1 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
4 * ira-color.c (update_costs_from_copies): Add new parameter. Use
5 it for calling update_costs_from_allocno.
6 (assign_hard_reg): Call restore_costs_from_copies only for
7 !retry_p. Pass new argument to update_costs_from_copies.
8 (color_pass): Pass new argument to update_costs_from_copies.
9 (ira_mark_allocation_change): Ditto.
11 2013-10-30 Sharad Singhai <singhai@google.com>
14 * opts.c (common_handle_option): Remove deprecated option
15 -ftree-vectorizer-verbose.
16 * doc/invoke.texi (Debugging Options): Ditto.
17 * opts-global.c (handle_common_deferred_options): Ditto.
18 (dump_remap_tree_vectorizer_verbose): Delete.
19 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
21 2013-10-30 DJ Delorie <dj@redhat.com>
23 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
24 that take no arguments.
26 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
29 * reload1.c (update_eliminables_and_spill): New function, broken
31 (reload): Use it. Check for frame size change after frame
32 size alignment, and call update_eliminables_and_spill first
35 2013-10-30 Cong Hou <congh@google.com>
38 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
39 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
40 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
42 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
44 PR rtl-optimization/58369
45 * reload1.c (compute_reload_subreg_offset): New function.
46 (choose_reload_regs): Use it to pass endian-correct
47 offset to subreg_regno_offset.
49 2013-10-30 Tobias Burnus <burnus@net-b.de>
52 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
55 2013-10-30 Jason Merrill <jason@redhat.com>
57 * configure.ac (loose_warn): Add -Wno-format if
58 --disable-build-format-warnings.
60 2013-10-30 David Malcolm <dmalcolm@redhat.com>
62 * cgraphunit.c (analyze_functions): Split symtab_node
63 declarations onto multiple lines to make things easier
66 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
67 (symtab_semantically_equivalent_p): Likewise.
69 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
72 * lra.c (check_rtl): Remove address check before LRA work.
74 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
76 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
77 POINTER_PLUS_EXPR in the defining statement.
79 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
82 * tree-pass.h (make_pass_regmove): Remove.
83 * timevar.def (TV_REGMOVE): Remove.
84 * passes.def (pass_regmove): Remove.
85 * opts.c (default_options_table): Remove entry for regmove.
86 * doc/passes.texi: Remove regmove pass description.
87 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
89 (-fdump-rtl-regmove): Ditto.
90 * common.opt (foptimize-register-move, fregmove): Ignore.
91 * Makefile.in (OBJS): Remove regmove.o.
93 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
95 (struct ira_allocno) New member allocno_prefs.
96 (ALLOCNO_PREFS): New macro.
97 (ira_prefs, ira_prefs_num): New external vars.
98 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New
100 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref):
102 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
104 (ira_add_allocno_copy_to_list): Remove prototype.
105 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
106 (ira_pref_iterator): New type.
107 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
108 (FOR_EACH_PREF): New macro.
109 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
110 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
112 (ira_setup_alts): New function.
113 (decrease_live_ranges_number): New function.
114 (ira): Call the above function.
115 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
116 (ira_create_allocno): Initialize allocno prefs.
117 (pref_pool, pref_vec): New static vars.
118 (initiate_prefs, find_allocno_pref, ira_create_pref): New
120 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
121 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
122 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
123 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
124 (ira_add_allocno_copy_to_list): Make static. Rename to
125 add_allocno_copy_to_list.
126 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
127 swap_allocno_copy_ends_if_necessary.
128 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
129 ira_remove_allocno_prefs.
130 (ira_flattening): Ditto.
131 (ira_build): Call initiate_prefs, print_prefs.
132 (ira_destroy): Call finish_prefs.
133 * ira-color.c (struct update_cost_record): New.
134 (struct allocno_color_data): Add new member update_cost_records.
135 (update_cost_record_pool): New static var.
136 (init_update_cost_records, get_update_cost_record): New functions.
137 (free_update_cost_record_list, finish_update_cost_records): Ditto.
138 (struct update_cost_queue_elem): Add member from.
139 (initiate_cost_update): Call init_update_cost_records.
140 (finish_cost_update): Call finish_update_cost_records.
141 (queue_update_cost, get_next_update_cost): Add new param from.
142 (Update_allocno_cost, update_costs_from_allocno): New functions.
143 (update_costs_from_prefs): Ditto.
144 (update_copy_costs): Rename to update_costs_from_copies.
145 (restore_costs_from_copies): New function.
146 (update_conflict_hard_regno_costs): Don't go back.
147 (assign_hard_reg): Call restore_costs_from_copies. Add printing
149 (pop_allocnos): Add priniting more debug info.
