941 (length): Consider compression in computing length.
19942 (shift_compression): New code attribute.
19943 (*add<mode>3): New operands. Record compression.
19944 (sub<mode>3): Likewise.
19945 (one_cmpl<mode>2): Likewise.
19946 (*and<mode>3): Likewise.
19947 (*ior<mode>3): Likewise.
19948 (unnamed pattern for xor): Likewise.
19949 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
19950 (*<optab><mode>3): Likewise.
19951 (*mov<mode>_internal: Likewise.
19952 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
19953 (mips_unsigned_immediate_p): New.
19954 (umips_lwsp_swsp_address_p): New.
19955 (m16_based_address_p): New.
19956 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
19957 (mips_unsigned_immediate_p): New prototype.
19958 (lwsp_swsp_address_p): New prototype.
19959 (m16_based_address_p): New prototype.
19960 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
19961 (mips_signed_immediate_p): New function.
19962 (m16_based_address_p): New function.
19963 (lwsp_swsp_address_p): New function.
19964 (mips_print_operand_punctuation): Recognize short delay slot insns
19965 for microMIPS.add<mode>3"
19967 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19970 * config/arm/iterators.md (v_cmp_result): New mode attribute.
19971 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
19973 2013-03-25 Richard Biener <rguenther@suse.de>
19975 PR tree-optimization/56689
19976 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
19979 2013-03-25 Richard Biener <rguenther@suse.de>
19981 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
19983 (memory_references): Likewise.
19984 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
19985 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
19986 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
19987 (gather_mem_refs_in_loops): Fold into ...
19988 (analyze_memory_references): ... this. Move initialization
19989 to tree_ssa_lim_initialize.
19990 (fill_always_executed_in): Rename to ...
19991 (fill_always_executed_in_1): ... this.
19992 (fill_always_executed_in): Move contains_call computation to
19993 this new function from ...
19994 (tree_ssa_lim_initialize): ... here.
19995 (tree_ssa_lim): Call fill_always_executed_in.
19997 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
19999 * postreload.c (reload_combine): Fix code detecting returns.
20001 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
20003 * function.c (emit_use_return_register_into_block): On cc0 targets,
20004 do not emit the sequence between cc0 setter and user.
20006 2013-03-25 Kai Tietz <ktietz@redhat.com>
20008 * config/i386/predicates.md (local_symbolic_operand): Interpret
20009 dll-imported symbols as none-local.
20011 2013-03-25 Richard Biener <rguenther@suse.de>
20013 * tree-ssa-loop-im.c (struct depend): Remove.
20014 (struct lim_aux_data): Make depends a vec of gimples.
20015 (free_lim_aux_data): Adjust.
20016 (add_dependency): Likewise.
20017 (set_level): Likewise.
20019 2013-03-25 Richard Biener <rguenther@suse.de>
20021 PR middle-end/56434
20022 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
20023 the pointer returned by calls with ECF_MALLOC set.
20025 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
20027 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
20029 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
20031 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
20032 using MMXMODE mode iterator.
20033 (*move<mode>_internal): Merge with *movv2sf_internal and
20034 *movv2sf_internal_rex64 using MMXMODE mode iterator.
20036 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
20038 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
20039 (record_last_mem_set_info): Likewise.
20041 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
20042 of XNEWVEC followed by memset.
20043 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
20045 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
20047 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
20048 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
20049 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
20050 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
20051 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
20052 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
20053 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
20054 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
20055 BARRIER_P instead of GET_CODE.
20057 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20059 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
20060 inaccuracy in the probing code.
20062 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
20063 (ctrapdi4): Likewise.
20065 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20067 * calls.c (expand_call): Add missing guard to code handling return
20068 of non-BLKmode structures in MSB.
20069 * function.c (expand_function_end): Likewise.
20071 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20073 * combine.c (try_combine): Adjust comment. Do not add the set of
20074 insn #0 if the destination indirectly is set or dies in insn #2.
20075 Tidy up code to distribute a new note.
20077 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
20079 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
20080 also for alternatives 16 and 17.
20082 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
20084 * config/i386/sse.md (*mov<mode>_internal): Merge with
20085 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
20086 Emit insn template depending on type attribute. Use
20087 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
20088 movd instead of movq mnemonic for interunit moves. Rewrite mode
20089 attribute calculation. Remove unit attribute calculation.
20090 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
20091 Set prefix_data16 attribute for DImode ssemov types.
20092 Use Ym instead of y for SSE-MMX conversion alternatives.
20093 Reorder operand constraints.
20095 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
20097 * df.h (df_insn_delete): Adjust prototype.
20098 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
20099 and let it decide whether mark the basic block dirty.
20100 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
20101 * df-scan.c (df_insn_info_delete): New helper function, split
20102 off from df_insn_delete.
20103 (df_scan_free_bb_info): Use it.
20104 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
20106 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
20107 that the insn is actually an insn and it has a non-NULL basic block.
20108 Do not mark basic block dirty if only deleting a DEBUG_INSN.
20110 2013-03-22 Richard Biener <rguenther@suse.de>
20112 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
20114 (mem_ref_alloc): Do not allocate them.
20115 (refs_independent_p): Do not query or maintain a cache.
20117 2013-03-22 Richard Biener <rguenther@suse.de>
20119 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
20120 (gather_mem_refs_in_loops): Do not compute it.
20121 (analyze_memory_references): Do not allocate it.
20122 (tree_ssa_lim_finalize): Do not free it.
20123 (for_all_locs_in_loop): Do not query all_refs_in_loop.
20125 2013-03-22 Richard Biener <rguenther@suse.de>
20127 * is-a.h (as_a): Use gcc_checking_assert.
20129 2013-03-22 Ian Bolton <ian.bolton@arm.com>
20131 * config/aarch64/aarch64.c (aarch64_print_operand): New
20132 format specifier for printing a constant in hex.
20133 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
20134 format specifier for printing second operand.
20136 2013-03-22 Richard Biener <rguenther@suse.de>
20138 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
20140 (gather_mem_refs_in_loops): Perform store accumulation here.
20141 (create_vop_ref_mapping_loop): Remove.
20142 (create_vop_ref_mapping): Likewise.
20143 (analyze_memory_references): Initialize refs_stored_in_loop.
20144 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
20145 (record_indep_loop): Remove.
20146 (record_dep_loop): New function.
20147 (ref_indep_loop_p_1): Adjust to only walk over references
20148 in the loop, not its subloops.
20149 (ref_indep_loop_p): Rename to ...
20150 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
20151 maintaining a more fine-grained cache.
20152 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
20153 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
20155 2013-03-22 Richard Biener <rguenther@suse.de>
20157 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
20158 (struct mem_ref): Make accesses_in_loop a vec of a vec of
20159 aggregate mem_ref_loc.
20160 (free_mem_ref_locs): Inline into ...
20161 (memref_free): ... this and adjust.
20162 (mem_ref_alloc): Adjust.
20163 (mem_ref_locs_alloc): Remove.
20164 (record_mem_ref_loc): Adjust.
20165 (get_all_locs_in_loop): Rewrite into ...
20166 (for_all_locs_in_loop): ... this iterator.
20167 (rewrite_mem_ref_loc): New functor.
20168 (rewrite_mem_refs): Use for_all_locs_in_loop.
20169 (sm_set_flag_if_changed): New functor.
20170 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
20171 (ref_always_accessed): New functor.
20172 (ref_always_accessed_p): Use for_all_locs_in_loop.
20174 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20176 * tree-pass.h (PROP_gimple_lvec): New.
20177 * passes.c (dump_properties): Handle PROP_gimple_lvec.
20178 (init_optimization_passes): Move pass_lower_vector.
20179 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
20181 (pass_lower_vector): Provide PROP_gimple_lvec.
20182 (pass_lower_vector_ssa): Likewise.
20183 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
20185 2013-03-21 Mark Wielaard <mjw@redhat.com>
20187 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
20189 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
20191 * config/i386/i386.md (*movdi_internal): Disparage slightly
20192 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
20193 conversion alternatives.
20195 2013-03-21 Jakub Jelinek <jakub@redhat.com>
20197 PR middle-end/48087
20198 * diagnostic.def (DK_WERROR): New kind.
20199 * diagnostic.h (werrorcount): Define.
20200 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
20201 promoted to DK_ERROR, increment DK_WERROR counter instead of
20203 * toplev.c (toplev_main): Call print_ignored_options even if
20204 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
20205 even if just werrorcount is non-zero.
20208 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
20210 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
20211 (string_cst_pool_decl): New function.
20212 (optimize_one_addr_into_implicit_ptr): New function.
20213 (resolve_addr_in_expr): Optimize DWARF location expression
20214 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
20215 which doesn't live in memory, but has DW_AT_location or
20216 DW_AT_const_value, or refers to a string literal, into
20217 DW_OP_GNU_implicit_pointer.
20218 (optimize_location_into_implicit_ptr): New function.
20219 (resolve_addr): If removing DW_AT_location of a variable because
20220 it was DW_OP_addr of address of the variable, but the variable doesn't
20221 live in memory, try to emit const value attribute for the initializer.
20223 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20225 * tree.h (VECTOR_TYPE_P): New macro.
20226 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
20227 TYPE_MODE): Use it.
20228 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
20229 VEC_COND_EXPR cannot be lvalues.
20230 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
20232 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20234 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
20235 Restrict the transformation to equal modes.
20237 2013-03-21 Richard Biener <rguenther@suse.de>
20239 PR tree-optimization/39326
20240 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
20241 (MEM_ANALYZABLE): Adjust.
20242 (record_mem_ref_loc): Move bitmap ops ...
20243 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
20244 unanalyzable refs, do not record locations for it.
20245 (analyze_memory_references): Allocate ref zero as shared
20247 (refs_independent_p): Do not test for unanalyzed mems here.
20248 (ref_indep_loop_p_1): Special-case disambiguation against
20249 the unanalyzed ref.
20250 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
20252 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
20254 * config/arm/arm-protos.h (tune_params): Add
20255 prefer_neon_for_64bits field.
20256 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
20257 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
20258 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
20259 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
20260 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
20261 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
20262 (arm_option_override): Handle -mneon-for-64bits new option.
20263 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
20264 (prefer_neon_for_64bits): Declare new variable.
20265 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
20266 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
20267 (arch_enabled): Handle new arch types. Remove support for onlya8
20269 (one_cmpldi2): Use new arch names.
20270 (zero_extend<mode>di2, extend<mode>di2): Ditto.
20271 * config/arm/arm.opt (mneon-for-64bits): Add option.
20272 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
20273 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
20274 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
20276 * doc/invoke.texi (-mneon-for-64bits): Document.
20278 2013-03-21 Richard Biener <rguenther@suse.de>
20280 PR tree-optimization/39326
20281 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
20282 (sort_bbs_in_loop_postorder_cmp): New function.
20283 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
20285 2013-03-21 Richard Biener <rguenther@suse.de>
20287 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
20288 (vect_insert_into_interleaving_chain): Likewise.
20289 (vect_drs_dependent_in_basic_block): Inline ...
20290 (vect_slp_analyze_data_ref_dependence): ... here. New function,
20292 (vect_analyze_data_ref_dependence): ... here. Simplify.
20293 (vect_check_interleaving): Simplify.
20294 (vect_analyze_data_ref_dependences): Likewise. Split out ...
20295 (vect_slp_analyze_data_ref_dependences): ... this new function.
20296 (dr_group_sort_cmp): New function.
20297 (vect_analyze_data_ref_accesses): Compute data-reference groups
20298 here instead of in vect_analyze_data_ref_dependence. Use
20299 a more efficient algorithm.
20300 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
20301 vect_slp_analyze_data_ref_dependences. Call
20302 vect_analyze_data_ref_accesses earlier.
20303 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
20304 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
20305 (vect_slp_analyze_data_ref_dependences): New prototype.
20307 2013-03-21 Richard Biener <rguenther@suse.de>
20309 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
20310 ref is stored in the loop.
20311 (find_refs_for_sm): Walk only over all stores.
20312 (store_motion_loop): Allocate from lim_bitmap_obstack.
20313 (store_motion): Likewise.
20315 2013-03-21 Richard Biener <rguenther@suse.de>
20317 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
20318 Update virtual SSA form.
20320 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20322 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
20323 * configure: Regenerate.
20324 * config.in: Regenerate.
20325 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
20326 if !HAVE_LD_EH_FRAME_CIEV3.
20328 2013-03-21 Richard Biener <rguenther@suse.de>
20330 * tree-cfg.c (verify_expr_no_block): New function.
20331 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
20332 nor DECL_VALUE_EXPR have locations with associated blocks.
20333 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
20334 (clear_unused_block_pointer): Remove code dealing with
20335 blocks in DECL_DEBUG_EXPR locations.
20337 2013-03-21 Richard Biener <rguenther@suse.de>
20339 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
20340 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
20341 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
20342 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
20343 instead of DECL_DEBUG_EXPR_IS_FROM.
20344 * gimplify.c (gimplify_modify_expr): Likewise.
20345 * tree-cfg.c (verify_expr_location_1): Likewise.
20346 * tree-complex.c (create_one_component_var): Likewise.
20347 * tree-sra.c (create_access_replacement): Likewise.
20348 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
20349 (clear_unused_block_pointer): Likewise.
20350 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
20351 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
20352 * var-tracking.c (var_debug_decl): Likewise.
20353 (track_expr_p): Likewise.
20354 * tree-inline.c (add_local_variables): Likewise. Set
20355 DECL_HAS_DEBUG_EXPR_P after copying it.
20356 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
20357 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
20359 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
20362 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
20363 * configure: Regenerate.
20364 * config.in: Regenerate.
20365 * config/i386/i386.md (*movdf_internal): Use
20366 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
20367 movd instead of movq mnemonic for interunit moves.
20368 (*movdi_internal): Ditto.
20370 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20372 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
20373 (abd<mode>_3): New pattern.
20374 (aba<mode>_3): New pattern.
20375 (fabd<mode>_3): New pattern.
20377 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20379 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
20380 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
20381 occurrence of REGISTER_PREFIX as its empty string.
20383 2013-03-20 Jeff Law <law@redhat.com>
20385 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
20386 addititional equivalences for equality comparisons between an SSA_NAME
20387 and a constant where the SSA_NAME was set from a widening conversion.
20389 2013-03-20 Walter Lee <walt@tilera.com>
20391 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
20393 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20395 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
20396 depending on type attribute.
20397 (*movti_internal): Ditto.
20398 (*movtf_internal): Ditto.
20399 (*movxf_internal): Ditto.
20400 (*movdf_internal): Ditto.
20401 (*movsf_internal): Ditto.
20403 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20405 * config/i386/i386.md (*movti_internal): Set prefix attribute to
20406 maybe_vex for sselog1 and ssemov types.
20407 (*movdi_internal): Reorder operand constraints.
20408 (*movsi_internal): Ditto. Set prefix attribute to
20409 maybe_vex for sselog1 and ssemov types.
20410 (*movtf_internal): Set prefix attribute to maybe_vex
20411 for sselog1 and ssemov types.
20412 (*movdf_internal): Ditto. Set prefix_data16 attribute for
20413 DImode ssemov types. Reorder operand constraints.
20414 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
20415 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
20416 attribute for SImode ssemov types. Reorder operand constraints.
20418 2013-03-20 Martin Jambor <mjambor@suse.cz>
20420 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
20421 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
20423 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
20425 * config/rs6000/predicates.md (indexed_address, update_address_mem
20426 update_indexed_address_mem): New predicates.
20427 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
20428 attribute for load/store instructions.
20429 * config/rs6000/dfp.md (movsd_store): Likewise.
20430 (movsd_load): Likewise.
20431 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
20432 (unnamed HI->DI extend define_insn): Likewise.
20433 (unnamed SI->DI extend define_insn): Likewise.
20434 (unnamed QI->SI extend define_insn): Likewise.
20435 (unnamed QI->HI extend define_insn): Likewise.
20436 (unnamed HI->SI extend define_insn): Likewise.
20437 (unnamed HI->SI extend define_insn): Likewise.
20438 (extendsfdf2_fpr): Likewise.
20439 (movsi_internal1): Likewise.
20440 (movsi_internal1_single): Likewise.
20441 (movhi_internal): Likewise.
20442 (movqi_internal): Likewise.
20443 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
20444 attribute for load/store instructions.
20445 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
20447 (mov<mode>_softfloat): Likewise.
20448 (mov<mode>_hardfloat32): Likewise.
20449 (mov<mode>_hardfloat64): Likewise.
20450 (mov<mode>_softfloat64): Likewise.
20451 (movdi_internal32): Likewise.
20452 (movdi_internal64): Likewise.
20453 (probe_stack_<mode>): Likewise.
20455 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
20457 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
20458 floating point, and decimal floating point to reload iterator.
20460 * config/rs6000/constraints.md (wl constraint): New constraints to
20461 return FLOAT_REGS if certain options are used to reduce the number
20462 of separate patterns that exist in the file.
20463 (wx constraint): Likewise.
20464 (wz constraint): Likewise.
20466 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
20467 -mdebug=reg, print wg, wl, wx, and wz constraints.
20468 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
20469 Initialize the reload functions for 64-bit binary/decimal floating
20471 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
20472 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
20473 create the buffer on the stack to overcome not having a 32-bit
20475 (rs6000_emit_move): Likewise.
20476 (rs6000_secondary_memory_needed_rtx): Likewise.
20477 (rs6000_alloc_sdmode_stack_slot): Likewise.
20478 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
20479 via xxlxor, just like DFmode 0.0.
20481 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
20482 define as 1 if we are running on a power7 or newer.
20483 (enum r6000_reg_class_enum): Add new constraints.
20485 * config/rs6000/dfp.md (movsd): Delete, combine with binary
20486 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
20487 with other moves by using conditional constraits (wg). Use LFIWZX
20488 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
20489 (movsd splitter): Likewise.
20490 (movsd_hardfloat): Likewise.
20491 (movsd_softfloat): Likewise.
20493 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
20494 binary and decimal floating point moves.
20495 (fmove_ok): New attributes to combine binary and decimal floating
20496 point moves, and to combine power6x (mfpgpr) moves along normal
20498 (real_value_to_target): Likewise.
20499 (f32_lr): Likewise.
20500 (f32_lm): Likewise.
20501 (f32_li): Likewise.
20502 (f32_sr): Likewise.
20503 (f32_sm): Likewise.
20504 (f32_si): Likewise.
20505 (movsf): Combine binary and decimal floating point moves. Combine
20506 power6x (mfpgpr) moves with other moves by using conditional
20507 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
20508 (mov<mode> for SFmode/SDmode); Likewise.
20509 (SFmode/SDmode splitters): Likewise.
20510 (movsf_hardfloat): Likewise.
20511 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
20512 (movsf_softfloat): Likewise.
20513 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
20515 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
20516 wx and wz constraints.
20518 * config/rs6000/constraints.md (wg constraint): New constraint to
20519 return FLOAT_REGS if -mmfpgpr (power6x) was used.
20521 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
20524 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
20525 -mdebug=reg, print wg, wl, wx, and wz constraints.
20526 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
20527 Initialize the reload functions for 64-bit binary/decimal floating
20529 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
20530 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
20531 create the buffer on the stack to overcome not having a 32-bit
20533 (rs6000_emit_move): Likewise.
20534 (rs6000_secondary_memory_needed_rtx): Likewise.
20535 (rs6000_alloc_sdmode_stack_slot): Likewise.
20536 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
20537 via xxlxor, just like DFmode 0.0.
20539 * config/rs6000/dfp.md (movdd): Delete, combine with binary
20540 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
20541 with other moves by using conditional constraits (wg). Use LFIWZX
20542 and STFIWX for loading SDmode on power7.
20543 (movdd splitters): Likewise.
20544 (movdd_hardfloat32): Likewise.
20545 (movdd_softfloat32): Likewise.
20546 (movdd_hardfloat64_mfpgpr): Likewise.
20547 (movdd_hardfloat64): Likewise.
20548 (movdd_softfloat64): Likewise.
20550 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
20551 64-bit binary and decimal floating point moves.