150 (color_allocnos): Remove prefs for conflicting hard regs.
151 Call update_costs_from_prefs.
152 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
153 (get_dup_num): Rename, modify, and move to ira.c
154 (process_regs_for_copy): Add prefs.
155 (add_insn_allocno_copies): Put src as first arg of
156 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
157 * ira-costs.c (record_reg_classes): Modify and move code into
158 record_operands_costs.
159 (find_costs_and_classes): Create prefs for the hard reg of small
161 (process_bb_node_for_hard_reg_moves): Add prefs.
163 2013-10-30 Richard Biener <rguenther@suse.de>
166 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
167 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
169 (pre_and_rev_post_order_compute): ... which now wraps it.
170 * graph.c (draw_cfg_nodes_no_loops): Use
171 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
174 2013-10-30 Christian Bruel <christian.bruel@st.com>
176 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
177 (sh_expand_cmpstr): Handle known align and schedule improvements.
178 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
179 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
181 2013-10-30 Martin Jambor <mjambor@suse.cz>
183 PR rtl-optimization/10474
184 * ira.c (find_moveable_pseudos): Do not calculate dominance info
186 (interesting_dest_for_shprep): New function.
187 (split_live_ranges_for_shrink_wrap): Likewise.
188 (ira): Calculate dominance info and df analysis. Call
189 split_live_ranges_for_shrink_wrap.
191 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
194 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
196 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
198 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
199 * tree.h (POINTER_BOUNDS_P): New.
200 (BOUNDED_TYPE_P): New.
202 (pointer_bounds_type_node): New.
203 * tree.c (build_common_tree_nodes): Initialize
204 pointer_bounds_type_node.
205 * gimple.h (gimple_call_get_nobnd_arg_index): New.
206 (gimple_call_num_nobnd_args): New.
207 (gimple_call_nobnd_arg): New.
208 (gimple_return_retbnd): New.
209 (gimple_return_set_retbnd): New
210 * gimple.c (gimple_build_return): Increase number of ops
211 for return statement.
212 (gimple_call_get_nobnd_arg_index): New.
213 * gimple-pretty-print.c (dump_gimple_return): Print second op.
215 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
217 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
218 with "chkp ctor" and "bnd_legacy" attributes.
219 * gimplify.c (gimplify_init_constructor): Avoid infinite
220 loop during gimplification of bounds initializer.
222 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
224 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
225 (handle_bnd_legacy): New.
226 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
227 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
230 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
232 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
233 (BT_FN_PTR_CONST_PTR): New.
234 (BT_FN_CONST_PTR_CONST_PTR): New.
235 (BT_FN_PTR_CONST_PTR_SIZE): New.
236 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
237 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
238 (BT_FN_VOID_CONST_PTR_SIZE): New.
239 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
240 * chkp-builtins.def: New.
241 * builtins.def: include chkp-builtins.def.
242 (DEF_CHKP_BUILTIN): New.
243 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
244 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
245 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
246 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
247 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
248 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
249 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
250 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
251 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
252 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
253 * common.opt (fcheck-pointer-bounds): New.
254 * toplev.c (process_options): Check Pointer Bounds Checker is supported.
255 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
257 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
259 * target.def (builtin_chkp_function): New.
260 (chkp_bound_type): New.
261 (chkp_bound_mode): New.
262 (fn_abi_va_list_bounds_size): New.
263 (load_bounds_for_arg): New.
264 (store_bounds_for_arg): New.
265 * targhooks.h (default_load_bounds_for_arg): New.
266 (default_store_bounds_for_arg): New.
267 (default_fn_abi_va_list_bounds_size): New.
268 (default_chkp_bound_type): New.
269 (default_chkp_bound_mode): New.
270 (default_builtin_chkp_function): New.
271 * targhooks.c (default_load_bounds_for_arg): New.
272 (default_store_bounds_for_arg): New.
273 (default_fn_abi_va_list_bounds_size): New.
274 (default_chkp_bound_type): New.
275 (default_chkp_bound_mode); New.
276 (default_builtin_chkp_function): New.
277 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
278 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
279 (TARGET_STORE_BOUNDS_FOR_ARG): New.
280 (TARGET_BUILTIN_CHKP_FUNCTION): New.
281 (TARGET_CHKP_BOUND_TYPE): New.
282 (TARGET_CHKP_BOUND_MODE): New.
283 * doc/tm.texi: Regenerated.
284 * langhooks.h (lang_hooks): Add chkp_supported field.
285 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
286 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
288 2013-10-29 Andrew Pinski <apinski@cavium.com>
290 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
291 (ifcombine_ifandif): Handle cases where
292 maybe_fold_and_comparisons fails, combining the branches
294 (tree_ssa_ifcombine): Inverse the order of
295 the basic block walk, increases the number of combinings.
296 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
298 2013-10-29 Mike Stump <mikestump@comcast.net>
300 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
301 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
302 (make_vector_mode): Increase namebuf to 16.
303 (emit_insn_modes_h): When processing BImode, don't
304 also match partial int modes.
305 (emit_class_narrowest_mode): Likewise.
307 * config/bfin/bfin-modes.def: Add precision to PDI.
308 * config/m32c/m32c-modes.def: Add precision to PSI.
309 * config/msp430/msp430-modes.def: Add precision to PSI.
310 * config/rs6000/rs6000-modes.def: Add precision to PTI.
311 * config/sh/sh-modes.def: Add precision to PSI and PDI.
313 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
316 * config/sh/sh.md (*addc): Rename existing variations to ...
317 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
318 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
319 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
320 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
322 2013-10-29 Teresa Johnson <tejohnson@google.com>
325 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
326 insanities when updating probabilities.
328 2013-10-29 David Malcolm <dmalcolm@redhat.com>
330 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
331 prettyprinter for cgraph_node to reflect the conversion of the
332 symtable types to a C++ class hierarchy: it now *is* a
333 symtab_node_base, rather than having one (named "symbol").
335 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
337 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
338 __cilkrts_pop_frame. If matched, then return true for built-in
340 (expand_builtin): Added BUILT_IN_CILK_DETACH and
341 BUILT_IN_CILK_POP_FRAME case.
342 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
343 (lhs_cilk_detect_spawn): Likewise.
344 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
345 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
346 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
347 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
348 (LANG_HOOKS_CILKPLUS): Likewise.
349 * tree.h (CILK_SPAWN_FN): Likewise.
350 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
351 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
352 (OBJS): Added cilk-common.o.
353 (BUILTINS_DEF): Added cilk-builtins.def.
354 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
355 (lhd_cilk_detect_spawn): Likewise.
356 * langhooks.h (lang_hooks_for_cilkplus): New struct.
357 (struct lang_hooks): Added new field called "cilkplus."
358 * cilk-common.c: New file.
360 * cilk-builtins.def: Likewise.
361 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
362 "__cilk" macro and set it to 200.
363 * function.h (struct function::cilk_frame_decl): New field.
364 (struct function::is_cilk_function): Likewise.
365 (struct function::calls_cilk_spawn): Likewise.
366 * gimplify.c (gimplify_call_expr): Added a check if the function call
367 being gimplified is a spawn detach point. If so, then add pop_frame
368 and detach function calls.
369 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
370 for gimplifying _Cilk_spawn and _Cilk_sync statements.
371 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
372 function. If so, added a _Cilk_sync and gimplified it.
373 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
374 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
375 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
377 (can_inline_edge_p): Prevent inling of spawnee function.
378 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
379 for functions that use Cilk keywords.
380 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
381 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
382 CILK_SYNC_STMT cases.
383 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
385 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
386 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
387 * passes.texi (Cilk Keywords): New section that describes the compiler
388 code changes for handling Cilk Keywords.
390 2013-10-29 David Malcolm <dmalcolm@redhat.com>
392 Patch autogenerated by refactor_symtab.py from
393 https://github.com/davidmalcolm/gcc-refactoring-scripts
394 revision 58bb219cc090b2f4516a9297d868c245495ee622
396 * asan.c (asan_finish_file): Update for conversion of symtab types to
397 a true class hierarchy.
398 * cfgexpand.c (estimated_stack_frame_size): Likewise.
399 * cgraph.c (cgraph_get_body): Likewise.
400 (cgraph_get_create_real_symbol_node): Likewise.
401 (verify_cgraph_node): Likewise.
402 (verify_edge_corresponds_to_fndecl): Likewise.
403 (verify_edge_count_and_frequency): Likewise.
404 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
405 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
406 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
407 (cgraph_node_cannot_return): Likewise.
408 (cgraph_set_pure_flag_1): Likewise.
409 (cgraph_set_const_flag_1): Likewise.
410 (cgraph_set_nothrow_flag_1): Likewise.
411 (cgraph_make_node_local_1): Likewise.
412 (cgraph_for_node_and_aliases): Likewise.
413 (cgraph_for_node_thunks_and_aliases): Likewise.
414 (cgraph_node_can_be_local_p): Likewise.
415 (cgraph_node_cannot_be_local_p_1): Likewise.
416 (cgraph_function_body_availability): Likewise.
417 (dump_cgraph_node): Likewise.