20552 (FMOVE64X): Likewise.
20553 (movdf): Combine 64-bit binary and decimal floating point moves.
20554 Combine power6x (mfpgpr) moves with other moves by using
20555 conditional constraits (wg).
20556 (mov<mode> for DFmode/DDmode): Likewise.
20557 (DFmode/DDmode splitters): Likewise.
20558 (movdf_hardfloat32): Likewise.
20559 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
20560 (movdf_softfloat32): Likewise.
20561 (movdf_hardfloat64_mfpgpr): Likewise.
20562 (movdf_hardfloat64): Likewise.
20563 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
20564 (movdf_softfloat64): Likewise.
20565 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
20566 (reload_<mode>_load): Move to later in the file so they aren't in
20567 the middle of the floating point move insns.
20568 (reload_<mode>_store): Likewise.
20570 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
20573 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
20574 constraint if -mdebug=reg.
20575 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
20576 Enable using dd reload support if needed.
20578 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
20579 binary and decimal floating point moves in rs6000.md.
20580 (movtd_internal): Likewise.
20582 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
20583 decimal floating point moves.
20585 (movtf_internal): Likewise.
20586 (mov<mode>_internal, TDmode/TFmode): Likewise.
20587 (movtf_softfloat): Likewise.
20588 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
20590 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
20591 movdi_internal64, using wg constraint for move direct operations.
20592 (movdi_internal64): Likewise.
20594 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
20595 MODES_TIEABLE_P for selected modes. Print the numerical value of
20596 the various virtual registers. Use GPR/FPR first/last values,
20597 instead of hard coding the register numbers. Print which modes
20598 have reload functions registered.
20599 (rs6000_option_override_internal): If -mdebug=reg, trace the options
20600 settings before/after setting cpu, target and subtarget settings.
20601 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
20602 and for secondary reload failures in rs6000_secondary_reload_inner.
20603 (rs6000_secondary_reload_fail): Likewise.
20604 (rs6000_secondary_reload_inner): Likewise.
20606 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
20607 macros for first/last GPR and FPR registers.
20608 (LAST_GPR_REGNO): Likewise.
20609 (FIRST_FPR_REGNO): Likewise.
20610 (LAST_FPR_REGNO): Likewise.
20612 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
20613 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
20614 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
20615 (vcond<mode><mode>): Likewise.
20616 (vcondu<mode><mode>): Likewise.
20617 (vector_gtu<mode>): Likewise.
20618 (vector_gte<mode>): Likewise.
20619 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
20620 to prevent the compiler from converting DImode operations to TImode.
20621 (ior<mode>3): Likewise.
20622 (and<mode>3): Likewise.
20623 (one_cmpl<mode>2): Likewise.
20624 (nor<mode>3): Likewise.
20625 (andc<mode>3): Likewise.
20627 * config/rs6000/constraints.md (wt constraint): New constraint
20628 that returns VSX_REGS if TImode is allowed in VSX registers.
20630 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
20631 constant under VSX.
20633 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
20634 similar to TImode, but it is restricted to being in the GPRs.
20636 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
20637 TImode to occupy a single VSX register.
20639 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
20640 -mvsx-timode for power7/power8.
20641 (power7 cpu): Likewise.
20642 (power8 cpu): Likewise.
20644 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
20645 sure that TFmode/TDmode take up two registers if they are ever
20646 allowed in the upper VSX registers.
20647 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
20649 (rs6000_init_hard_regno_mode_ok): Likewise.
20650 (rs6000_debug_reg_global): Add debugging for PTImode and wt
20651 constraint. Print if LRA is turned on.
20652 (rs6000_option_override_internal): Give an error if -mvsx-timode
20653 and VSX is not enabled.
20654 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
20655 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
20656 to reg+offset addressing. Use PTImode when checking offset
20657 addresses for validity.
20658 (reg_offset_addressing_ok_p): Likewise.
20659 (rs6000_legitimate_offset_address_p): Likewise.
20660 (rs6000_legitimize_address): Likewise.
20661 (rs6000_legitimize_reload_address): Likewise.
20662 (rs6000_legitimate_address_p): Likewise.
20663 (rs6000_eliminate_indexed_memrefs): Likewise.
20664 (rs6000_emit_move): Likewise.
20665 (rs6000_secondary_reload): Likewise.
20666 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
20667 reloads to fpr registers to continue to use reg+offset addressing,
20668 but 64-bit reloads to altivec registers need reg+reg addressing.
20669 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
20670 it. Treat LO_SUM like a PLUS operation.
20671 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
20672 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
20673 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
20674 registers to share a register with a smaller sized type, since VSX
20675 puts scalars in the upper 64-bits.
20676 (print_operand): Add support for PTImode.
20677 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
20678 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
20679 registers, but don't have arithmetic support.
20680 (rs6000_memory_move_cost): Add test for VSX.
20681 (rs6000_opt_masks): Add -mvsx-timode.
20683 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
20686 (VSr): Use wt constraint for TImode.
20687 (VSv): Drop TImode support.
20688 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
20689 (vsx_movti_64bit): Likewise.
20690 (vsx_movti_32bit): Likewise.
20691 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
20692 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
20693 one '?' on the appropriate output constraint. Do not allow TImode
20694 logical operations on 32-bit systems.
20695 (vsx_ior<mode>3): Likewise.
20696 (vsx_xor<mode>3): Likewise.
20697 (vsx_one_cmpl<mode>2): Likewise.
20698 (vsx_nor<mode>3): Likewise.
20699 (vsx_andc<mode>3): Likewise.
20700 (vsx_concat_<mode>): Likewise.
20701 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
20703 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
20704 OPTION_MASK_VSX_TIMODE.
20705 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
20706 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
20708 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
20709 (TI2 iterator): New iterator for TImode, PTImode.
20710 (wd mode attribute): Add values for vector types.
20711 (movti_string): Replace TI move operations with operations for TImode
20712 and PTImode. Add support for TImode being allowed in VSX registers.
20713 (mov<mode>_string, TImode/PTImode): Likewise.
20714 (movti_ppc64): Likewise.
20715 (mov<mode>_ppc64, TImode/PTImode): Likewise.
20716 (TI mode splitters): Likewise.
20718 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
20721 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
20723 PR tree-optimization/56355
20724 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
20725 Also handle integers with undefined overflow.
20727 2013-03-20 Catherine Moore <clm@codesourcery.com>
20728 Maciej W. Rozycki <macro@codesourcery.com>
20729 Tom de Vries <tom@codesourcery.com>
20730 Nathan Sidwell <nathan@codesourcery.com>
20731 Iain Sandoe <iain@codesourcery.com>
20732 Nathan Froyd <froydnj@codesourcery.com>
20733 Chao-ying Fu <fu@mips.com>
20735 * doc/extend.texi: (micromips, nomicromips, nocompression):
20736 Document new function attributes.
20737 * doc/invoke.texi (minterlink-compressed, mmicromips,
20738 m14k, m14ke, m14kec): Document new options.
20739 (minterlink-mips16): Update documentation.
20740 * doc/md.texi (ZC, ZD): Document new constraints.
20741 * configure.ac (gcc_cv_as_micromips): Check if linker
20742 supports the .set micromips directive.
20743 * configure: Regenerate.
20744 * config.in: Regenerate.
20745 * config/mips/mips-tables.opt: Regenerate.
20746 * config/mips/micromips.md: New file.
20747 * constraints.md (ZC, ZD): New constraints.
20748 * config/mips/predicates.md (movep_src_register): New predicate.
20749 (movep_src_operand): New predicate.
20750 (non_volatile_mem_operand): New predicate.
20751 * config/mips/mips.md (multimem): New type.
20752 (length): Differentiate between 17-bit and 18-bit branch offsets.
20753 (MOVEP1, MOVEP2): New mode iterator.
20754 (mov_<load>l): Use ZC constraint.
20755 (mov_<load>r): Likewise.
20756 (mov_<store>l): Likewise.
20757 (mov_<store>r): Likewise.
20758 (*branch_equality<mode>_inverted): Add microMIPS support.
20759 (*branch_equality<mode>): Likewise.
20760 (*jump_absolute): Likewise.
20761 (indirect_jump_<mode>): Likewise.
20762 (tablejump_<mode>): Likewise.
20763 (<optab>_internal): Likewise.
20764 (sibcall_internal): Likewise.
20765 (sibcall_value_internal): Likewise.
20766 (prefetch): Use constraint ZD.
20767 * config/mips/mips.opt (minterlink-compressed): New option.
20768 (minterlink-mips16): Now an alias for minterlink-compressed.
20769 (mmicromips): New option.
20770 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20771 (compare_and_swap_12): Likewise.
20772 (sync_add<mode>): Likewise.
20773 (sync_<optab>_12): Likewise.
20774 (sync_old_<optab>_12): Likewise.
20775 (sync_new_<optab>_12): Likewise.
20776 (sync_nand_12): Likewise.
20777 (sync_old_nand_12): Likewise.
20778 (sync_new_nand_12): Likewise.
20779 (sync_sub<mode>): Likewise.
20780 (sync_old_add<mode>): Likewise.
20781 (sync_old_sub<mode>): Likewise.
20782 (sync_new_add<mode>): Likewise.
20783 (sync_new_sub<mode>): Likewise.
20784 (sync_<optab><mode>): Likewise.
20785 (sync_old_<optab><mode>): Likewise.
20786 (sync_new_<optab><mode>): Likewise.
20787 (sync_nand<mode>): Likewise.
20788 (sync_old_nand<mode>): Likewise.
20789 (sync_new_nand<mode>): Likewise.
20790 (sync_lock_test_and_set<mode>): Likewise.
20791 (test_and_set_12): Likewise.
20792 (atomic_compare_and_swap<mode>): Likewise.
20793 (atomic_exchange<mode>_llsc): Likewise.
20794 (atomic_fetch_add<mode>_llsc): Likewise.
20795 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20796 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20797 (umips_save_restore_pattern_p): Likewise.
20798 (umips_load_store_pair_p): Likewise.
20799 (umips_output_load_store_pair): Likewise.
20800 (umips_movep_target_p): Likewise.
20801 (umips_12bit_offset_address_p): Likewise.
20802 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20803 (mips_base_mips16): Rename this...
20804 (mips_base_compression_flags): ...to this. Update all uses.
20805 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20806 (mips_mips16_decl_p): Delete.
20807 (mips_nomips16_decl_p): Delete.
20808 (mips_get_compress_on_flags): New function.
20809 (mips_get_compress_off_flags): New function.
20810 (mips_get_compress_mode): New function.
20811 (mips_get_compress_on_name): New function.
20812 (mips_get_compress_off_name): New function.
20813 (mips_insert_attributes): Support multiple compression types.
20814 (mips_merge_decl_attributes): Likewise.
20815 (umips_12bit_offset_address_p): New function.
20816 (mips_start_function_definition): Emit .set micromips directive.
20817 (mips_call_may_need_jalx_p): New function.
20818 (mips_function_ok_for_sibcall): Add microMIPS support.
20819 (mips_print_operand_punctuation): Support short delay slots and
20821 (umips_swm_mask, umips_swm_encoding): New.
20822 (umips_build_save_restore): New function.
20823 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20824 (was_mips16_p): Remove.
20825 (old_compression_mode): New.
20826 (mips_set_compression_mode): New function.
20827 (mips_set_current_function): Add microMIPS support.
20828 (mips_option_override): Likewise.
20829 (umips_save_restore_pattern_p): New function.
20830 (umips_output_save_restore): New function.
20831 (umips_load_store_pair_p_1): New function.
20832 (umips_load_store_pair_p): New function.
20833 (umips_output_load_store_pair_1): New function.
20834 (umips_output_load_store_pair): New function.
20835 (umips_movep_target_p) New function.
20836 (mips_prepare_pch_save): Add microMIPS support.
20837 * config/mips/mips.h (TARGET_COMPRESSION): New.
20838 (TARGET_CPU_CPP_BUILTINS): Update macro
20839 to use new compression flags and to support microMIPS.
20840 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20841 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20842 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20843 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20844 (ASM_SPEC): Support mmicromips and mno-micromips.
20845 (M16STORE_REG_P): New macro.
20846 (MIPS_CALL): Support TARGET_MICROMIPS.
20847 (MICROMIPS_J): New macro.
20848 (mips_base_mips16): Rename this...
20849 (mips_base_compression_flags): ...to this.
20850 (UMIPS_12BIT_OFFSET_P): New macro.
20851 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20852 (MULTILIB_DIRNAMES): Likewise.
20853 2013-03-20 Richard Biener <rguenther@suse.de>
20855 PR tree-optimization/56661
20856 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
20857 the result does not have to be distinct.
20859 2013-03-20 Richard Biener <rguenther@suse.de>
20861 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
20864 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20865 Steven Bosscher <steven@gcc.gnu.org>
20867 PR rtl-optimization/56605
20868 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
20870 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20873 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
20874 that require movd instead of movq.
20876 2013-03-20 Richard Biener <rguenther@suse.de>
20878 * tree-ssa-structalias.c (struct variable_info): Add pointer
20879 to the first field of an aggregate with sub-vars. Make
20880 this and the pointer to the next subfield its ID.
20881 (vi_next): New function.
20882 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
20883 storedanything_id, integer_id): Increment by one.
20884 (new_var_info, get_call_vi, lookup_call_clobber_vi,
20885 get_call_clobber_vi): Adjust.
20886 (solution_set_expand): Simplify and speedup.
20887 (solution_set_add): Inline into ...
20888 (set_union_with_increment): ... this. Adjust accordingly.
20889 (do_sd_constraint): Likewise.
20890 (do_ds_constraint): Likewise.
20891 (do_complex_constraint): Simplify.
20892 (build_pred_graph): Adjust.
20893 (solve_graph): Likewise. Simplify and speedup.
20894 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
20895 get_constraint_for_component_ref, get_constraint_for_1,
20896 first_vi_for_offset, first_or_preceding_vi_for_offset,
20897 create_function_info_for, create_variable_info_for_1,
20898 create_variable_info_for, intra_create_variable_infos): Adjust.
20899 (init_base_vars): Push NULL for ID zero.
20900 (compute_points_to_sets): Adjust.
20902 2013-03-20 Richard Biener <rguenther@suse.de>
20904 * cfgloop.c (verify_loop_structure): Streamline and avoid
20905 ICEing on corrupt loop tree.
20906 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
20909 2013-03-20 Richard Biener <rguenther@suse.de>
20911 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
20912 check whether an SSA update is needed.
20914 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
20916 * config/mips/constraints.md (T): Rename to...
20920 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
20921 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
20923 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20925 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
20926 (*subsi3_carryin_uxtw): Likewise.
20928 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20930 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
20931 (*rorsi3_insn_uxtw): Likewise.
20933 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20935 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
20936 (*extrsi5_insn_uxtw): Likewise.
20938 2013-03-19 Richard Biener <rguenther@suse.de>
20940 PR tree-optimization/56273
20941 * passes.c (init_optimization_passes): Move second VRP after DOM.
20943 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
20945 * config/i386/i386.md (*movti_internal): Merge from
20946 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
20947 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
20948 nox64 isa attributes.
20950 2013-03-18 Richard Biener <rguenther@suse.de>
20952 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
20954 (merge_node_constraints): Likewise.
20955 (build_succ_graph): Likewise.
20956 (valid_graph_edge): Inline into single caller.
20957 (unify_nodes): Likewise. Use bitmap_set_bit return value
20959 (scc_visit): Fix formatting and variable use.
20960 (do_sd_constraint): Use gcc_checking_assert.
20961 (do_ds_constraint): Likewise.
20962 (do_complex_constraint): Likewise.
20963 (condense_visit): Likewise. Cleanup.
20964 (dump_pred_graph): New function.
20965 (perform_var_substitution): Dump the pred-graph before
20966 variable substitution.
20967 (find_equivalent_node): Use gcc_checking_assert.
20968 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
20970 2013-03-18 Richard Biener <rguenther@suse.de>
20972 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
20973 Remove cond_expr_stmt_list argument and do not gimplify the
20975 (vect_loop_versioning): Adjust.
20976 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
20977 Cleanup to use less temporaries.
20978 (vect_create_data_ref_ptr): Cleanup.
20980 2013-03-18 Jakub Jelinek <jakub@redhat.com>
20982 PR tree-optimization/56635
20983 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
20984 require types_compatible_p types.
20986 2013-03-18 Nick Clifton <nickc@redhat.com>
20988 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
20989 spurious backslash.
20991 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
20992 Add missing line to comment describing function.
20994 2013-03-18 Richard Biener <rguenther@suse.de>
20996 PR tree-optimization/56210
20997 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20998 Handle string / character search functions.
20999 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
21001 2013-03-18 Richard Biener <rguenther@suse.de>
21003 PR middle-end/56483
21004 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
21005 and implement properly.
21006 * gimple.h (gimple_cond_single_var_p): Remove.
21008 2013-03-18 Richard Biener <rguenther@suse.de>
21010 * tree-data-ref.h (find_data_references_in_loop): Declare.
21011 * tree-data-ref.c (get_references_in_stmt): Use a stack
21012 vector pre-allocated in the callers.
21013 (find_data_references_in_stmt): Adjust.
21014 (graphite_find_data_references_in_stmt): Likewise.
21015 (create_rdg_vertices): Likewise.
21016 (find_data_references_in_loop): Export.
21017 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
21018 Compute dependences here...
21019 (vect_analyze_data_refs): ...not here. When we encounter
21020 a non-vectorizable data reference in basic-block vectorization
21021 truncate the data reference vector. Do not bother to
21022 fixup data-dependence information for gather loads.
21023 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
21024 of data references, as reported.
21026 2013-03-18 Richard Biener <rguenther@suse.de>
21028 PR tree-optimization/3713
21029 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
21030 has_constants and expr.
21031 (stmt_has_constants): Properly valueize SSA names when deciding
21032 whether the stmt has constants.
21034 2013-03-18 Richard Biener <rguenther@suse.de>
21036 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
21037 whole function when there is nothing to do.
21038 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
21039 * tree-vectorizer.c (vectorize_loops): Update virtual and
21040 loop-closed SSA once.
21041 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
21043 2013-03-18 Richard Biener <rguenther@suse.de>
21045 PR middle-end/56113
21046 * domwalk.c (bb_postorder): New global static.
21047 (cmp_bb_postorder): New function.
21048 (walk_dominator_tree): Replace scheme imposing an order for
21049 visiting dominator sons by one sorting them at the time they
21050 are pushed on the stack.
21052 2013-03-18 Richard Biener <rguenther@suse.de>
21054 PR tree-optimization/39326
21055 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
21056 (struct mem_ref): Replace mem member with ao_ref typed member.
21057 (MEM_ANALYZABLE): Adjust.
21058 (memref_eq): Likewise.
21059 (mem_ref_alloc): Likewise.
21060 (gather_mem_refs_stmt): Likewise.
21061 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
21062 (execute_sm_if_changed_flag_set): Adjust.
21063 (execute_sm): Likewise.
21064 (ref_always_accessed_p): Likewise.
21065 (refs_independent_p): Likewise.
21066 (can_sm_ref_p): Likewise.
21068 2013-03-18 Jakub Jelinek <jakub@redhat.com>
21071 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
21072 return 1 even for !unsignedp.
21074 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
21076 * config/i386/i386.md (isa): Add x64 and nox64.
21077 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
21078 (*pushtf): Enable *roF alternative for x64 isa only.
21079 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
21080 mode attribute of integer alternatives to DImode for TARGET_64BIT.
21081 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
21082 (*movtf_internal): Merge from *movtf_internal_rex64 and
21083 *movtf_internal_sse. Use x64 and nox64 isa attributes.
21084 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
21085 nox64 isa attributes.
21086 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
21087 nox64 isa attributes.
21088 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
21090 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
21092 * config/alpha/alpha.c (TARGET_LRA_P): New define.
21094 2013-03-17 Jakub Jelinek <jakub@redhat.com>
21097 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
21098 class names. Remove trailing comma after "ALL_REGS".
21100 2013-03-16 Jan Hubicka <jh@suse.cz>
21102 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
21103 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
21104 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
21105 of cgraph_get_create_node.
21106 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
21108 2013-03-16 Jason Merrill <jason@redhat.com>
21111 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
21112 with DW_AT_default_value.
21114 2013-03-16 Jakub Jelinek <jakub@redhat.com>
21116 * BASE-VER: Set to 4.9.0.
21118 2013-03-14 Andi Kleen <ak@linux.intel.com>
21121 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
21122 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
21123 Document _x* TSX intrinsics.
21125 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
21126 David Holsgrove <david.holsgrove@xilinx.com>
21128 * configure.ac: Add MicroBlaze TLS support detection.
21129 * configure: Regenerate.
21130 * config/microblaze/microblaze-protos.h
21131 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
21132 symbol_mentioned_p, label_mentioned_p): Add prototypes.
21133 * config/microblaze/microblaze.c (microblaze_address_type): Add
21134 ADDRESS_TLS and tls_reloc address types.
21135 (microblaze_address_info): Add tls_reloc.
21136 (TARGET_HAVE_TLS): Define.
21137 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
21138 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
21139 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
21140 load_tls_operand, microblaze_call_tls_get_addr,
21141 microblaze_legitimize_tls_address): New functions.
21142 (microblaze_classify_unspec): Handle UNSPEC_TLS.
21143 (get_base_reg): Use microblaze_tls_symbol_p.
21144 (microblaze_classify_address): Handle TLS.
21145 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
21146 label_mentioned_p and microblaze_tls_referenced_p.
21147 (microblaze_legitimize_address): Handle TLS.
21148 (microblaze_address_insns): Handle ADDRESS_TLS.
21149 (pic_address_needs_scratch): Handle TLS.
21150 (print_operand_address): Handle TLS.
21151 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
21152 (microblaze_expand_move): Handle TLS.
21153 (microblaze_legitimate_constant_p): Check
21154 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
21155 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
21156 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
21157 (PIC_OFFSET_TABLE_REGNUM): Set.
21158 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
21159 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
21160 (addsi3, movsi_internal2, movdf_internal): Update constraints
21161 * config/microblaze/predicates.md (arith_plus_operand): Define
21162 (move_operand): Redefine as move_src_operand,
21163 check microblaze_tls_referenced_p.
21165 2013-03-14 Ian Bolton <ian.bolton@arm.com>
21167 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
21168 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
21170 2013-03-14 Ian Bolton <ian.bolton@arm.com>
21172 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
21175 2013-03-14 Jakub Jelinek <jakub@redhat.com>
21177 PR tree-optimization/53265
21178 * common.opt (Waggressive-loop-optimizations): New option.
21179 * tree-ssa-loop-niter.c: Include tree-pass.h.
21180 (do_warn_aggressive_loop_optimizations): New function.
21181 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
21182 if number_of_latch_executions returned constant.
21183 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
21184 early. If number_of_latch_executions returned constant, set
21185 nb_iterations_upper_bound back to it.
21186 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
21188 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
21189 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
21191 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
21192 (MULTILIB_OSDIRNAMES): Set.
21193 * genmultilib: If defaultosdirname doesn't start with :: , set
21194 defaultosdirname2 instead, clear it and emit two . multilib_raw
21195 entries instead of just one.
21197 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
21199 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
21200 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
21201 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
21202 (SUBTARGET_OVERRIDE_OPTIONS): New.
21204 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
21207 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
21208 (musermode): Convert to Var(TARGET_USERMODE).
21209 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
21210 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
21211 * config/sh/sh.c (sh_option_override): Use
21212 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
21213 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
21215 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
21217 (udivsi3_i4_single, divsi3_i4_single): Use
21218 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
21220 2013-03-13 Dave Korn <dave.korn.cygwin@....>
21222 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
21225 2013-03-13 Richard Biener <rguenther@suse.de>
21227 PR tree-optimization/56608
21228 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
21229 calls when vectorizing basic-blocks.
21231 2013-03-13 Jakub Jelinek <jakub@redhat.com>
21234 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
21237 2013-03-12 Jakub Jelinek <jakub@redhat.com>
21239 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
21241 2013-03-11 Jan Hubicka <jh@suse.cz>
21244 * lto-streamer-out.c (output_symbol_p): Skip references from
21245 constructors of external variables.
21247 2013-03-11 Jan Hubicka <jh@suse.cz>
21249 PR middle-end/56571
21250 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
21252 * emit-rtl.c (verify_rtx_sharing): Likewise.
21253 (copy_insn_1): Likewise.
21254 * rtl.c (copy_rtx): Likewise.
21256 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
21259 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
21260 output_operand_lossage message.
21262 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
21265 * arm.c (shift_op): Validate RTL pattern on the fly.
21266 (arm_print_operand, case 'S'): Don't use shift_operator to validate
21269 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21272 * config/pa/pa.md (call_value): Check for calls to powf and direct to
21273 new call patterns that clobber %fr12.
21274 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
21275 split and postreload patterns.
21276 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
21277 registers %fr12 and %fr12R as call used.
21279 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
21281 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
21282 (canon_address, record_store, replace_read, check_mem_read_rtx,
21283 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
21284 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
21285 rest_of_handle_dse): Likewise.
21287 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
21289 PR middle-end/56524
21290 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
21292 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
21293 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
21294 (save_optabs_if_changed): Replace with...
21295 (init_tree_optimization_optabs): ...this.
21296 * optabs.c (save_optabs_if_changed): Rename to...
21297 (init_tree_optimization_optabs): ...this. Take the optimization node
21298 as argument. Do nothing if the base optabs are already correct.
21299 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
21300 to recompute optabs.
21301 * function.h (function): Remove optabs field.
21302 * function.c (invoke_set_current_function_hook): Call
21303 init_tree_optimization_optabs. Use the result to initialize
21306 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
21308 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
21309 if GTMA_HAS_NO_INSTRUMENTATION.
21310 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
21311 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
21312 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
21313 * gimple-pretty-print.c (dump_gimple_transaction): Handle
21314 GTMA_HAS_NO_INSTRUMENTATION.
21316 2013-03-08 Jakub Jelinek <jakub@redhat.com>
21318 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
21321 2013-03-08 Marek Polacek <polacek@redhat.com>
21322 Jakub Jelinek <jakub@redhat.com>
21324 PR tree-optimization/56478
21325 * predict.c (is_comparison_with_loop_invariant_p): Change the
21326 type of loop_step to tree.
21327 (predict_loops): Adjust.
21328 (predict_iv_comparison): Perform the computations on double_ints.
21330 2013-03-08 Richard Biener <rguenther@suse.de>
21332 PR tree-optimization/56570
21333 * tree-cfg.c (verify_expr_location_1): Verify locations for
21335 * tree-sra.c (create_access_replacement): Strip locations
21336 from DECL_DEBUG_EXPRs.
21338 2013-03-08 Richard Biener <rguenther@suse.de>
21340 * tree-inline.c (expand_call_inline): Do not associate
21341 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
21342 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
21344 2013-03-08 Richard Biener <rguenther@suse.de>
21346 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
21347 or block changes with -Og. Fix for location / block encoding
21348 changes and PHI arguments with locations.
21350 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
21352 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
21354 (struct output_info): Likewise.
21355 (register_overhead): Remove bad gcc_assert.
21356 (bitmap_find_bit): If there is only a single bitmap element, do not
21357 count a miss as a search.
21358 (print_statistics): Update for counter type changes.
21359 (dump_bitmap_statistics): Likewise. Print headers such that they
21360 are properly lined up with the printed counters.
21362 2013-03-07 Jakub Jelinek <jakub@redhat.com>
21364 PR tree-optimization/56559
21365 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
21366 check that it has only a single use.
21368 2013-03-07 Richard Biener <rguenther@suse.de>
21370 * doc/invoke.texi (fwhole-program): Discourage use in combination
21373 2013-03-06 Jakub Jelinek <jakub@redhat.com>
21375 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
21377 PR tree-optimization/56539
21378 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
21379 instead of GSI_CONTINUE_LINKING as last argument to
21380 force_gimple_operand_gsi. Adjust function comment.
21382 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
21385 PR middle-end/56548
21386 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
21387 promoted mode, convert the result back to the original mode.
21389 2013-03-06 Richard Biener <rguenther@suse.de>
21391 PR middle-end/56294
21392 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
21393 (insert_updated_phi_nodes_compare_uids): New function.
21394 (update_ssa): Sort symbols_to_rename after UID before
21395 traversing it to insert PHI nodes.
21397 2013-03-06 Richard Biener <rguenther@suse.de>
21399 PR middle-end/50494
21400 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
21401 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
21404 2013-02-13 Richard Biener <rguenther@suse.de>
21407 * varasm.c (output_constant_def_1): Get the decl representing
21408 the constant as argument.
21409 (output_constant_def): Wrap output_constant_def_1.
21410 (make_decl_rtl): Use output_constant_def_1 with the decl
21411 representing the constant.
21412 (build_constant_desc): Optionally re-use a decl already
21413 representing the constant.
21414 (tree_output_constant_def): Adjust.
21416 2013-03-06 Joey Ye <joey.ye@arm.com>
21419 * gcc.c (convert_white_space): New function.
21420 (main): Handles white space in function name.
21422 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
21425 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
21426 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
21427 to SH_DIV_CALL_TABLE for TARGET_SH2.
21428 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
21430 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
21431 call-table options.
21433 2013-03-05 Sterling Augustine <saugustine@google.com>
21434 Cary Coutant <ccoutant@google.com>
21437 * dwarf2out.c (resolve_addr): Don't call
21438 remove_loc_list_addr_table_entries a second time for the same
21441 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21444 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
21445 (avoid_complex_debug_insns): New function.
21446 (expand_debug_locations): Call it.
21448 PR rtl-optimization/56484
21449 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
21450 lifetimes of hard registers on small register class machines.
21452 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21454 * config/microblaze/microblaze-protos.h: Rename
21455 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
21456 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
21458 (microblaze_fast_interrupt_function_p): New function.
21459 (microblaze_is_interrupt_handler): Rename to
21460 microblaze_is_interrupt_variant and add fast_interrupt check.
21461 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
21462 (save_restore_insns): Likewise.
21463 (compute_frame_size): Likewise.
21464 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
21465 (microblaze_globalize_label): Likewise.
21466 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
21467 * config/microblaze/microblaze.md: Use wrapper
21468 microblaze_is_interrupt_variant.
21470 2013-03-05 Kai Tietz <ktietz@redhat.com>
21472 * sdbout.c (sdbout_one_type): Switch to current function's section
21473 supporting cold/hot.
21475 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21477 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
21480 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21482 PR middle-end/56461
21483 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
21484 if VALGRIND_GET_VBITS is defined, temporarily make object
21485 memory all defined, and restore previous valgrind addressability
21486 and definability afterwards. Free this_object at the end.
21488 PR middle-end/56461
21489 * lra.c (lra): Call lra_clear_live_ranges if live_p,
21490 right before calling lra_create_live_ranges, also call it
21491 when clearing live_p. Only call lra_clear_live_ranges
21492 at the end if live_p.
21494 PR middle-end/56461
21495 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
21497 2013-03-05 Richard Biener <rguenther@suse.de>
21499 PR tree-optimization/56521
21500 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
21503 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
21506 * except.h (remove_unreachable_eh_regions): New prototype.
21507 * except.c (remove_eh_handler_splicer): New function, split out
21508 of remove_eh_handler.
21509 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
21510 warning about running it on many EH regions one at a time.
21511 (remove_unreachable_eh_regions_worker): New function, walk the
21512 EH tree in depth-first order and remove non-marked regions.
21513 (remove_unreachable_eh_regions): New function.
21514 * tree-eh.c (mark_reachable_handlers): New function, split out
21515 from remove_unreachable_handlers.
21516 (remove_unreachable_handlers): Use mark_reachable_handlers and
21517 remove_unreachable_eh_regions.
21518 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
21519 and remove_unreachable_eh_regions.
21521 2013-03-05 Richard Biener <rguenther@suse.de>
21523 PR middle-end/56525
21524 * loop-init.c (fix_loop_structure): Remove loops in two stages,
21525 not freeing them until the end.
21527 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21529 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
21531 2013-03-05 Richard Biener <rguenther@suse.de>
21533 PR tree-optimization/56270
21534 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
21535 of loads after scheduling an SLP instance.
21537 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21539 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
21541 (check_gcc_parallelize): Run guality.exp as a separate job from
21542 vect.exp with unsorted.exp and $(dg_target_exps) separately from
21543 struct-layout-1.exp with stackalign.exp.
21545 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
21547 PR middle-end/56461
21548 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
21549 load_index sbitmap even if some bit in it isn't set.
21551 PR middle-end/56461
21552 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
21553 (discover_iteration_bound_by_body_walk): Change queues to
21554 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
21555 spelling in comment. Call safe_push on queues[bound_index] directly.
21556 Release queues[queue_index] in every iteration unconditionally.
21557 Release bounds vector.
21559 PR middle-end/56461
21560 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
21561 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
21562 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
21565 2013-03-05 Richard Biener <rguenther@suse.de>
21568 * tree-inline.c (remap_blocks_to_null): New function.
21569 (expand_call_inline): When expanding a call stmt without
21570 an associated block inline remap all callee blocks to NULL.
21572 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21574 PR rtl-optimization/56494
21575 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
21576 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
21577 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
21579 PR middle-end/56461
21580 * sel-sched-ir.c (free_sched_pools): Release
21581 succs_info_pool.stack[succs_info_pool.max_top] vectors too
21582 if succs_info_pool.max_top isn't -1.
21585 * opts.c (opts_obstack, opts_concat): Moved to...
21586 * opts-common.c (opts_obstack, opts_concat): ... here.
21588 2013-03-04 Jakub Jelinek <jakub@redhat.com>
21590 PR middle-end/56461
21591 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
21593 2013-03-04 Martin Jambor <mjambor@suse.cz>
21595 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
21596 all appropriate places.
21598 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
21600 PR tree-optimization/56424
21601 * ipa-split.c (split_function): Do not set the RSO flag if result is
21602 not by reference and its type is a register type.
21604 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
21606 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
21607 (microblaze_legitimate_pic_operand): Likewise
21608 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
21609 new function microblaze_legitimate_pic_operand
21610 * config/microblaze/microblaze-protos.h
21611 (microblaze_legitimate_pic_operand): Declare.
21613 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21615 * config/microblaze/predicates.md (call_insn_simple_operand):
21616 New predicate for supported rtx code types.
21617 * config/microblaze/microblaze.md (call_internal1): Use
21618 call_insn_simple_operand predicate.
21620 2013-03-04 Jakub Jelinek <jakub@redhat.com>
21622 PR middle-end/56461
21623 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
21624 partitions.ordered_remove.
21626 PR middle-end/56461
21627 * tree-vect-stmts.c (vectorizable_conversion): Don't call
21628 vec_oprnds0.create (1) for modifier == NONE.
21630 PR middle-end/56461
21631 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
21632 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
21633 vec_oprnds1 right before pushing anything to it for
21636 PR middle-end/56461
21637 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
21638 set nbbs to 0 instead of having separate code path.
21639 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
21640 instead of false as last argument if returning NULL.
21642 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
21644 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
21645 the attribute is now called "target" instead of "option".
21646 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
21647 * doc/tm.texi.in (Target Attributes): Likewise document the correct
21648 attribute/pragma name for TARGET_OPTION_VALID_P and
21649 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
21650 * doc/tm.texi: Regenerated.
21652 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
21654 * config/microblaze/microblaze.c:
21655 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
21656 * config/microblaze/microblaze.h: Add -mxl-reorder to
21658 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
21659 instructions emitted if TARGET_REORDER.
21660 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
21661 or 0 for -m/-mno case, but initialises as 2 to detect default use case
21664 2013-03-01 Xinliang David Li <davidxl@google.com>
21666 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
21669 2013-03-01 Jakub Jelinek <jakub@redhat.com>
21671 PR middle-end/56461
21672 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
21673 vector even when returning true. Fix up function comment formatting.
21675 PR middle-end/56461
21676 * ira-build.c (ira_loop_nodes_count): New variable.
21677 (create_loop_tree_nodes): Initialize it.
21678 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
21680 PR middle-end/56461
21681 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
21682 method on dr_chain and result_chain.
21683 * tree-vect-stmts.c (vectorizable_store): Only call
21684 result_chain.create if j == 0.
21686 PR middle-end/56461
21687 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
21688 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
21689 before overwriting it.
21691 2013-03-01 Tobias Burnus <burnus@net-b.de>
21693 * doc/extended.texi (C Extensions): Change order in @menu
21695 (Other MIPS Built-in Functions): Move last MIPS entry before
21696 "picoChip Built-in Functions".
21697 (SH Built-in Functions): Move after RX Built-in Functions.
21698 * doc/gcc.texi (Introduction): Change order in @menu
21700 * doc/md.texi (Constraints): Ditto.
21701 * gty.texi (Type Information): Ditto.
21702 (User-provided marking routines for template types): Make
21704 * doc/invoke.texi (AArch64 Options): Move before
21705 "Adapteva Epiphany Options".
21707 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
21708 Jakub Jelinek <jakub@redhat.com>
21711 * asan.c (gate_asan): Lookup no_sanitize_address instead of
21712 no_address_safety_analysis attribute.
21713 * doc/extend.texi (no_address_safety_attribute): Rename to
21714 no_sanitize_address attribute, mention no_address_safety_analysis
21715 attribute as deprecated alias.
21717 2013-02-28 Jakub Jelinek <jakub@redhat.com>
21719 PR middle-end/56461
21720 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
21721 type to vec<vec<tree> > *.
21722 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
21723 to be vec<tree> instead of vec<tree> *, set vec_defs
21724 to vNULL and call vec_defs.create (number_of_vects), adjust other
21726 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
21727 vectorizable_condition): Adjust vect_get_slp_defs callers.
21729 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21731 * config/aarch64/aarch64.c
21732 (aarch64_float_const_representable): Remove unused variable.
21734 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21736 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
21738 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21740 * config/aarch64/aarch64-builtins.c
21741 (aarch64_init_simd_builtins): Make static.
21743 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21745 * config/aarch64/aarch64.c
21746 (aarch64_simd_make_constant): Make static.
21748 2013-02-28 Martin Jambor <mjambor@suse.cz>
21750 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
21751 with no initialization to the RHS of debug statements.
21753 2013-02-28 Martin Jambor <mjambor@suse.cz>
21755 PR tree-optimization/56294
21756 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
21758 (get_access_replacement): Do not call create_access_replacement.
21759 Assert a replacement exists.
21760 (get_repl_default_def_ssa_name): Create the replacement declaration
21763 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21765 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
21766 final_end_function.
21768 2013-02-28 Marek Polacek <polacek@redhat.com>
21770 PR rtl-optimization/56466
21771 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
21772 if we're changing a loop.
21773 (peel_loops_completely): Likewise.
21775 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
21778 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
21780 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21783 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
21784 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
21785 INTX_FTYPE_FX, FX_FTYPE_INTX.
21786 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
21788 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21790 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
21791 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
21792 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
21793 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
21794 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
21795 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
21796 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
21797 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
21798 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
21799 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
21800 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
21801 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
21802 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
21803 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
21804 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
21805 (avrxmega6): Increase max flash segments from 5 to 6.
21806 * config/avr/t-multilib: Regenerate.
21807 * config/avr/avr-tables.opt: Regenerate.
21808 * doc/avr-mmcu.texi: Regenerate.
21810 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21812 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
21813 (avr_device_to_arch): Rename to avr_device_to_ld.
21814 (avr_device_to_as): New prototype.
21815 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
21816 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
21817 * config/avr/driver-avr.c (avr_device_to_as): New.
21818 (avr_device_to_arch): Rename to avr_device_to_ld.
21820 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21822 PR middle-end/56461
21823 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
21824 method on dr_chain and result_chain.
21826 PR middle-end/56461
21827 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
21828 pointer_set_destroy on not_executed_last_iteration.
21830 PR middle-end/56461
21831 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
21833 PR middle-end/56461
21834 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
21835 FOR_EACH_DEFINED_FUNCTION when freeing state.
21837 PR middle-end/56461
21838 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
21840 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
21843 PR middle-end/56461
21844 * ipa-cp.c (decide_whether_version_node): Call vec_free on
21845 known_aggs[i].items and release known_aggs vector.
21847 PR middle-end/56461
21848 * ipa-reference.c (propagate): Free node_info even for alias nodes.
21850 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21852 * config/microblaze/microblaze.c (microblaze_emit_compare):
21853 Use xor for EQ/NE comparisions.
21854 * config/microblaze/microblaze.md (cstoresf4): Add constraints
21855 (cbranchsf4): Adjust operator to comparison_operator.
21857 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21859 PR middle-end/56461
21860 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
21862 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
21863 vec_safe_push, always update *slot.
21864 (redirect_edge_var_map_clear): Use vec_free.
21865 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
21866 (free_var_map_entry): Use vec_free.
21867 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
21868 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
21870 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
21872 PR middle-end/45472
21873 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
21874 when the may_trap_p bit of the exprs being merged differs.
21875 Reorder tests for speculativeness in the logical and operator.
21877 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21879 * incpath.c (add_standard_paths): Use reconcat instead of concat
21880 where appropriate and avoid leaking memory.
21882 * opts.h: Include obstack.h.
21883 (opts_concat): New prototype.
21884 (opts_obstack): New declaration.
21885 * opts.c (opts_concat): New function.
21886 (opts_obstack): New variable.
21887 (init_options_struct): Call gcc_init_obstack on opts_obstack.
21888 (finish_options): Use opts_concat instead of concat
21889 and XOBNEWVEC instead of XNEWVEC.
21890 * opts-common.c (generate_canonical_option, decode_cmdline_option,
21891 generate_option): Likewise.
21892 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
21893 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
21896 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
21897 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
21899 2013-02-26 Jakub Jelinek <jakub@redhat.com>
21901 PR middle-end/56461
21902 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
21904 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
21906 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
21907 (arm_block_move_unaligned_straight): Likewise.
21908 (arm_adjust_block_mem): Likewise.
21910 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
21913 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
21914 temp, cond and label.
21915 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
21918 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
21919 * config/c6x/c6x.h (dbx_register_map): Update declaration.
21922 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
21923 of prologue/epilogue functions.
21926 * config/tilegx/tilegx.c (tilegx_expand_prologue):
21927 Remove unused variable cfa_offset.
21928 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
21931 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
21932 type promotion to unsigned.
21935 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
21936 for HOST_WIDE_INT of 32 bit / same size as int.
21937 (arm_block_move_unaligned_straight): Likewise.
21938 (arm_adjust_block_mem): Likewise.
21941 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
21944 2013-02-26 Marek Polacek <polacek@redhat.com>
21946 PR tree-optimization/56426
21947 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
21949 2013-02-26 Richard Biener <rguenther@suse.de>
21952 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
21953 unused variable loops.
21955 2013-02-26 Jakub Jelinek <jakub@redhat.com>
21957 PR tree-optimization/56448
21958 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
21959 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
21960 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
21961 later operands of the references, or even first operand for
21962 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
21964 PR tree-optimization/56443
21965 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
21966 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
21967 to type_for_mode langhook.
21969 2013-02-25 Matt Turner <mattst88@gmail.com>
21971 * doc/invoke.texi: Document r4700.
21973 2013-02-25 Richard Biener <rguenther@suse.de>
21975 PR tree-optimization/56175
21976 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
21978 (simplify_bitwise_binary): ... here. Also guard the conversion
21979 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
21981 2013-02-25 Catherine Moore <clm@codesourcery.com>
21984 2013-02-24 Catherine Moore <clm@codesourcery.com>
21985 Maciej W. Rozycki <macro@codesourcery.com>
21986 Tom de Vries <tom@codesourcery.com>
21987 Nathan Sidwell <nathan@codesourcery.com>
21988 Iain Sandoe <iain@codesourcery.com>
21989 Nathan Froyd <froydnj@codesourcery.com>
21990 Chao-ying Fu <fu@mips.com>
21992 * doc/extend.texi: (micromips, nomicromips, nocompression):
21993 Document new function attributes.
21994 * doc/invoke.texi (minterlink-compressed, mmicromips,
21995 m14k, m14ke, m14kec): Document new options.
21996 (minterlink-mips16): Update documentation.
21997 * doc/md.texi (ZC, ZD): Document new constraints.
21998 * configure.ac (gcc_cv_as_micromips): Check if linker
21999 supports the .set micromips directive.
22000 * configure: Regenerate.
22001 * config.in: Regenerate.
22002 * config/mips/mips-tables.opt: Regenerate.
22003 * config/mips/micromips.md: New file.
22004 * constraints.md (ZC, AD): New constraints.
22005 * config/mips/predicates.md (movep_src_register): New predicate.
22006 (movep_src_operand): New predicate.
22007 (non_volatile_mem_operand): New predicate.
22008 * config/mips/mips.md (multimem): New type.
22009 (length): Differentiate between 17-bit and 18-bit branch offsets.
22010 (MOVEP1, MOVEP2): New mode iterator.
22011 (mov_<load>l): Use ZC constraint.
22012 (mov_<load>r): Likewise.
22013 (mov_<store>l): Likewise.
22014 (mov_<store>r): Likewise.
22015 (*branch_equality<mode>_inverted): Add microMIPS support.
22016 (*branch_equality<mode>): Likewise.
22017 (*jump_absolute): Likewise.
22018 (indirect_jump_<mode>): Likewise.
22019 (tablejump_<mode>): Likewise.
22020 (<optab>_internal): Likewise.
22021 (sibcall_internal): Likewise.
22022 (sibcall_value_internal): Likewise.
22023 (prefetch): Use constraint ZD.
22024 * config/mips/mips.opt (minterlink-compressed): New option.
22025 (minterlink-mips16): Now an alias for minterlink-compressed.
22026 (mmicromips): New option.
22027 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22028 (compare_and_swap_12): Likewise.
22029 (sync_add<mode>): Likewise.
22030 (sync_<optab>_12): Likewise.
22031 (sync_old_<optab>_12): Likewise.
22032 (sync_new_<optab>_12): Likewise.
22033 (sync_nand_12): Likewise.
22034 (sync_old_nand_12): Likewise.
22035 (sync_new_nand_12): Likewise.
22036 (sync_sub<mode>): Likewise.
22037 (sync_old_add<mode>): Likewise.
22038 (sync_old_sub<mode>): Likewise.
22039 (sync_new_add<mode>): Likewise.
22040 (sync_new_sub<mode>): Likewise.
22041 (sync_<optab><mode>): Likewise.
22042 (sync_old_<optab><mode>): Likewise.
22043 (sync_new_<optab><mode>): Likewise.
22044 (sync_nand<mode>): Likewise.
22045 (sync_old_nand<mode>): Likewise.
22046 (sync_new_nand<mode>): Likewise.
22047 (sync_lock_test_and_set<mode>): Likewise.
22048 (test_and_set_12): Likewise.
22049 (atomic_compare_and_swap<mode>): Likewise.
22050 (atomic_exchange<mode>_llsc): Likewise.
22051 (atomic_fetch_add<mode>_llsc): Likewise.
22052 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22053 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22054 (umips_save_restore_pattern_p): Likewise.
22055 (umips_load_store_pair_p): Likewise.
22056 (umips_output_load_store_pair): Likewise.
22057 (umips_movep_target_p): Likewise.
22058 (umips_12bit_offset_address_p): Likewise.
22059 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22060 (mips_base_mips16): Rename this...
22061 (mips_base_compression_flags): ...to this. Update all uses.
22062 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22063 (mips_mips16_decl_p): Delete.
22064 (mips_nomips16_decl_p): Delete.
22065 (mips_get_compress_on_flags): New function.
22066 (mips_get_compress_off_flags): New function.
22067 (mips_get_compress_mode): New function.
22068 (mips_get_compress_on_name): New function.
22069 (mips_get_compress_off_name): New function.
22070 (mips_insert_attributes): Support multiple compression types.
22071 (mips_merge_decl_attributes): Likewise.
22072 (umips_12bit_offset_address_p): New function.
22073 (mips_start_function_definition): Emit .set micromips directive.
22074 (mips_call_may_need_jalx_p): New function.
22075 (mips_function_ok_for_sibcall): Add microMIPS support.
22076 (mips_print_operand_punctuation): Support short delay slots and
22078 (umips_swm_mask, umips_swm_encoding): New.
22079 (umips_build_save_restore): New function.
22080 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22081 (was_mips16_p): Remove.
22082 (old_compression_mode): New.
22083 (mips_set_compression_mode): New function.
22084 (mips_set_current_function): Add microMIPS support.
22085 (mips_option_override): Likewise.
22086 (umips_save_restore_pattern_p): New function.
22087 (umips_output_save_restore): New function.
22088 (umips_load_store_pair_p_1): New function.
22089 (umips_load_store_pair_p): New function.
22090 (umips_output_load_store_pair_1): New function.
22091 (umips_output_load_store_pair): New function.
22092 (umips_movep_target_p) New function.
22093 (mips_prepare_pch_save): Add microMIPS support.
22094 * config/mips/mips.h (TARGET_COMPRESSION): New.
22095 (TARGET_CPU_CPP_BUILTINS): Update macro
22096 to use new compression flags and to support microMIPS.
22097 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22098 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22099 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22100 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22101 (ASM_SPEC): Support mmicromips and mno-micromips.
22102 (M16STORE_REG_P): New macro.
22103 (MIPS_CALL): Support TARGET_MICROMIPS.
22104 (MICROMIPS_J): New macro.
22105 (mips_base_mips16): Rename this...
22106 (mips_base_compression_flags): ...to this.
22107 (UMIPS_12BIT_OFFSET_P): New macro.
22108 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22109 (MULTILIB_DIRNAMES): Likewise.
22111 2013-02-25 Tom de Vries <tom@codesourcery.com>
22113 PR rtl-optimization/56131
22114 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
22115 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
22116 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
22118 2013-02-25 Tobias Burnus <burnus@net-b.de>
22120 * doc/invoke.texi (-fsanitize=): Move from optimization
22121 to debugging options.
22123 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
22125 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
22127 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
22128 Alexander Monakov <amonakov@ispras.ru>
22130 PR middle-end/56077
22131 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
22132 flush pending lists also on non-jumps. Adjust comment.
22134 2013-02-24 Catherine Moore <clm@codesourcery.com>
22135 Maciej W. Rozycki <macro@codesourcery.com>
22136 Tom de Vries <tom@codesourcery.com>
22137 Nathan Sidwell <nathan@codesourcery.com>
22138 Iain Sandoe <iain@codesourcery.com>
22139 Nathan Froyd <froydnj@codesourcery.com>
22140 Chao-ying Fu <fu@mips.com>
22142 * doc/extend.texi: (micromips, nomicromips, nocompression):
22143 Document new function attributes.
22144 * doc/invoke.texi (minterlink-compressed, mmicromips,
22145 m14k, m14ke, m14kec): Document new options.
22146 (minterlink-mips16): Update documentation.
22147 * doc/md.texi (ZC, ZD): Document new constraints.
22148 * configure.ac (gcc_cv_as_micromips): Check if linker
22149 supports the .set micromips directive.
22150 * configure: Regenerate.
22151 * config.in: Regenerate.
22152 * config/mips/mips-tables.opt: Regenerate.
22153 * config/mips/micromips.md: New file.
22154 * constraints.md (ZC, AD): New constraints.
22155 * config/mips/predicates.md (movep_src_register): New predicate.
22156 (movep_src_operand): New predicate.
22157 (non_volatile_mem_operand): New predicate.
22158 * config/mips/mips.md (multimem): New type.
22159 (length): Differentiate between 17-bit and 18-bit branch offsets.
22160 (MOVEP1, MOVEP2): New mode iterator.
22161 (mov_<load>l): Use ZC constraint.
22162 (mov_<load>r): Likewise.
22163 (mov_<store>l): Likewise.
22164 (mov_<store>r): Likewise.
22165 (*branch_equality<mode>_inverted): Add microMIPS support.
22166 (*branch_equality<mode>): Likewise.
22167 (*jump_absolute): Likewise.
22168 (indirect_jump_<mode>): Likewise.
22169 (tablejump_<mode>): Likewise.
22170 (<optab>_internal): Likewise.
22171 (sibcall_internal): Likewise.
22172 (sibcall_value_internal): Likewise.
22173 (prefetch): Use constraint ZD.
22174 * config/mips/mips.opt (minterlink-compressed): New option.
22175 (minterlink-mips16): Now an alias for minterlink-compressed.
22176 (mmicromips): New option.
22177 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22178 (compare_and_swap_12): Likewise.
22179 (sync_add<mode>): Likewise.
22180 (sync_<optab>_12): Likewise.
22181 (sync_old_<optab>_12): Likewise.
22182 (sync_new_<optab>_12): Likewise.
22183 (sync_nand_12): Likewise.
22184 (sync_old_nand_12): Likewise.
22185 (sync_new_nand_12): Likewise.
22186 (sync_sub<mode>): Likewise.
22187 (sync_old_add<mode>): Likewise.
22188 (sync_old_sub<mode>): Likewise.
22189 (sync_new_add<mode>): Likewise.
22190 (sync_new_sub<mode>): Likewise.
22191 (sync_<optab><mode>): Likewise.
22192 (sync_old_<optab><mode>): Likewise.
22193 (sync_new_<optab><mode>): Likewise.
22194 (sync_nand<mode>): Likewise.
22195 (sync_old_nand<mode>): Likewise.
22196 (sync_new_nand<mode>): Likewise.
22197 (sync_lock_test_and_set<mode>): Likewise.
22198 (test_and_set_12): Likewise.
22199 (atomic_compare_and_swap<mode>): Likewise.
22200 (atomic_exchange<mode>_llsc): Likewise.
22201 (atomic_fetch_add<mode>_llsc): Likewise.
22202 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22203 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22204 (umips_save_restore_pattern_p): Likewise.
22205 (umips_load_store_pair_p): Likewise.
22206 (umips_output_load_store_pair): Likewise.
22207 (umips_movep_target_p): Likewise.
22208 (umips_12bit_offset_address_p): Likewise.
22209 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22210 (mips_base_mips16): Rename this...
22211 (mips_base_compression_flags): ...to this. Update all uses.
22212 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22213 (mips_mips16_decl_p): Delete.
22214 (mips_nomips16_decl_p): Delete.
22215 (mips_get_compress_on_flags): New function.
22216 (mips_get_compress_off_flags): New function.
22217 (mips_get_compress_mode): New function.
22218 (mips_get_compress_on_name): New function.
22219 (mips_get_compress_off_name): New function.
22220 (mips_insert_attributes): Support multiple compression types.
22221 (mips_merge_decl_attributes): Likewise.
22222 (umips_12bit_offset_address_p): New function.
22223 (mips_start_function_definition): Emit .set micromips directive.
22224 (mips_call_may_need_jalx_p): New function.
22225 (mips_function_ok_for_sibcall): Add microMIPS support.
22226 (mips_print_operand_punctuation): Support short delay slots and
22228 (umips_swm_mask, umips_swm_encoding): New.
22229 (umips_build_save_restore): New function.
22230 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22231 (was_mips16_p): Remove.
22232 (old_compression_mode): New.
22233 (mips_set_compression_mode): New function.
22234 (mips_set_current_function): Add microMIPS support.
22235 (mips_option_override): Likewise.
22236 (umips_save_restore_pattern_p): New function.
22237 (umips_output_save_restore): New function.
22238 (umips_load_store_pair_p_1): New function.
22239 (umips_load_store_pair_p): New function.
22240 (umips_output_load_store_pair_1): New function.
22241 (umips_output_load_store_pair): New function.
22242 (umips_movep_target_p) New function.
22243 (mips_prepare_pch_save): Add microMIPS support.
22244 * config/mips/mips.h (TARGET_COMPRESSION): New.
22245 (TARGET_CPU_CPP_BUILTINS): Update macro
22246 to use new compression flags and to support microMIPS.
22247 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22248 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22249 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22250 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22251 (ASM_SPEC): Support mmicromips and mno-micromips.
22252 (M16STORE_REG_P): New macro.
22253 (MIPS_CALL): Support TARGET_MICROMIPS.
22254 (MICROMIPS_J): New macro.
22255 (mips_base_mips16): Rename this...
22256 (mips_base_compression_flags): ...to this.
22257 (UMIPS_12BIT_OFFSET_P): New macro.
22258 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22259 (MULTILIB_DIRNAMES): Likewise.
22261 2013-02-24 Jakub Jelinek <jakub@redhat.com>
22264 * target-globals.c (save_target_globals): For init_reg_sets and
22265 target_reinit remporarily set this_fn_optabs to this_target_optabs.
22267 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
22269 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
22270 * config/aarch64/t-aarch64
22271 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
22273 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
22275 PR inline-asm/56148
22276 * lra-constraints.c (process_alt_operands): Reload operand
22277 conflicting with earlier clobber only if no more other conflicting
22280 2013-02-22 Jakub Jelinek <jakub@redhat.com>
22283 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
22284 if not linking a shared library.
22286 2013-02-22 Seth LaForge <sethml@google.com>
22288 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
22290 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
22292 * config/arm/arm.md (split for extendsidi): Update condition.
22293 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
22294 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
22295 (qhs_zextenddi_cstr): Likewise.
22297 2013-02-21 Jakub Jelinek <jakub@redhat.com>
22299 PR middle-end/56420
22300 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
22301 avoid signed wrapping.
22302 (expand_mult): Handle properly multiplication by
22303 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
22304 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
22305 in the compiler if coeff is HOST_WIDE_INT_MIN.
22306 (expand_divmod): Don't make ext_op1 static, change it's type to
22307 uhwi. Avoid undefined behavior in -INTVAL (op1).
22309 PR rtl-optimization/50339
22310 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
22312 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
22313 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
22314 into splitting_ashiftrt field.
22315 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
22317 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
22320 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
22322 PR middle-end/56108
22323 * trans-mem.c (execute_tm_mark): Do not expand transactions that
22324 are sure to go irrevocable.
22326 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
22328 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
22329 scalars are valid operands.
22331 2013-02-21 Martin Jambor <mjambor@suse.cz>
22333 PR tree-optimization/56310
22334 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
22335 only matching indices and non-negative final offsets.
22336 (intersect_aggregates_with_edge): Pass src_idx to
22337 agg_replacements_to_vector. Pass src_idx insstead of index to
22338 intersect_with_agg_replacements.
22340 2013-02-21 Martin Jambor <mjambor@suse.cz>
22342 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
22343 instead of hard-wired defaults.
22345 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
22347 * doc/invoke.texi (MIPS Options): Update documentation of the
22348 floating-point multiply-accumulate instruction restrictions.
22350 2013-02-21 Kostya Serebryany <kcc@google.com>
22352 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
22353 asan_shadow_offset on x86_64 linux.
22355 2013-02-21 Richard Biener <rguenther@suse.de>
22357 PR tree-optimization/56415
22359 2013-02-11 Richard Biener <rguenther@suse.de>
22361 PR tree-optimization/56273
22362 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
22365 2013-02-21 Jakub Jelinek <jakub@redhat.com>
22368 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
22371 PR inline-asm/56405
22372 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
22373 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
22375 2013-02-20 Jan Hubicka <jh@suse.cz>
22377 PR tree-optimization/56265
22378 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
22379 when target is referenced for first time.
22381 2013-02-20 Richard Biener <rguenther@suse.de>
22383 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
22384 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
22385 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
22386 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
22387 not return anything.
22388 (rename_ssa_copies): Do not remove unused locals.
22389 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
22390 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
22391 * passes.c (execute_function_todo): Do not schedule unused locals
22392 removal if cleanup_tree_cfg did something.
22393 * tree-ssa-live.c (remove_unused_locals): Dump statistics
22394 about the number of removed locals.
22396 2013-02-20 Richard Biener <rguenther@suse.de>
22398 PR tree-optimization/56398
22399 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
22401 2013-02-20 Martin Jambor <mjambor@suse.cz>
22403 PR tree-optimization/55334
22404 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
22405 restricted pointers to arrays.
22407 2013-02-20 Richard Biener <rguenther@suse.de>
22408 Jakub Jelinek <jakub@redhat.com>
22410 PR tree-optimization/56396
22411 * tree-ssa-ccp.c (n_const_val): New static variable.
22412 (get_value): Return NULL for SSA names we don't have a lattice
22414 (ccp_initialize): Initialize n_const_val.
22415 * tree-ssa-copy.c (n_copy_of): New static variable.
22416 (init_copy_prop): Initialize n_copy_of.
22417 (get_value): Return NULL_TREE for SSA names we don't have a
22420 2013-02-20 Martin Jambor <mjambor@suse.cz>
22422 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
22424 2013-02-20 Richard Biener <rguenther@suse.de>
22426 * genpreds.c (write_lookup_constraint): Do not compare first
22427 letter of the constraint again.
22429 2013-02-20 Richard Biener <rguenther@suse.de>
22431 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
22433 (get_use_iv_cost): Terminate hashtable walk when coming across
22436 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
22438 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
22439 reassociation for avx2 targets.
22441 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22443 * config/microblaze/microblaze.c: microblaze_has_clz = 0
22444 Add version check for v8.10.a to enable microblaze_has_clz
22445 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
22446 version and TARGET_PATTERN_COMPARE check
22447 * config/microblaze/microblaze.md: New clzsi2 instruction
22449 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22451 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
22452 function before branching.
22454 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22456 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
22458 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
22460 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22462 PR middle-end/55889
22463 * sel-sched.c: Include ira.h.
22464 (implicit_clobber_conflict_p): New function.
22465 (moveup_expr): Use it.
22466 * Makefile.in (sel-sched.o): Depend on ira.h.
22468 2013-02-19 Richard Biener <rguenther@suse.de>
22470 PR tree-optimization/56384
22471 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
22472 (vn_hash_type): Split out from ...
22473 (vn_hash_constant_with_type): ... here.
22474 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
22475 (vn_phi_eq): Compare types from vn_phi_s structure.
22476 (vn_phi_lookup): Populate vn_phi_s type.
22477 (vn_phi_insert): Likewise.
22479 2013-02-19 Jakub Jelinek <jakub@redhat.com>
22481 PR tree-optimization/56350
22482 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
22483 if haven't found reduction or nested cycle operand, rather than
22484 asserting we must find it.
22486 PR tree-optimization/56381
22487 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
22490 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
22491 Jakub Jelinek <jakub@redhat.com>
22494 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
22495 (swap_optab_enable): Same.
22496 (init_all_optabs): Use argument instead of global.
22497 * tree.h (struct tree_optimization_option): New field target_optabs.
22498 * expr.h (init_all_optabs): Add argument to prototype.
22499 (TREE_OPTIMIZATION_OPTABS): New.
22500 (save_optabs_if_changed): Protoize.
22501 * optabs.h: Declare this_fn_optabs.
22502 * optabs.c (save_optabs_if_changed): New.
22503 Declare this_fn_optabs.
22504 (init_optabs): Add argument to init_all_optabs() call.
22505 * function.c (invoke_set_current_function_hook): Handle per
22507 * function.h (struct function): New field optabs.
22508 * config/mips/mips.c (mips_set_mips16_mode): Handle when
22509 optimization_current_node has changed.
22510 * target-globals.h (save_target_globals_default_opts): Protoize.
22511 * target-globals.c (save_target_globals_default_opts): New.
22513 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22516 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
22517 registers %fr12 and %fr12R as call used.
22520 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
22521 and HImode, require all displacements to be an integer multiple of
22523 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
22524 only allow QImode and HImode when reload is in progress and strict is
22525 true. Likewise for symbolic addresses. Use base14_operand to check
22526 displacements in REG+BASE addresses.
22528 2013-02-18 Richard Biener <rguenther@suse.de>
22530 PR tree-optimization/56366
22531 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22532 handle sign-conversion of outer-loop initial induction value.
22534 2013-02-18 Richard Biener <rguenther@suse.de>
22536 PR middle-end/56349
22537 * cfghooks.c (merge_blocks): If we merge a latch into another
22538 block adjust references to it.
22539 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
22540 (verify_loop_structure): Verify that a recorded latch is in fact
22543 2013-02-18 Richard Biener <rguenther@suse.de>
22545 PR tree-optimization/56321
22546 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
22547 order SSA name release and virtual operand unlinking.
22549 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22551 * config/microblaze/microblaze.md (save_stack_block): Define.
22552 (restore_stack_block): Likewise.
22554 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22556 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
22557 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
22558 * config/microblaze/microblaze.c (microblaze_option_override):
22559 Bail out early for PIC modes when target does not support PIC.
22561 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22563 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
22564 Replace with a microblaze version.
22565 (microblaze_trampoline_init): Adapt for microblaze.
22566 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
22569 2013-02-16 Jakub Jelinek <jakub@redhat.com>
22570 Dodji Seketeli <dodji@redhat.com>
22573 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
22574 (instrument_mem_region_access): Do not forget to always put
22575 instrumentation of the of 'base' and 'base + len' in a "if (len !=
22576 0) statement, even for cases where either 'base' or 'base + len'
22577 are not instrumented -- because they have been previously
22578 instrumented. Simplify the logic by putting all the statements
22579 instrument 'base + len' inside a sequence, and then insert that
22580 sequence right before the current insertion point. Then, to
22581 instrument 'base + len', just get an iterator on that statement.
22582 And do not forget to update the pointer to iterator the function
22583 received as argument.
22585 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
22587 PR rtl-optimization/56348
22588 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
22590 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
22592 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
22593 (clean_graph_dump_file): Pass base to start_graph_dump.
22595 2013-02-14 Richard Henderson <rth@redhat.com>
22598 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
22600 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
22602 * collect2-aix.h: Define F_LOADONLY.
22604 2013-02-14 Richard Biener <rguenther@suse.de>
22607 * varasm.c (output_constant_def_1): Get the decl representing
22608 the constant as argument.
22609 (output_constant_def): Wrap output_constant_def_1.
22610 (make_decl_rtl): Use output_constant_def_1 with the decl
22611 representing the constant.
22612 (build_constant_desc): Optionally re-use a decl already
22613 representing the constant.
22614 (tree_output_constant_def): Adjust.
22616 2013-02-14 Dodji Seketeli <dodji@redhat.com>
22619 * asan.c (instrument_builtin_call): Really put the length of the
22620 second source argument into src1_len.
22622 2013-02-13 Jakub Jelinek <jakub@redhat.com>
22624 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
22625 argument. If it is false, don't create edge from then_bb to
22627 (insert_if_then_before_iter): Pass true to it.
22628 (build_check_stmt): Pass false to it.
22629 (transform_statements): Flush hash table only on extended basic
22630 block boundaries, rather than at the beginning of every bb.
22631 Don't flush hash table on nonfreeing_call_p calls.
22632 * tree-flow.h (nonfreeing_call_p): New prototype.
22633 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
22635 2013-02-13 David S. Miller <davem@davemloft.net>
22637 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
22639 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
22642 * ira.c (max_regno_before_ira): Move from ...
22644 (fix_reg_equiv_init): Use max_regno_before_ira instead of
22647 2013-02-13 Jakub Jelinek <jakub@redhat.com>
22649 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
22651 2013-02-13 Richard Biener <rguenther@suse.de>
22654 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
22655 globals in MEM_REFs.
22657 2013-02-13 Richard Biener <rguenther@suse.de>
22659 * loop-init.c (loop_optimizer_init): Clear loop state when
22660 re-initializing preserved loops.
22661 * loop-unswitch.c (unswitch_single_loop): Return whether
22662 we unswitched the loop. Do not verify loop state here.
22663 (unswitch_loops): When we unswitched a loop discover new loops.
22665 2013-02-13 Kostya Serebryany <kcc@google.com>
22667 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
22669 * sanitizer.def: Rename __asan_init to __asan_init_v1.
22671 2013-02-12 Dodji Seketeli <dodji@redhat.com>
22673 Avoid instrumenting duplicated memory access in the same basic block
22674 * Makefile.in (asan.o): Add new dependency on hash-table.h
22675 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
22676 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
22677 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
22678 (free_mem_ref_resources, has_mem_ref_been_instrumented)
22679 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
22680 (get_mem_ref_of_assignment): New functions.
22681 (get_mem_refs_of_builtin_call): Extract from
22682 instrument_builtin_call and tweak a little bit to make it fit with
22684 (instrument_builtin_call): Use the new
22685 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
22686 of is_gimple_builtin_call.
22687 (instrument_derefs, instrument_mem_region_access): Insert the
22688 instrumented memory reference into the hash table.
22689 (maybe_instrument_assignment): Renamed instrument_assignment into
22690 this, and change it to advance the iterator when instrumentation
22691 actually happened and return true in that case. This makes it
22692 homogeneous with maybe_instrument_assignment, and thus give a
22693 chance to callers to be more 'regular'.
22694 (transform_statements): Clear the memory reference hash table
22695 whenever we enter a new BB, when we cross a function call, or when
22696 we are done transforming statements. Use
22697 maybe_instrument_assignment instead of instrumentation. No more
22698 need to special case maybe_instrument_assignment and advance the
22699 iterator after calling it; it's now handled just like
22700 maybe_instrument_call. Update comment.
22702 2013-02-13 Richard Biener <rguenther@suse.de>
22704 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
22705 Fix loop discovery code.
22707 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
22709 PR inline-asm/56148
22710 * lra-constraints.c (process_alt_operands): Match early clobber
22711 operand with itself. Check conflicts with earlyclobber only if
22712 the operand is not reloaded. Prefer to reload conflicting operand
22713 if earlyclobber and matching operands are the same.
22715 2013-02-12 Richard Biener <rguenther@suse.de>
22718 * lto-streamer-out.c (write_symbol): Do not output symbols
22719 for hard register variables.
22721 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
22724 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
22725 (umulsidi3_insn, mulsidi3_insn): New insns.
22727 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
22729 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
22730 (struct tune_params): Add vec_costs field.
22731 * config/arm/arm.c (arm_builtin_vectorization_cost)
22732 (arm_add_stmt_cost): New functions.
22733 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
22734 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
22735 (arm_default_vec_cost): New struct of type cpu_vec_costs.
22736 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
22737 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
22738 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
22739 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
22741 2013-02-12 Richard Biener <rguenther@suse.de>
22744 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
22745 decls again if possible.
22747 2013-02-12 Richard Biener <rguenther@suse.de>
22749 PR middle-end/56288
22750 * tree-ssa.c (verify_ssa_name): Fix check, move
22751 SSA_NAME_IN_FREE_LIST check up.
22753 2013-02-12 Jakub Jelinek <jakub@redhat.com>
22754 Steven Bosscher <steven@gcc.gnu.org>
22756 PR rtl-optimization/56151
22757 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
22758 equal to op0 or op1, and last_insn pattern is CODE operation
22759 with MEM dest and one of the operands matches that MEM.
22761 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
22763 * doc/extend.texi: Document Function Multiversioning and "default"
22764 parameter string to target attribute.
22765 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
22766 target attribute parameter is "default".
22767 (ix86_compare_version_priority): Remove checks for target attribute.
22768 (ix86_mangle_function_version_assembler_name): Change error to sorry.
22769 Remove check for target attribute equal to NULL. Add assert.
22770 (ix86_generate_version_dispatcher_body): Change error to sorry.
22772 2013-02-11 Iain Sandoe <iain@codesourcery.com>
22773 Jack Howarth <howarth@bromo.med.uc.edu>
22774 Patrick Marlier <patrick.marlier@gmail.com>
22777 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
22778 define ENDFILE_SPEC as TM_DESTRUCTOR.
22779 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
22781 2013-02-11 Alexander Potapenko <glider@google.com>
22782 Jack Howarth <howarth@bromo.med.uc.edu>
22783 Jakub Jelinek <jakub@redhat.com>
22786 * config/darwin.c (cdtor_record): Rename ctor_record.
22787 (sort_cdtor_records): Rename sort_ctor_records.
22788 (finalize_dtors): New routine to sort destructors by
22789 priority before use in assemble_integer.
22790 (machopic_asm_out_destructor): Use finalize_dtors if needed.
22792 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
22794 PR rtl-optimization/56275
22795 * simplify-rtx.c (avoid_constant_pool_reference): Check that
22796 offset is non-negative and less than cmode size before
22797 calling simplify_subreg.
22799 2013-02-11 Richard Biener <rguenther@suse.de>
22801 PR tree-optimization/56264
22802 * cfgloop.h (fix_loop_structure): Adjust prototype.
22803 * loop-init.c (fix_loop_structure): Return the number of
22804 newly discovered loops.
22805 * tree-cfgcleanup.c (repair_loop_structures): When new loops
22806 are discovered, do a full loop-closed SSA rewrite.
22808 2013-02-11 Richard Biener <rguenther@suse.de>
22810 PR tree-optimization/56273
22811 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
22813 (check_array_ref): Fix missing newline in dumps.
22814 (search_for_addr_array): Likewise.
22816 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
22818 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
22820 2013-02-09 Jakub Jelinek <jakub@redhat.com>
22823 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
22825 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
22827 PR rtl-optimization/56246
22828 * lra-constraints.c (simplify_operand_subreg): Try to reuse
22830 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
22831 constraints are satisfied.
22833 2013-02-08 Jeff Law <law@redhat.com>
22836 * emit-rtl.c (reg_is_parm_p): New function.
22837 * regs.h (reg_is_parm_p): New prototype.
22838 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
22839 callee-clobbered registers.
22841 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
22844 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
22845 If there is no implicit builtin declaration, just return NULL.
22847 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22849 * config/i386/sse.md (FMAMODEM): New mode iterator.
22850 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
22851 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
22853 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22855 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
22856 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
22857 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
22859 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22861 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
22862 (microblaze*-*-elf): Likewise.
22863 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
22865 * config/microblaze/microblaze-c.c: Add builtin defines for
22866 _LITTLE_ENDIAN and _BIG_ENDIAN.
22867 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
22868 add to TARGET_DEFAULT flags.
22869 Expand ASM_SPEC and LINK_SPEC.
22870 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
22871 * config/microblaze/microblaze.md: Update extendsidi2 and
22872 movdi_internal instructions to use low-order / high-order reg
22874 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
22875 options and inversemask / mask of LITTLE_ENDIAN.
22876 * config/microblaze/t-microblaze: Expand multilib options to
22877 include mlittle-endian (le) and update exceptions patterns.
22879 2013-02-08 Jakub Jelinek <jakub@redhat.com>
22881 PR rtl-optimization/56195
22882 * lra-constraints.c (get_reload_reg): Don't reuse regs
22883 if they have smaller mode than requested, if they have
22884 wider mode than requested, try to return a SUBREG.
22886 PR tree-optimization/56250
22887 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
22888 if type is unsigned and code isn't MULT_EXPR.
22890 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
22892 PR tree-optimization/56064
22893 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
22894 bits according to mode.
22895 * fixed-value.h (fixed_from_double_int)
22896 (const_fixed_from_double_int): Adjust comments.
22898 2013-02-08 Richard Biener <rguenther@suse.de>
22901 * lto-streamer.h (struct data_in): Remove current_file, current_line
22902 and current_col members.
22903 * lto-streamer-out.c (lto_output_location): Stream changed bits
22904 en-block for efficiency.
22905 * lto-streamer-in.c (clear_line_info): Remove.
22906 (lto_input_location): Cache current file, line and column
22907 globally via local statics. Read changed bits en-block.
22908 (input_function): Do not call clear_line_info.
22909 (lto_read_body): Likewise.
22910 (lto_input_toplevel_asms): Likewise.
22912 2013-02-08 Michael Matz <matz@suse.de>
22914 PR tree-optimization/52448
22915 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
22916 (nt_call_phase): New static.
22917 (add_or_mark_expr): Only mark accesses with newer phase than any
22919 (nonfreeing_call_p): New.
22920 (nt_init_block): Update nt_call_phase, mark blocks as visited.
22921 (nt_fini_block): Keep blocks marked as visited.
22922 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
22924 2013-02-08 Richard Biener <rguenther@suse.de>
22926 * ira.c (ira): Free broken dominator information.
22928 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22930 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
22932 2013-02-08 Marek Polacek <polacek@redhat.com>
22934 * cfgloop.c (verify_loop_structure): Add more checking of headers.
22936 2013-02-08 Richard Biener <rguenther@suse.de>
22938 PR middle-end/56181
22939 * cfgloop.h (flow_loops_find): Adjust.
22940 (bb_loop_header_p): Declare.
22941 * cfgloop.c (bb_loop_header_p): New function split out from ...
22942 (flow_loops_find): ... here. Adjust function signature,
22943 support incremental loop structure update.
22944 (verify_loop_structure): Cleanup. Verify a loop is a loop.
22945 * cfgloopmanip.c (fix_loop_structure): Move ...
22946 * loop-init.c (fix_loop_structure): ... here.
22947 (apply_loop_flags): Split out from ...
22948 (loop_optimizer_init): ... here.
22949 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
22950 in incremental mode, only remove dead loops here.
22952 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
22955 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
22956 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
22957 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
22958 (*round<mode>3.libgcc): New insns for fixed-modes.
22959 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
22960 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
22961 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
22962 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
22963 implementations. Define to __builtin_avr_absFX,
22964 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
22965 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
22966 __builtin_avr_countlsFX, respectively.
22967 * config/avr/avr-c.c (target.h): Include it.
22968 (enum avr_builtin_id): New enum.
22969 (avr_resolve_overloaded_builtin): New static function.
22970 (avr_register_target_pragmas): Use it to set
22971 targetm.resolve_overloaded_builtin.
22972 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
22973 tree nodes used by DEF_BUILTIN.
22974 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
22975 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
22976 <AVR_BUILTIN_xxBITS>: Same.
22978 2013-02-08 Richard Biener <rguenther@suse.de>
22980 * cfgloop.c (verify_loop_structure): Properly handle
22981 a loop exiting to another loop header.
22982 * ira-int.h (ira_loops): Remove.
22983 * ira.c (ira_loops): Remove.
22984 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
22985 (do_reload): Use loop_optimizer_finalize.
22986 * ira-build.c (create_loop_tree_nodes): Use get_loops and
22987 number_of_loops to access the loop tree.
22988 (more_one_region_p): Likewise.
22989 (finish_loop_tree_nodes): Likewise.
22990 (rebuild_regno_allocno_maps): Likewise.
22991 (mark_loops_for_removal): Likewise.
22992 (mark_all_loops_for_removal): Likewise.
22993 (remove_unnecessary_regions): Likewise.
22994 (ira_build): Likewise.
22995 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22997 2013-02-08 Richard Biener <rguenther@suse.de>
22999 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
23000 * ipa-pure-const.c (analyze_function): Avoid calling
23001 mark_irreducible_loops twice.
23002 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
23004 2013-02-07 David S. Miller <davem@davemloft.net>
23006 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
23008 * var-tracking.c (vt_add_function_parameter): Test the presence of
23009 HAVE_window_save properly and do not remap argument registers when
23010 we have a leaf function.
23012 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
23015 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
23017 * config/i386/i386.c (ix86_print_operand): Ditto.
23019 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
23021 * lra-constraints.c (process_alt_operands): Fix recently added comment.
23023 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
23025 PR rtl-optimization/56225
23026 * lra-constraints.c (process_alt_operands): Check that reload hard
23027 reg can hold value for strict_low_part.
23029 2013-02-07 Jakub Jelinek <jakub@redhat.com>
23032 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
23033 dwarf2out_end_function.
23034 (in_first_function_p, maybe_at_text_label_p,
23035 first_loclabel_num_not_at_text_label): New variables.
23036 (dwarf2out_var_location): In the first function find out
23037 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
23038 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
23041 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
23043 PR rtl-optimization/56178
23044 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
23045 SUBREG of a register. Tidy up related block of code.
23046 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
23047 note if the source is a register or a SUBREG of a register.
23049 2013-02-07 Jakub Jelinek <jakub@redhat.com>
23052 * config/rs6000/rs6000.md (ptrm): New mode attr.
23053 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
23054 call_value_indirect_aix<pttrsize>,
23055 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
23058 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
23060 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
23061 if -bnortl. Convert to strcmp and strncmp.
23063 2013-02-07 Alan Modra <amodra@gmail.com>
23066 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
23067 addresses won't wrap when offsetting.
23068 (rs6000_secondary_reload): Provide secondary reloads needed for
23069 wrapping LO_SUM addresses.
23071 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
23073 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
23074 MACH, just __MACH__.
23076 2013-02-06 Richard Biener <rguenther@suse.de>
23078 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
23079 instead of calling fix_loop_structure.
23081 2013-02-06 Jakub Jelinek <jakub@redhat.com>
23083 PR middle-end/56217
23084 * omp-low.c (use_pointer_for_field): Return false if
23085 lower_send_shared_vars doesn't generate any copy-out code.
23087 2013-02-06 Tom de Vries <tom@codesourcery.com>
23089 PR rtl-optimization/56131
23090 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
23091 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
23092 of the label is NULL. Add comment.
23094 2013-02-05 Jakub Jelinek <jakub@redhat.com>
23096 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
23099 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
23100 (STATIC_LIBTSAN_LIBS): Likewise.
23101 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
23102 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
23103 is defined, don't add anything else beyond that.
23104 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
23105 (LINK_COMMAND_SPEC): Use them.
23107 PR tree-optimization/56205
23108 * tree-stdarg.c (check_all_va_list_escapes): Return true if
23109 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
23110 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
23112 2013-02-05 Richard Biener <rguenther@suse.de>
23114 PR tree-optimization/53342
23115 PR tree-optimization/53185
23116 * tree-vectorizer.h (vect_check_strided_load): Remove.
23117 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
23118 not disallow peeling for vectorized strided loads.
23119 (vect_check_strided_load): Make static and simplify.
23120 (vect_analyze_data_refs): Adjust.
23121 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
23122 correctly when vectorizing strided loads.
23124 2013-02-05 Richard Biener <rguenther@suse.de>
23126 * doc/install.texi: Refer to ISL, not PPL.
23128 2013-02-05 Jan Hubicka <jh@suse.cz>
23130 PR tree-optimization/55789
23131 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
23133 2013-02-05 Jan Hubicka <jh@suse.cz>
23135 PR tree-optimization/55789
23136 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
23137 the dead call anyway.
23139 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
23142 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
23144 2013-02-04 Alexander Potapenko <glider@google.com>
23145 Jack Howarth <howarth@bromo.med.uc.edu>
23146 Jakub Jelinek <jakub@redhat.com>
23149 * config/darwin.c (sort_ctor_records): Stabilized qsort
23150 on constructor priority by using original position.
23151 (finalize_ctors): New routine to sort constructors by
23152 priority before use in assemble_integer.
23153 (machopic_asm_out_constructor): Use finalize_ctors if needed.
23155 2013-02-04 Jakub Jelinek <jakub@redhat.com>
23158 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
23159 about visibility on artificial decls.
23160 * config/sol2.c (solaris_assemble_visibility): Likewise.
23162 2013-02-04 Kai Tietz <ktietz@redhat.com>
23165 * config/i386/i386.c (function_value_ms_64): Add additional valtype
23166 argument and improve checking of return-argument types for 16-byte
23168 (ix86_function_value_1): Add additional valtype argument on call
23169 of function_value_64.
23170 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
23171 handling infunction_value_64 function.
23173 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
23175 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
23177 2013-02-04 Richard Biener <rguenther@suse.de>
23179 PR tree-optimization/56188
23180 * tree-ssa-structalias.c (label_visit): Consider case with
23181 initially non-empty points-to set.
23182 (perform_var_substitution): Dump node mapping and clean up.
23184 2013-02-04 Richard Guenther <rguenther@suse.de>
23187 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
23188 node prevail as last resort.
23189 (lto_symtab_merge_decls): Remove guard on LTRANS here.
23190 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
23192 2013-02-04 Richard Biener <rguenther@suse.de>
23194 PR tree-optimization/56113
23195 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
23197 (equiv_class_lookup_or_add): ... this.
23198 (label_visit): Adjust and fix error in previous patch.
23199 (perform_var_substitution): Adjust.
23201 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
23203 * config/sh/divtab.c: Fix formatting and comments throughout the file.
23204 * config/sh/sh4-300.md: Likewise.
23205 * config/sh/sh4a.md: Likewise.
23206 * config/sh/constraints.md: Likewise.
23207 * config/sh/sh.md: Likewise.
23208 * config/sh/netbsd-elf.h: Likewise.
23209 * config/sh/predicates.md: Likewise.
23210 * config/sh/sh-protos.h: Likewise.
23211 * config/sh/ushmedia.h: Likewise.
23212 * config/sh/linux.h: Likewise.
23213 * config/sh/sh.c: Likewise.
23214 * config/sh/superh.h: Likewise.
23215 * config/sh/elf.h: Likewise.
23216 * config/sh/sh4.md: Likewise.
23217 * config/sh/sh.h: Likewise.
23219 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23221 * config/pa/constraints.md: Adjust unused letters. Change "T"
23222 constraint to match_test floating_point_store_memory_operand().
23223 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
23224 (base14_operand): New.
23225 (floating_point_store_memory_operand): New.
23226 (integer_store_memory_operand): Revise to use base14_operand and
23227 reg_plus_base_memory_operand.
23228 (move_dest_operand): Allow symbolic_memory_operands.
23229 (symbolic_memory_operand): Check for LO_SOM.
23230 (symbolic_operand): Change default case to break.
23231 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
23232 CONST_DOUBLE values to be reloaded by putting them into memory when
23233 the destination is a floating point register.
23234 (movdf): Remove code to handle CONST_DOUBLE.
23236 (reload_indf_r1): New.
23237 (reload_insf_r1): New.
23238 Consistently use "Q" and "T" constraints with integer and floating
23239 point move instructions, respectively.
23240 (movdi): Remove FAIL.
23241 Change predicate for source operand unamed DImode move from
23242 general_operand to move_src_operand.
23243 (umulsidi3): Change predicate for destination operand to
23245 Likewise for similar unamed patterns.
23246 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
23247 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
23248 (hppa_legitimize_address): Simplify mask calculation.
23249 (pa_emit_move_sequence): Revised handling of secondary reloads from
23250 REG+D addresses for floating point loads and stores. Directly handle
23251 loading CONST0_RTX (mode) to a floating point register.
23252 (pa_secondary_reload): Handle reloading DF and SFmode constant values
23253 to floating point registers. Don't restrict secondary reloads to
23254 floating point registers to integer modes. Revise some comments and
23256 (TARGET_LEGITIMATE_ADDRESS_P): Define.
23257 (pa_legitimate_address_p): New.
23258 (pa_legitimize_reload_address): New.
23259 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
23260 (STRICT_REG_OK_FOR_BASE_P): New.
23261 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
23262 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
23264 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
23265 Andrew Dixie <andrewd@gentrack.com>
23267 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
23270 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
23272 * expmed.c (extract_bit_field_1): Pass the full width of the
23273 structure to get_best_reg_extraction_insn.
23275 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
23278 * configure.ac (use_cxa_atexit): Add AIX.
23279 * configure: Regenerate.
23281 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
23283 2013-02-01 Jakub Jelinek <jakub@redhat.com>
23286 * final.c (need_profile_function): New variable.
23287 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
23288 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
23289 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
23290 notes, targetm.asm_out.function_prologue doesn't emit anything,
23291 HAVE_prologue and profiler should be emitted before prologue,
23292 set need_profile_function instead of emitting it.
23293 (final_scan_insn): If need_profile_function, emit
23294 profile_function on the first NOTE_INSN_BASIC_BLOCK or
23295 NOTE_INSN_FUNCTION_BEG note.
23297 2013-02-01 Richard Henderson <rth@redhat.com>
23299 * config/rs6000/rs6000.md (smulditi3): New.
23302 * config/alpha/alpha.md (umulditi3): New.
23304 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
23306 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
23307 (ASM_OUTPUT_ALIGNED_LOCAL): New.
23309 2013-02-01 Richard Biener <rguenther@suse.de>
23311 PR tree-optimization/56113
23312 * tree-ssa-structalias.c (label_visit): Reduce work for
23313 single-predecessor nodes.
23315 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
23317 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
23318 range isn't testing for zero.
23320 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
23322 PR middle-end/56113
23323 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
23325 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
23326 Nick Clifton <nickc@redhat.com>
23328 * config/v850/constraints.md (Q): Define as a memory constraint.
23329 * config/v850/predicates.md (label_ref_operand): New predicate.
23330 (e3v5_shift_operand): New predicate.
23331 (ior_operator): New predicate.
23332 * config/v850/t-v850: Add e3v5 multilib.
23333 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
23334 (v850_gen_movdi): Prototype.
23335 * config/v850/v850.c: Add support for e3v5 architecture.
23336 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
23338 (construct_save_jarl): Add e3v5 long JARL support.
23339 (v850_adjust_insn_length): New function. Adjust length of call
23340 insns when using e3v5 instructions.
23341 (v850_gen_movdi): New function: Generate instructions to move a
23343 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
23344 (CPP_SPEC): Define __v850e3v5__ as appropriate.
23345 (TARGET_USE_FPU): Enable for e3v5.
23346 (CONST_OK_FOR_W): New macro.
23347 (ADJUST_INSN_LENGTH): Define.
23348 * config/v850/v850.md (UNSPEC_LOOP): Define.
23349 (attr cpu): Add v850e3v5.
23350 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
23351 (movdi): New pattern.
23352 (movdi_internal): New pattern.
23353 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
23354 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
23355 (cstoresf4): Likewise.
23356 (cstoredf4): Likewise.
23357 (insv): New pattern.
23358 (rotlso3_a): New pattern.
23359 (rotlsi3_b): New pattern
23360 (rotlsi3_v850e3v5): New pattern.
23361 (doloop_begin): New pattern.
23362 (fix_loop_counter): New pattern.
23363 (doloop_end): New pattern.
23364 (branch_normal): Add e3v5 long branch support.
23365 (branch_invert): Likewise.
23366 (branch_z_normal): Likewise.
23367 (branch_z_invert): Likewise.
23368 (branch_nz_normal): Likewise.
23369 (branch_nz_invert): Likewise.
23370 (call_internal_short): Add e3v5 register-indirect JARL support.
23371 (call_internal_long): Likewise.
23372 (call_value_internal_short): Likewise.
23373 (call_value_internal_long): Likewise.
23374 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
23375 (mloop): New option.
23376 * config.gcc: Add support for configuring v840e3v5 target.
23377 * doc/invoke.texi: Document new v850 specific command line options.
23379 2013-01-31 Paul Koning <ni1d@arrl.net>
23383 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
23384 children if parent is a class.
23385 (prune_unused_types_prune): Don't add DW_AT_declaration.
23387 2013-01-31 Richard Biener <rguenther@suse.de>
23389 PR tree-optimization/56157
23390 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
23391 match up operand with SLP child.
23393 2013-01-31 Jason Merrill <jason@redhat.com>
23396 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
23397 parameters the first time.
23398 (gen_scheduled_generic_parms_dies): Check completeness here.
23400 2013-01-31 Richard Biener <rguenther@suse.de>
23402 PR middle-end/53073
23403 * common.opt (faggressive-loop-optimizations): New flag,
23404 enabled by default.
23405 * doc/invoke.texi (faggressive-loop-optimizations): Document.
23406 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
23407 infer_loop_bounds_from_undefined by it.
23409 2013-01-31 Richard Biener <rguenther@suse.de>
23411 PR tree-optimization/56150
23412 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
23413 visit virtual operands.
23414 (find_uses_to_rename_bb): Likewise.
23416 2013-01-31 Richard Biener <rguenther@suse.de>
23418 PR tree-optimization/56150
23419 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
23420 mixed store non-store stmts.
23422 2013-01-30 Jakub Jelinek <jakub@redhat.com>
23425 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
23426 LIBASAN_EARLY_SPEC is defined.
23427 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
23428 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
23430 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
23433 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
23434 invalid args instead of ICEing on it.
23435 (ix86_valid_target_attribute_tree): Return error_mark_node if
23436 ix86_valid_target_attribute_inner_p failed.
23437 (ix86_valid_target_attribute_p): Return false only if
23438 ix86_valid_target_attribute_tree returned error_mark_node. Allow
23439 target("default") attribute.
23440 (sorted_attr_string): Change argument from const char * to tree,
23441 merge in all target attribute arguments rather than just one.
23442 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
23443 instead of free. Avoid using strcat.
23444 (ix86_mangle_function_version_assembler_name): Mangle
23445 target("default") as if no target attribute is present. Adjust
23446 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
23447 instead of xmalloc and XDELETEVEC instead of free.
23448 (ix86_function_versions): Don't return true if one of the decls
23449 doesn't have target attribute. If they don't and one of the decls
23450 is DECL_FUNCTION_VERSIONED, report an error. Adjust
23451 sorted_attr_string caller. Use XDELETEVEC instead of free.
23452 (ix86_supports_function_versions): Remove.
23453 (make_name): Fix up formatting.
23454 (make_dispatcher_decl): Remove resolver_name and its initialization.
23455 Avoid leaking memory.
23456 (is_function_default_version): Return true if there is
23457 target("default") attribute rather than no target attribute at all.
23458 (make_resolver_func): Avoid leaking memory.
23459 (ix86_generate_version_dispatcher_body): Likewise.
23460 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
23461 * target.def (supports_function_versions): Remove.
23462 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
23463 * doc/tm.texi: Regenerated.
23465 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
23467 PR rtl-optimization/56144
23468 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
23469 for values with side effects.
23471 2013-01-30 Richard Biener <rguenther@suse.de>
23473 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
23474 (sparseset_pop): Likewise.
23475 * cfganal.c (compute_idf): Likewise. Increase work-stack size
23476 to be able to use quick_push in the worker loop.
23478 2013-01-30 Marek Polacek <polacek@redhat.com>
23480 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
23482 2013-01-30 Richard Biener <rguenther@suse.de>
23485 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
23487 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
23489 PR tree-optimization/56064
23490 * fixed-value.c (fixed_from_double_int): New function.
23491 * fixed-value.h (fixed_from_double_int): New prototype.
23492 (const_fixed_from_double_int): New static inline function.
23493 * fold-const.c (native_interpret_fixed): New static function.
23494 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
23495 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
23496 (native_encode_fixed): New static function.
23497 (native_encode_expr) <FIXED_CST>: Use it.
23498 (native_interpret_int): Move double_int worker code to...
23499 * double-int.c (double_int::from_buffer): ...this new static method.
23500 * double-int.h (double_int::from_buffer): Prototype it.
23502 2013-01-30 Richard Biener <rguenther@suse.de>
23504 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
23505 New pointer-map and obstack.
23506 (init_alias_vars): Allocate pointer-map and obstack.
23507 (delete_points_to_sets): Free them.
23508 (find_what_var_points_to): Cache result.
23509 (find_what_p_points_to): Adjust for changed interface of
23510 find_what_var_points_to.
23511 (compute_points_to_sets): Likewise.
23512 (ipa_pta_execute): Likewise.
23514 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23516 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
23517 * configure: Regenerate.
23518 * config.in: Regenerate.
23519 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
23520 #nobits/#progbits if supported.
23522 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
23525 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
23526 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
23527 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
23529 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23531 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
23532 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
23534 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23536 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
23537 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
23539 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23541 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
23543 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
23544 * config/arm/cortex-a7.md: New bypasses using
23545 arm_mac_accumulator_is_result.
23547 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23549 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
23550 (cortex_a7_neon_mla): Likewise.
23551 (cortex_a7_fpfmad): New reservation.
23552 (cortex_a7_fpmacs): Use ffmas and update required units.
23553 (cortex_a7_fpmuld): Update required units and latency.
23554 (cortex_a7_fpmacd): Likewise.
23555 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
23556 (cortex_a7_neon). Likewise.
23557 (bypass) Update participating units.
23559 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23561 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
23562 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
23564 * config/arm/vfp11.md (vfp_farith): Use ffmas.
23565 (vfp_fmul): Use ffmad.
23566 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
23567 (cortex_r4_fmacd): Use ffmad.
23568 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
23569 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
23570 (cortex_a9_fmacd): Use ffmad.
23571 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
23572 (cortex_a8_vfp_macd): Use ffmad.
23573 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
23574 (cortex_a5_fpmacd): Use ffmad.
23575 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
23576 (cortex_a15_vfp_macd): Use ffmad.
23577 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
23579 2013-01-29 Jason Merrill <jason@redhat.com>
23582 * varasm.c (default_assemble_visibility): Don't warn about
23583 visibility on artificial decls.
23585 2013-01-29 Richard Biener <rguenther@suse.de>
23587 PR tree-optimization/56113
23588 * tree-ssa-structalias.c (equiv_class_lookup): Also return
23590 (label_visit): Free duplicate bitmaps and record the leader instead.
23591 (perform_var_substitution): Adjust.
23593 2013-01-29 Richard Biener <rguenther@suse.de>
23595 PR tree-optimization/55270
23596 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
23597 the CFG, schedule loops for fixup.
23599 2013-01-29 Nick Clifton <nickc@redhat.com>
23601 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
23604 2013-01-28 Leif Ekblad <leif@rdos.net>
23606 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
23607 * config/i386/i386.h (TARGET_RDOS): New macro.
23608 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
23609 * config/i386/i386.c (ix86_option_override_internal): For 64bit
23610 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
23611 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
23612 DEFAULT_LARGE_SECTION_THRESHOLD.
23613 * config/i386/i386.md (R14_REG, R15_REG): New constants.
23614 * config/i386/rdos.h: New file.
23615 * config/i386/rdos64.h: New file.
23617 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
23620 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
23623 2013-01-28 Jakub Jelinek <jakub@redhat.com>
23625 PR rtl-optimization/56117
23626 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
23627 call cselib_lookup_from_insn on the MEM before calling
23628 add_insn_mem_dependence.
23630 2013-01-28 Richard Biener <rguenther@suse.de>
23632 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
23633 to a stmt that didn't have one.
23634 (copy_phis_for_bb): Likewise for PHI arguments.
23635 (copy_debug_stmt): Likewise for debug stmts.
23637 2013-01-28 Richard Biener <rguenther@suse.de>
23639 PR tree-optimization/56034
23640 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
23641 (partition_builtin_p): Adjust.
23642 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
23643 it is the last partition.
23644 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
23645 up the vertex for the definition.
23646 (classify_partition): Classify whether a partition is a
23647 PKIND_REDUCTION, thus has uses outside of the loop.
23648 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
23649 Merge all PKIND_REDUCTION partitions into the last partition.
23650 (tree_loop_distribution): Seed partitions from reductions as well.
23652 2013-01-28 Jakub Jelinek <jakub@redhat.com>
23654 PR tree-optimization/56125
23655 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
23656 pow(x,c) into sqrt(x) * powi(x, n/2) or
23657 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
23658 optimizing for size.
23659 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
23660 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
23663 PR tree-optimization/56094
23664 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
23665 to UNKNOWN_LOCATION while gimplifying expr.
23667 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
23670 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
23671 operand 0 in movabs insn template for -masm=intel asm alternative.
23672 (*movabs<mode>_2): Ditto for operand 1.
23674 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
23677 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
23680 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23682 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
23685 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23687 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
23688 Undef to avoid warning.
23690 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
23692 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
23693 * configure: Regenerate.
23695 2013-01-25 Jakub Jelinek <jakub@redhat.com>
23697 PR tree-optimization/56098
23698 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
23699 for stmts with volatile ops.
23700 (cond_store_replacement): Don't optimize if assign has volatile ops.
23701 (cond_if_else_store_replacement_1): Don't optimize if either
23702 then_assign or else_assign have volatile ops.
23703 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
23706 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23708 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
23710 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23712 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
23713 missing ':' in asm example.
23715 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
23717 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
23718 entries into lane and laneq entries.
23719 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
23720 Remove AdvSIMD scalar modes.
23721 (aarch64_sq<r>dmulh_laneq<mode>): New.
23722 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
23724 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
23725 builtin implementations to relfect changes in RTL in aarch64-simd.md.
23726 * config/aarch64/iterators.md (VCOND): New.
23729 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23732 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
23733 Add NULL LIBNAME argument to existing definitions.
23734 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
23735 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
23736 * config/avr/avr.c (DEF_BUILTIN): Same.
23737 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
23738 (avr_expand_builtin): Expand to a vanilla call if a libgcc
23739 implementation is available (DECL_ASSEMBLER_NAME is set).
23740 (avr_fold_absfx): New static function.
23741 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
23742 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
23743 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
23744 AVR_BUILTIN_ABSLLK.
23745 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
23746 (abshk, absk, abslk, absllk): Provide as static inline functions.
23748 2013-01-25 Marek Polacek <polacek@redhat.com>
23750 PR tree-optimization/56035
23751 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
23753 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
23755 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
23756 (*movtf_internal_rex64): Add (!o,C) alternative
23757 (*movxf_internal_rex64): Ditto.
23758 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
23760 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
23762 * doc/invoke.texi: fix typo.
23763 * doc/objc.texi: fix typo.
23765 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
23767 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
23768 for the first two alternatives.
23770 2013-01-24 Diego Novillo <dnovillo@google.com>
23772 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
23773 (ggc-zone.o): Remove.
23774 * configure.ac: Remove option --with-gc.
23775 * configure: Re-generate.
23776 * doc/install.texi: Remove documentation for --with-gc.
23777 * gengtype.c (write_enum_defn): Remove. Update all users.
23778 (write_Types_process_field): Remove generation of gt_e_* argument.
23779 (output_type_enum): Remove. Update all users.
23780 (write_enum_defn): Remove. Update all users.
23781 (enum alloc_zone): Remove. Update all users.
23782 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
23783 * ggc-common.c (ggc_splay_alloc): Remove first argument.
23784 Update all callers.
23785 (struct ptr_data): Remove field TYPE. Update all users.
23786 (gt_pch_note_object): Remove argument TYPE. Update all users.
23787 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
23789 * ggc-none.c (ggc_alloc_typed_stat): Remove.
23790 (struct alloc_zone): Remove.
23791 (ggc_internal_alloc_zone_stat): Remove.
23792 (ggc_internal_cleared_alloc_zone_stat): Remove.
23793 * ggc-page.c (ggc_alloc_typed_stat): Remove.
23794 (ggc_pch_count_object): Remove last argument. Update all users.
23795 (ggc_pch_alloc_object): Remove last argument. Update all users.
23796 (struct alloc_zone): Remove.
23797 * ggc-zone.c: Remove.
23798 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
23799 (struct alloc_zone): Remove.
23800 (ggc_alloc_typed_stat): Remove.
23801 (ggc_alloc_typed): Remove.
23802 (ggc_splay_alloc): Remove first argument.
23803 (rtl_zone): Remove. Update all users.
23804 (tree_zone): Remove. Update all users.
23805 (tree_id_zone): Remove. Update all users.
23806 (ggc_internal_zone_alloc_stat): Remove. Update all users.
23807 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
23808 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
23809 * tree-ssanames.c: Remove references to zone allocator in comments.
23811 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
23813 * config/avr/avr.c (avr_out_fract): Make register numbers that
23814 might be outside of source operand signed.
23816 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
23818 * config/i386/constraints.md (Yf): New constraint.
23819 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
23820 of f constraint to conditionaly disable x87 register preferences.
23821 (*movdf_internal): Ditto.
23822 (*movsf_internal): Ditto.
23824 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
23826 PR inline-asm/55934
23827 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
23828 that have operands with impossible constraints.
23829 Add a FIXME for a speed-up opportunity.
23830 * lra-constraints.c (process_alt_operands): Verify that a class
23831 selected from constraints on asms is valid for the operand mode.
23832 (curr_insn_transform): Remove incorrect comment.
23834 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
23836 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
23837 TOC operand is a valid symbol ref in the constant pool.
23839 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23841 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
23843 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
23846 * config/avr/stdfix.h: New file.
23847 * t-avr (stdfix-gcc.h): New rule to build it.
23848 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
23850 2013-01-23 Kostya Serebryany <kcc@google.com>
23852 * config/darwin.h: remove dependency on
23853 CoreFoundation (asan on Mac OS).
23855 2013-01-23 Jakub Jelinek <jakub@redhat.com>
23858 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
23859 instead of cmpdi_operand for first comparison operand.
23860 Don't assert that comparison operands aren't both constants.
23862 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
23864 * doc/install.texi (Downloading the Source): Update references to
23865 downloading separate components.
23867 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
23869 * doc/extend.texi (__int128): Improve grammar.
23871 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
23874 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
23875 alternative to (o,r).
23876 (*movdi_internal_rex64): Remove (!o,n) alternative.
23877 (DImode immediate->memory splitter): Remove.
23878 (DImode immediate->memory peephole2): Remove.
23879 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
23880 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
23881 alternative to (!o,*r).
23882 (*movtf_internal_sse): New pattern.
23883 (*movxf_internal_rex64): New pattern.
23884 (*movxf_internal): Disable for TARGET_64BIT.
23885 (*movdf_internal_rex64): Remove (!o,F) alternative.
23887 2013-01-22 Jakub Jelinek <jakub@redhat.com>
23889 PR middle-end/56074
23890 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
23891 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
23892 * tree-vect-loop-manip.c (find_loop_location): Also ignore
23893 stmt locations where LOCATION_LOCUS of the stmt location is
23894 UNKNOWN_LOCATION or BUILTINS_LOCATION.
23897 * config/i386/i386.md (UNSPEC_STOS): New.
23898 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
23899 *strsetqi_1): Add UNSPEC_STOS.
23901 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
23904 * doc/invoke.texi: Remove left over -Wsynth example.
23906 2013-01-21 Jakub Jelinek <jakub@redhat.com>
23908 PR tree-optimization/56051
23909 * fold-const.c (fold_binary_loc): Don't fold
23910 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
23911 a narrowing conversion, or widening conversion from signed
23914 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
23916 PR rtl-optimization/56023
23917 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
23918 dependent on debug instruction.
23920 2013-01-21 Martin Jambor <mjambor@suse.cz>
23922 PR middle-end/56022
23923 * function.c (allocate_struct_function): Call
23924 invoke_set_current_function_hook earlier.
23926 2013-01-21 Jakub Jelinek <jakub@redhat.com>
23928 * reload1.c (init_reload): Only initialize reload_obstack
23929 during the first call.
23931 2013-01-21 Marek Polacek <polacek@redhat.com>
23933 * cfgloop.c (verify_loop_structure): Fix up grammar.
23935 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
23937 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
23938 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
23940 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23943 * config/arm/marvell-pj4.md: Update copyright year.
23944 Fix up use of alu to alu_reg and simple_alu_imm.
23946 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
23948 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
23950 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
23953 * lra-constraints.c (curr_insn_transform): Don't reuse original
23954 insn for secondary memory move when memory mode should be different.
23956 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23958 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
23959 atomic_storedi_1): New patterns.
23961 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23963 btver2 pipeline descriptions.
23964 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
23966 * config/i386/i386.md (btver2_decode): New type attributes.
23967 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
23969 * config/i386/btver2.md: New file describing btver2 pipelines.
23971 2013-01-19 Andrew Pinski <apinski@cavium.com>
23973 PR tree-optimization/52631
23974 * tree-ssa-sccvn (visit_use): Before looking up the original
23975 statement, try looking up the simplified expression.
23977 2013-01-19 Anthony Green <green@moxielogic.com>
23979 * config/moxie/moxie.c (moxie_expand_prologue): Set
23980 current_function_static_stack_size.
23982 2013-01-18 Jakub Jelinek <jakub@redhat.com>
23984 PR tree-optimization/56029
23985 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
23986 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
23988 2013-01-18 Sharad Singhai <singhai@google.com>
23990 PR tree-optimization/55995
23991 * dumpfile.c (dump_loc): Print location only if available.
23992 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
23994 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
23997 * lra-constraints.c (curr_insn_transform): Reuse original insn for
23998 secondary memory move.
23999 (inherit_reload_reg): Use rclass instead of cl for
24000 check_secondary_memory_needed_p.
24002 2013-01-18 Jakub Jelinek <jakub@redhat.com>
24004 PR middle-end/56015
24005 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
24006 the case where writing real complex part of target modifies op1.
24008 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
24010 * config/aarch64/aarch64-simd.md
24011 (aarch64_vcond_internal<mode>): Handle unordered cases.
24012 * config/aarch64/iterators.md (v_cmp_result): New.
24014 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
24015 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24017 * config/arm/marvell-pj4.md: New file.
24018 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
24019 * config/arm/arm.md (generic_sched): Add marvell_pj4.
24020 (generic_vfp): Likewise.
24021 * config/arm/arm-cores.def: Add marvell-pj4.
24022 * config/arm/arm-tune.md: Regenerate.
24023 * config/arm/arm-tables.opt: Regenerate.
24024 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
24025 * doc/invoke.texi: Document marvell-pj4.
24027 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
24029 * config/aarch64/arm_neon.h: Map scalar types to standard types.
24031 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24036 * var-tracking.c (negative_power_of_two_p): New.
24037 (global_get_addr_cache, local_get_addr_cache): New.
24038 (get_addr_from_global_cache, get_addr_from_local_cache): New.
24039 (vt_canonicalize_addr): Rewrite using the above. Adjust the
24041 (vt_stack_offset_p): Remove.
24042 (vt_canon_true_dep): Always canonicalize loc's address.
24043 (clobber_overlapping_mems): Make sure we have a MEM.
24044 (local_get_addr_clear_given_value): New.
24045 (val_reset): Clear local cached entries.
24046 (compute_bb_dataflow): Create and release the local cache.
24047 Disable duplicate MEMs clobbering.
24048 (emit_notes_in_bb): Clobber MEMs likewise.
24049 (vt_emit_notes): Create and release the local cache.
24050 (vt_initialize, vt_finalize): Create and release the global
24051 cache, respectively.
24052 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
24054 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24056 PR libmudflap/53359
24057 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
24058 not found in the symtab.
24060 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24063 PR rtl-optimization/55547
24064 PR rtl-optimization/53827
24067 * alias.c (offset_overlap_p): New, factored out of...
24068 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
24069 the conservative special case for symbolic constants. Don't
24070 adjust zero sizes on alignment.
24072 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
24074 PR rtl-optimization/52573
24075 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
24076 REG_UNUSED for the same register.
24078 2013-01-17 Richard Biener <rguenther@suse.de>
24079 Marek Polacek <polacek@redhat.com>
24081 PR rtl-optimization/55833
24082 * loop-unswitch.c (unswitch_loops): Move loop verification...
24083 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
24084 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
24085 Set it to true when we're removing a loop from hierarchy tree in
24086 an irreducible region.
24087 (fix_bb_placements): Adjust caller.
24088 (fix_loop_placements): Likewise.
24090 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
24092 * config/avr/builtins.def (DEF_BUILTIN): Factor out
24093 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
24094 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
24095 Remove ID. Adjust comments.
24096 * config/avr/avr-c.c (avr_builtin_name): Remove.
24097 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
24098 * config/avr/avr.c (avr_tolower): New static function.
24099 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
24100 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
24101 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
24104 2013-01-17 Jan Hubicka <jh@suse.cz>
24106 PR tree-optimization/55273
24107 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
24109 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
24112 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
24113 store through atomic_store<mode>_1.
24114 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
24116 2013-01-17 Martin Jambor <mjambor@suse.cz>
24118 PR tree-optimizations/55264
24119 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
24120 for virtual methods.
24121 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
24122 virtual methods before inlining is over.
24123 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
24125 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
24128 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
24130 PR rtl-optimization/56005
24131 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
24132 pending reads for prefetch.
24134 2013-01-16 Ian Bolton <ian.bolton@arm.com>
24136 * config/aarch64/aarch64.md
24137 (*cstoresi_neg_uxtw): New pattern.
24138 (*cmovsi_insn_uxtw): New pattern.
24139 (*<optab>si3_uxtw): New pattern.
24140 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
24141 (*<optab>si3_insn_uxtw): New pattern.
24142 (*bswapsi2_uxtw): New pattern.
24144 2013-01-16 Richard Biener <rguenther@suse.de>
24146 * tree-inline.c (tree_function_versioning): Remove set but
24147 never used variable.
24149 2013-01-16 Richard Biener <rguenther@suse.de>
24151 PR tree-optimization/55964
24152 * tree-flow.h (rename_variables_in_loop): Remove.
24153 (rename_variables_in_bb): Likewise.
24154 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
24155 (copy_loop_before): Adjust and delete update-ssa status.
24156 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
24157 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
24158 (rename_variables_in_loop): Remove.
24159 (slpeel_update_phis_for_duplicate_loop): Likewise.
24160 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
24161 use available cfg machinery instead of duplicating it.
24162 Update PHI nodes and perform poor-mans SSA update here.
24163 (slpeel_tree_peel_loop_to_edge): Adjust.
24165 2013-01-16 Richard Biener <rguenther@suse.de>
24167 PR tree-optimization/54767
24168 PR tree-optimization/53465
24169 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
24170 (vrp_visit_phi_node): For PHI arguments coming via backedges
24171 drop all symbolical range information.
24172 (execute_vrp): Compute backedges.
24174 2013-01-16 Richard Biener <rguenther@suse.de>
24176 * doc/install.texi: Update CLooG and ISL requirements to
24179 2013-01-16 Christian Bruel <christian.bruel@st.com>
24182 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
24183 (broken_move): Handle UNSPECV_SP_SWITCH_B.
24184 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
24186 2013-01-16 DJ Delorie <dj@redhat.com>
24188 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
24189 (UNSPECV_SP_SWITCH_E): New.
24190 (sp_switch_1): Change to an unspec.
24191 (sp_switch_2): Change to an unspec. Don't use post-inc when we
24194 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
24196 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
24197 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
24198 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
24199 (expand_mem_thread_fence): Ditto.
24200 (expand_mem_signal_fence): Ditto.
24201 (expand_atomic_load): Ditto.
24202 (expand_atomic_store): Ditto.
24204 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
24206 PR rtl-optimization/55547
24207 PR rtl-optimization/53827
24210 * alias.c (memrefs_conflict_p): Set sizes to negative after
24213 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24216 * function.c (thread_prologue_and_epilogue_insns): Always
24217 add crtl->drap_reg to set_up_by_prologue.set, even if
24218 stack_realign_drap is false.
24220 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
24222 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
24223 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
24224 *call): Fix indention.
24226 2013-01-15 Tom de Vries <tom@codesourcery.com>
24229 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
24232 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
24234 PR rtl-optimization/55153
24235 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
24237 2013-01-15 Martin Jambor <mjambor@suse.cz>
24239 PR tree-optimization/55920
24240 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
24241 accesses as grp_to_be_debug_replaced.
24243 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24245 PR tree-optimization/55920
24246 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
24247 there is non-useless type conversion needed from debug rhs to lhs,
24248 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
24250 2013-01-15 Joseph Myers <joseph@codesourcery.com>
24251 Mikael Pettersson <mikpe@it.uu.se>
24254 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
24256 (ASM_OUTPUT_CASE_LABEL): Remove.
24257 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
24258 * final.c (shorten_branches): Update alignment of labels before
24259 jump tables if CASE_VECTOR_SHORTEN_MODE.
24261 2013-01-15 Richard Biener <rguenther@suse.de>
24264 * system.h: Do not include gmp.h for building host tools.
24266 2013-01-15 Richard Biener <rguenther@suse.de>
24268 PR middle-end/55882
24269 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
24270 account for bitpos when computing alignment.
24272 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
24274 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
24275 (ix86_target_macros_internal): Likewise.
24277 * config/i386/i386.c (m_CORE2I7): Removed.
24278 (m_CORE_HASWELL): New macro.
24279 (m_CORE_ALL): Likewise.
24280 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
24281 (initial_ix86_arch_features): Likewise.
24282 (processor_target_table): Initializations for Core avx2.
24283 (cpu_names): New names "core-avx2".
24284 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
24285 PROCESSOR_CORE_HASWELL.
24286 (ix86_issue_rate): New case.
24287 (ia32_multipass_dfa_lookahead): Likewise.
24288 (ix86_sched_init_global): Likewise.
24290 * config/i386/i386.h (TARGET_HASWELL): New macro.
24291 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
24292 (processor_type): New PROCESSOR_HASWELL.
24294 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24296 PR tree-optimization/55955
24297 * tree-vect-loop.c (vectorizable_reduction): Give up early on
24298 *SHIFT_EXPR and *ROTATE_EXPR codes.
24300 PR tree-optimization/48766
24301 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
24302 -ftrapv disable -fwrapv.
24304 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24307 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
24308 etc. to 1 and not to __flash.
24309 Use LL suffix for __INT24_MAX__ with -mint8.
24310 Use ULL suffix for __UINT24_MAX__ with -mint8.
24312 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24314 * config/avr/avr-arch.h
24315 (struct base_arch_s): Use typedef avr_arch_t instead.
24316 (struct arch_info_s): Use typedef avr_arch_info_t instead.
24317 (struct mcu_type_s): Use typedef avr_mcu_t instead.
24318 * config/avr/avr.c: Same.
24319 * config/avr/avr-devices.c: Same.
24320 * config/avr/driver-avr.c: Same.
24321 * config/avr/gen-avr-mmcu-texi.c: Same.
24322 * config/avr/avr-mcus.def: Adjust comment.
24324 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
24326 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
24327 * config/aarch64/iterators.md (VALLDI): New.
24329 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
24330 Andi Kleen <ak@linux.intel.com>
24333 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
24334 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
24337 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24339 * config/avr/avr-stdint.h: Remove trailing blanks.
24340 * config/avr/avr-log.h: Same.
24341 * config/avr/avr-arch.h: Same.
24342 * config/avr/avr-devices.c: Same.
24343 * config/avr/avr-dimode.md: Same.
24344 * config/avr/predicates.md: Same.
24345 * config/avr/avr-c.c: Same. And fix typo.
24347 * config/avr/avr-protos.h: Same. And:
24348 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
24349 (init_cumulative_args): Rename to avr_init_cumulative_args.
24350 (expand_prologue): Rename to avr_expand_prologue.
24351 (expand_epilogue): Rename to avr_expand_epilogue.
24352 (adjust_insn_length): Rename to avr_adjust_insn_length.
24353 (notice_update_cc): Rename to avr_notice_update_cc.
24354 (final_prescan_insn): Rename to avr_final_prescan_insn.
24355 * config/avr/avr.c: Same.
24356 * config/avr/avr.h: Same.
24357 * config/avr/avr.md: Remove trailing blanks.
24358 (prologue): Use avr_expand_prologue.
24359 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
24361 2013-01-14 Richard Biener <rguenther@suse.de>
24363 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
24364 verify_location, collect_subblocks): New functions.
24365 (verify_gimple_in_cfg): Verify that locations only reference
24366 BLOCKs in the functions BLOCK tree.
24368 2013-01-14 Richard Biener <rguenther@suse.de>
24370 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
24372 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
24374 (insert_out_of_ssa_copy_on_edge): Likewise.
24375 (rewrite_close_phi_out_of_ssa): Likewise.
24376 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
24378 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
24379 propagated constants.
24380 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
24383 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24385 * config/avr/avr-modes.def: Add GPL copyright notice.
24387 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
24389 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
24390 MEMMODEL_MASK to determine memory model.
24391 (atomic_store<mode>): Ditto from operands[2].
24392 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
24394 2013-01-13 Jakub Jelinek <jakub@redhat.com>
24397 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
24398 (fold_gimple_assign): Don't call unshare_expr here.
24399 (fold_ctor_reference): Call unshare_expr.
24401 2013-01-13 Terry Guo <terry.guo@arm.com>
24403 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
24404 * doc/fragments.texi: Document MULTILIB_REUSE.
24405 * gcc.c (multilib_reuse): New internal spec.
24406 (set_multilib_dir): Also search multilib from multilib_reuse.
24407 * genmultilib (tmpmultilib3): Refactor code.
24408 (tmpmultilib4): Ditto.
24409 (multilib_reuse): New multilib argument.
24411 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
24413 * Makefile.in: Update copyright.
24415 2013-01-12 Tom de Vries <tom@codesourcery.com>
24417 PR middle-end/55890
24418 * calls.c (expand_call): Check if arg_nr is valid.
24420 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
24422 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
24423 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
24424 documentation. Add missing '__' in front of
24425 __builtin_ia32_packssdw256.
24427 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24430 * config/s390/s390.c (s390_preferred_reload_class): Do not return
24431 NO_REGS for larl operands.
24432 (s390_reload_larl_operand): Use s390_load_address instead of
24435 2013-01-11 Richard Biener <rguenther@suse.de>
24437 * tree-cfg.c (verify_node_sharing_1): Split out from ...
24438 (verify_node_sharing): ... here.
24439 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
24441 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
24443 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
24444 Substitute TREECHECKING.
24445 * configure: Regenerate.
24446 * Makefile.in (TREECHECKING): New.
24448 2013-01-11 Richard Guenther <rguenther@suse.de>
24450 PR tree-optimization/44061
24451 * tree-vrp.c (extract_range_basic): Compute zero as
24452 value-range for __builtin_constant_p of function parameters.
24454 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
24456 Update copyright years.
24458 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
24460 PR rtl-optimization/55672
24461 * lra-eliminations.c (mark_not_eliminable): Permit addition with
24462 const to be eliminable.
24464 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
24466 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
24467 * configure: Regenerate.
24469 2013-01-10 Richard Biener <rguenther@suse.de>
24471 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
24473 2013-01-10 Richard Biener <rguenther@suse.de>
24476 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
24477 locations for virtual PHI arguments.
24478 (rewrite_update_phi_arguments): Likewise.
24480 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
24482 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
24485 2013-01-10 Jakub Jelinek <jakub@redhat.com>
24487 PR tree-optimization/55921
24488 * tree-complex.c (expand_complex_asm): New function.
24489 (expand_complex_operations_1): Call it for GIMPLE_ASM.
24491 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24494 * config/s390/s390.c (s390_symref_operand_p)
24495 (s390_loadrelative_operand_p): Merge the two functions.
24496 (s390_check_qrst_address, print_operand_address): Add parameters
24497 to s390_loadrelative_operand_p invokation.
24498 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
24499 (s390_reload_larl_operand, s390_secondary_reload): Use
24500 s390_loadrelative_operand_p instead of s390_symref_operand_p.
24501 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
24503 2013-01-09 Mike Stump <mikestump@comcast.net>
24505 * dse.c (record_store): Remove unnecessary assert.
24507 2013-01-09 Jan Hubicka <jh@suse.cz>
24509 PR tree-optimization/55569
24510 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
24511 * cfgloop.h (scale_loop_profile): Likewise.
24513 2013-01-09 Jan Hubicka <jh@suse.cz>
24516 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
24518 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
24520 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
24522 PR middle-end/55114
24523 * expr.h (maybe_emit_group_store): Declare.
24524 * expr.c (maybe_emit_group_store): New function.
24525 * builtins.c (expand_builtin_int_roundingfn): Call it.
24526 (expand_builtin_int_roundingfn_2): Likewise.
24528 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
24530 PR rtl-optimization/55829
24531 * lra-constraints.c (match_reload): Add code for absent output.
24532 (curr_insn_transform): Add code for reloads of matched inputs
24535 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
24537 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
24538 attribute of movddup insn to DF.
24539 (*vec_interleave_lowv2df): Ditto.
24540 (vec_dupv2df): Ditto.
24542 2013-01-09 Jan Hubicka <jh@suse.cz>
24544 PR tree-optimiation/55875
24545 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
24546 EVERY_ITERATION parameter.
24547 (number_of_iterations_exit): Check if exit is executed every iteration.
24548 (idx_infer_loop_bounds): Similarly here.
24549 (n_of_executions_at_most): Simplify
24550 to only test for cases where statement is dominated by the
24551 particular bound; handle correctly the "postdominance" test.
24552 (scev_probably_wraps_p): Use max loop iterations info
24553 as a global bound first.
24555 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
24556 Nick Clifton <nickc@redhat.com>
24558 * config/v850/v850.md (cbranchsf4): New pattern.
24559 (cstoresf4): New pattern.
24560 (cbranchdf4): New pattern.
24561 (cstoredf4): New pattern.
24562 (movsicc): Disallow floating point comparisons.
24563 (cmpsf_le_insn): Fix order of operators.
24564 (cmpsf_lt_insn): Likewise.
24565 (cmpsf_eq_insn): Likewise.
24566 (cmpdf_le_insn): Likewise.
24567 (cmpdf_lt_insn): Likewise.
24568 (cmpdf_eq_insn): Likewise.
24569 (cmpsf_ge_insn): Use LE comparison.
24570 (cmpdf_ge_insn): Likewise.
24571 (cmpsf_gt_insn): Use LT comparison.
24572 (cmpdf_gt_insn): Likewise.
24573 (cmpsf_ne_insn): Delete pattern.
24574 (cmpdf_ne_insn): Delete pattern.
24575 * config/v850/v850.c (v850_gen_float_compare): Use
24576 gen_cmpdf_eq_insn for NE comparison.
24577 (v850_float_z_comparison_operator)
24578 (v850_float_nz_comparison_operator): Move from here ...
24579 * config/v850/predicates.md: ... to here. Move GT and GE
24580 comparisons into v850_float_z_comparison_operator.
24581 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
24583 (v850_float_nz_comparison_operator): Likewise.
24585 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24587 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
24588 with calls to gen_insvsi/gen_insvdi.
24590 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
24592 * config/i386/i386.c (initial_ix86_tune_features): Set up
24593 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
24595 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
24596 Jakub Jelinek <jakub@redhat.com>
24598 PR tree-optimization/48189
24599 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
24600 If nitercst is 0, don't predict the exit edge.
24602 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24604 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
24605 in asm_fprintf with reg_names.
24606 (aarch64_print_operand_address): Likewise.
24607 (aarch64_return_addr): Likewise.
24608 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
24610 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24612 * config/pa/pa.h (VAL_U6_BITS_P): Define.
24613 (INT_U6_BITS): Likewise.
24614 * config/pa/predicates.md (uint6_operand): New predicate.
24615 (shift5_operand, shift6_operand): Likewise.
24616 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
24618 (lshrdi3): Use shift6_operand.
24619 (shrpsi4, shrpdi4): New insn patterns.
24620 (extzv): Delete expander.
24621 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
24622 predicates in unamed zero extract patterns. Tighten common constraint.
24623 (extv): Delete expander.
24624 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
24625 predicates in unamed sign extract patterns. Tighten common constraint.
24626 (insv): Delete expander.
24627 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
24628 predicates in unamed insert patterns. Tighten common constraint.
24629 Change uint32_operand predicate to uint6_operand predicate in unamed
24630 DImode pattern to insert constant values of type 1...1xxxx.
24632 2013-01-04 Jan Hubicka <jh@suse.cz>
24634 PR tree-optimization/55823
24635 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
24638 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24639 Uros Bizjak <ubizjak@gmail.com>
24641 PR rtl-optimization/55845
24642 * df-problems.c (can_move_insns_across): Stop scanning at
24643 volatile_insn_p source instruction or give up if
24644 across_from .. across_to range contains any volatile_insn_p
24647 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
24649 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
24650 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
24652 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
24653 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
24655 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24658 * asan.c (asan_clear_shadow): New function.
24659 (asan_emit_stack_protection): Use it.
24661 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
24663 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
24664 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
24665 with tab instead of space.
24667 2013-01-08 Nick Clifton <nickc@redhat.com>
24669 * config/rl78/rl78.c (rl78_expand_prologue): Always select
24670 register bank 0 at the start of an interrupt handler.
24671 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
24674 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
24676 * config/aarch64/aarch64-simd.md
24677 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
24678 (aarch64_simd_bsl): Likewise.
24679 (aarch64_vcond_internal<mode>): Likewise.
24680 (vcond<mode><mode>): Likewise.
24681 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
24682 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
24684 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
24686 * config/aarch64/aarch64-builtins.c
24687 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
24689 2013-01-08 Martin Jambor <mjambor@suse.cz>
24692 * tree-sra.c (analyze_access_subtree): Return true also after
24693 potentially creating a debug-only replacement.
24695 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24697 PR middle-end/55890
24698 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
24700 PR tree-optimization/54120
24701 * tree-vrp.c (range_fits_type_p): Don't allow
24702 src_precision < precision from signed vr to unsigned_p
24703 if vr->min or vr->max is negative.
24704 (simplify_float_conversion_using_ranges): Test can_float_p
24705 against CODE_FOR_nothing.
24707 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24708 Richard Biener <rguenther@suse.de>
24710 PR middle-end/55851
24711 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
24712 types instead of just INTEGER_TYPE types.
24714 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
24716 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
24719 2013-01-07 Steve Ellcey <sellcey@mips.com>
24722 * config/mips/mips.opt: Change mad to mmad to match documentation.
24724 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24727 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
24728 .progmemx.data now.
24730 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24733 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
24734 (avr_addrspace_t): Add .section_name field.
24735 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
24737 (avr_addrspace): Same. Initialize .section_name. Remove last
24738 NULL entry. Put __memx into .progmemx.data.
24739 (progmem_section_prefix): Remove.
24740 (avr_asm_init_sections): No need to initialize progmem_section.
24741 (avr_asm_named_section): Use avr_addrspace[].section_name to get
24742 section name prefix.
24743 (avr_asm_select_section): Ditto. And use get_unnamed_section to
24744 retrieve the progmem section.
24745 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
24746 boundary to run over avr_addrspace[].
24747 (avr_register_target_pragmas): Ditto.
24749 2013-01-06 Jakub Jelinek <jakub@redhat.com>
24751 * varasm.c (output_constant_def_contents): For asan_protect_global
24752 protected strings, adjust DECL_ALIGN if needed, before testing for
24754 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
24755 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
24757 (output_object_block): For asan protected decls, emit asan padding
24758 after their contents.
24759 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
24760 (asan_finish_file): Test it here instead.
24762 2013-01-07 Nick Clifton <nickc@redhat.com>
24763 Matthias Klose <doko@debian.org>
24764 Doug Kwan <dougkwan@google.com>
24765 H.J. Lu <hongjiu.lu@intel.com>
24768 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
24770 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
24772 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
24774 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
24776 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
24778 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24781 * doc/install.texi (Cross-Compiler-Specific Options): Document
24784 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
24786 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
24787 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
24788 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
24789 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
24790 vqmovun_high_s64): Fix source operand number and update copyright.
24792 2013-01-07 Richard Biener <rguenther@suse.de>
24794 PR middle-end/55890
24795 * gimple.h (gimple_call_builtin_p): New overload.
24796 * gimple.c (validate_call): New function.
24797 (gimple_call_builtin_p): Likewise.
24798 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
24799 Use gimple_call_builtin_p.
24800 (find_func_clobbers): Likewise.
24801 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
24802 (strlen_optimize_stmt): Likewise.
24804 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
24806 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
24807 (vld1q_dup_*): Likewise.
24808 (vld1_*): Likewise.
24809 (vld1q_*): Likewise.
24810 (vld1_lane_*): Likewise.
24811 (vld1q_lane_*): Likewise.
24813 2013-01-07 Richard Biener <rguenther@suse.de>
24815 * lto-streamer.h (LTO_minor_version): Bump to 2.
24817 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
24819 * config/aarch64/aarch64-protos.h
24820 (aarch64_const_double_zero_rtx_p): Rename to...
24821 (aarch64_float_const_zero_rtx_p): ...this.
24822 (aarch64_float_const_representable_p): New.
24823 (aarch64_output_simd_mov_immediate): Likewise.
24824 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
24825 move immediate case.
24826 * config/aarch64/aarch64.c
24827 (aarch64_const_double_zero_rtx_p): Rename to...
24828 (aarch64_float_const_zero_rtx_p): ...this.
24829 (aarch64_print_operand): Allow printing of new constants.
24830 (aarch64_valid_floating_const): New.
24831 (aarch64_legitimate_constant_p): Check for valid floating-point
24833 (aarch64_simd_valid_immediate): Likewise.
24834 (aarch64_vect_float_const_representable_p): New.
24835 (aarch64_float_const_representable_p): Likewise.
24836 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
24837 (aarch64_output_simd_mov_immediate): New.
24838 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
24839 (*movdf_aarch64): Likewise.
24840 * config/aarch64/constraints.md (Ufc): New.
24841 (Y): call aarch64_float_const_zero_rtx.
24842 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
24844 2013-01-07 Richard Biener <rguenther@suse.de>
24846 PR tree-optimization/55888
24847 PR tree-optimization/55862
24848 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
24849 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
24850 not if it is contained therein.
24852 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24854 * config/avr/t-avr: Typo.
24856 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24859 * config/avr/t-avr: Don't automatically rebuild
24860 $(srcdir)/config/avr/t-multilib
24861 $(srcdir)/config/avr/avr-tables.opt
24862 $(srcdir)/doc/avr-mmcu.texi
24863 (avr-mcus): New phony target to build them on request.
24864 (s-avr-mlib, s-avr-mmcu-texi): Remove.
24865 * avr/avr-mcus.def: Adjust comments.
24867 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
24869 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
24871 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
24873 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
24875 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
24877 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
24879 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
24881 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
24882 to generate profiling.
24883 * config/rs6000/aix64.h (LIB_SPEC): Same.
24885 2013-01-04 Andrew Pinski <apinski@cavium.com>
24887 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
24889 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
24891 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
24893 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
24895 (ix86_expand_move): Ditto.
24896 (ix86_zero_extend_to_Pmode): Ditto.
24897 (ix86_expand_call): Ditto.
24898 (ix86_expand_special_args_builtin): Ditto.
24899 (ix86_expand_builtin): Ditto.
24901 2013-01-04 Richard Biener <rguenther@suse.de>
24903 PR tree-optimization/55862
24904 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
24905 translating them through PHI nodes.
24907 2013-01-04 Martin Jambor <mjambor@suse.cz>
24909 PR tree-optimization/55755
24910 * tree-sra.c (sra_modify_assign): Do not check that an access has no
24911 children when trying to avoid producing a VIEW_CONVERT_EXPR.
24913 2013-01-04 Marek Polacek <polacek@redhat.com>
24915 PR middle-end/55859
24916 * opts.c (default_options_optimization): Clarify error message.
24918 2013-01-04 Richard Biener <rguenther@suse.de>
24920 PR middle-end/55863
24921 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
24924 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24927 * config/pa/pa.md (movsi): Revert previous change.
24928 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
24931 2013-01-03 Richard Henderson <rth@redhat.com>
24933 * config/i386/i386.c (ix86_expand_move): Always assign to op1
24934 after eliminating TLS symbols.
24936 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
24939 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
24940 * graphite-poly.c (debug_gmp_value): Likewise.
24942 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
24945 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
24946 selected code model, define __code_mode_small__, __code_model_medium__,
24947 __code_model_large__, __code_model_32__ or __code_model_kernel__.
24948 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
24949 xchg temporary register with %k. Declare temporary register as
24951 [__x86_64__]: For medium and large code models, preserve %rbx register.
24953 2013-01-03 Richard Biener <rguenther@suse.de>
24955 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
24956 (dump_subscript): Adjust.
24957 (finalize_ddr_dependent): Do not dump redundant info.
24958 (analyze_siv_subscript): Adjust.
24959 (subscript_dependence_tester): Likewise.
24960 (compute_affine_dependence): Likewise.
24962 2013-01-03 Richard Biener <rguenther@suse.de>
24965 2013-01-03 Richard Biener <rguenther@suse.de>
24967 PR tree-optimization/55857
24968 * tree-vect-stmts.c (vectorizable_load): Do not setup
24969 re-alignment for invariant loads.
24971 2013-01-02 Richard Biener <rguenther@suse.de>
24973 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
24974 invariant load do not generate a vector load from the scalar location.
24976 2013-01-03 Richard Biener <rguenther@suse.de>
24978 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
24979 for not vectorizing.
24980 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
24981 not build INDIRECT_REFs, call get_name once only.
24982 (vect_create_data_ref_ptr): Likewise. Dump base object kind
24983 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
24985 2013-01-03 Richard Biener <rguenther@suse.de>
24987 PR tree-optimization/55857
24988 * tree-vect-stmts.c (vectorizable_load): Do not setup
24989 re-alignment for invariant loads.
24991 2013-01-03 Richard Biener <rguenther@suse.de>
24994 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
24995 prefer a built-in decl.
24997 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24999 * gcc.c (process_command): Update copyright notice dates.
25000 * gcov.c (print_version): Likewise.
25001 * gcov-dump.c (print_version): Likewise.
25003 PR rtl-optimization/55838
25004 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
25005 iv0.step, iv1.step and step.
25007 2013-01-03 Jakub Jelinek <jakub@redhat.com>
25008 Marc Glisse <marc.glisse@inria.fr>
25010 PR tree-optimization/55832
25011 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
25012 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
25013 integer_{one,zero}_node.
25015 2013-01-03 Jakub Jelinek <jakub@redhat.com>
25018 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
25019 * var-tracking.c (reverse_op): Don't add reverse ops to
25020 VALUEs that have already
25021 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
25023 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
25025 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
25027 2013-01-02 Teresa Johnson <tejohnson@google.com>
25029 * dumpfile.c (dump_loc): Print filename with location.
25030 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
25031 new location_t parameter to emit complete unroll message with
25032 new dump framework.
25033 (canonicalize_loop_induction_variables): Compute loops location
25034 and pass to try_unroll_loop_completely.
25035 * loop-unroll.c (report_unroll_peel): New function.
25036 (peel_loops_completely): Use new dump format with location
25037 for main dumpfile message, and invoke report_unroll_peel on success.
25038 (decide_unrolling_and_peeling): Ditto.
25039 (decide_peel_once_rolling): Remove old dumpfile message subsumed
25040 by report_unroll_peel.
25041 (decide_peel_completely): Ditto.
25042 (decide_unroll_constant_iterations): Ditto.
25043 (decide_unroll_runtime_iterations): Ditto.
25044 (decide_peel_simple): Ditto.
25045 (decide_unroll_stupid): Ditto.
25046 * cfgloop.c (get_loop_location): New function.
25047 * cfgloop.h (get_loop_location): Declare.
25049 2013-01-02 Sriraman Tallam <tmsriram@google.com>
25051 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
25054 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25056 PR middle-end/55198
25057 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
25058 BLKmode objects when EXPAND_MEMORY is specified.
25060 2013-01-02 Sriraman Tallam <tmsriram@google.com>
25062 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
25064 (fold_builtin_cpu): Do not share cpu model decls across statements.
25066 2013-01-02 Jason Merrill <jason@redhat.com>
25069 * tree.c (build_array_type_1): Revert earlier change.
25071 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
25073 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
25075 * config/aarch64/aarch64-tune.md: Re-generate.
25077 2013-01-02 Richard Biener <rguenther@suse.de>
25079 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
25080 invariant load do not generate a vector load from the scalar location.
25082 2013-01-02 Richard Biener <rguenther@suse.de>
25085 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
25086 * configure: Regenerate.
25088 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
25090 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
25091 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
25092 (expand_builtin_int_roundingfn_2): Keep the original target around
25093 for the fallback case.
25095 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
25097 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
25098 to be clear for sign changes.
25100 2013-01-01 Jan Hubicka <jh@suse.cz>
25102 * ipa-inline-analysis.c: Fix formatting.
25104 2013-01-01 Jakub Jelinek <jakub@redhat.com>
25106 PR tree-optimization/55831
25107 * tree-vect-loop.c (get_initial_def_for_induction): Use
25108 gsi_after_labels instead of gsi_start_bb.
25110 Copyright (C) 2013 Free Software Foundation, Inc.
25112 Copying and distribution of this file, with or without modification,
25113 are permitted in any medium without royalty provided the copyright
25114 notice and this notice are preserved.