418 (cgraph_rtl_info): Likewise.
419 (cgraph_mark_address_taken_node): Likewise.
420 (cgraph_remove_node): Likewise.
421 (cgraph_release_function_body): Likewise.
422 (cgraph_update_edges_for_call_stmt_node): Likewise.
423 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
424 (cgraph_make_edge_direct): Likewise.
425 (cgraph_resolve_speculation): Likewise.
426 (cgraph_speculative_call_info): Likewise.
427 (cgraph_turn_edge_to_speculative): Likewise.
428 (cgraph_create_edge_1): Likewise.
429 (cgraph_set_call_stmt): Likewise.
430 (cgraph_node_for_asm): Likewise.
431 (cgraph_add_thunk): Likewise.
432 (cgraph_same_body_alias): Likewise.
433 (cgraph_create_function_alias): Likewise.
434 (cgraph_create_node): Likewise.
435 (cgraph_create_empty_node): Likewise.
436 (record_function_versions): Likewise.
437 (used_from_object_file_p): Likewise.
438 * cgraph.h (symtab_can_be_discarded): Likewise.
439 (symtab_real_symbol_p): Likewise.
440 (cgraph_mark_force_output_node): Likewise.
441 (cgraph_edge_recursive_p): Likewise.
442 (symtab_alias_target): Likewise.
443 (varpool_all_refs_explicit_p): Likewise.
444 (varpool_can_remove_if_no_refs): Likewise.
445 (cgraph_only_called_directly_or_aliased_p): Likewise.
446 (cgraph_next_function_with_gimple_body): Likewise.
447 (cgraph_first_function_with_gimple_body): Likewise.
448 (cgraph_function_with_gimple_body_p): Likewise.
449 (cgraph_next_function): Likewise.
450 (cgraph_first_function): Likewise.
451 (cgraph_next_defined_function): Likewise.
452 (cgraph_first_defined_function): Likewise.
453 (varpool_next_defined_variable): Likewise.
454 (varpool_first_defined_variable): Likewise.
455 (varpool_next_static_initializer): Likewise.
456 (varpool_first_static_initializer): Likewise.
457 (varpool_next_variable): Likewise.
458 (varpool_first_variable): Likewise.
459 (varpool_node_name): Likewise.
462 (is_a_helper <varpool_node>::test): Likewise.
463 (is_a_helper <cgraph_node>::test): Likewise.
464 (varpool_variable_node): Likewise.
465 (cgraph_function_or_thunk_node): Likewise.
466 (varpool_alias_target): Likewise.
467 (cgraph_alias_target): Likewise.
468 (cgraph_node_name): Likewise.
469 (varpool_node_asm_name): Likewise.
470 (cgraph_node_asm_name): Likewise.
471 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
472 (cgraph_rebuild_references): Likewise.
473 (rebuild_cgraph_edges): Likewise.
474 (record_eh_tables): Likewise.
475 (build_cgraph_edges): Likewise.
476 (mark_store): Likewise.
477 (mark_load): Likewise.
478 (mark_address): Likewise.
479 (record_type_list): Likewise.
480 (record_reference): Likewise.
481 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
482 (cgraph_materialize_clone): Likewise.
483 (cgraph_function_versioning): Likewise.
484 (cgraph_copy_node_for_versioning): Likewise.
485 (update_call_expr): Likewise.
486 (cgraph_find_replacement_node): Likewise.
487 (cgraph_create_virtual_clone): Likewise.
488 (cgraph_clone_node): Likewise.
489 * cgraphunit.c (compile): Likewise.
490 (output_weakrefs): Likewise.
491 (output_in_order): Likewise.
492 (expand_function): Likewise.
493 (assemble_thunks_and_aliases): Likewise.
494 (expand_thunk): Likewise.
495 (mark_functions_to_output): Likewise.
496 (handle_alias_pairs): Likewise.
497 (analyze_functions): Likewise.
498 (walk_polymorphic_call_targets): Likewise.
499 (varpool_finalize_decl): Likewise.
500 (process_function_and_variable_attributes): Likewise.
501 (cgraph_process_same_body_aliases): Likewise.
502 (analyze_function): Likewise.
503 (cgraph_add_new_function): Likewise.
504 (cgraph_finalize_function): Likewise.
505 (referred_to_p): Likewise.
506 (cgraph_reset_node): Likewise.
507 (cgraph_process_new_functions): Likewise.
508 (enqueue_node): Likewise.
509 (decide_is_symbol_needed): Likewise.
510 * coverage.c (coverage_compute_profile_id): Likewise.
511 * dbxout.c (dbxout_expand_expr): Likewise.
512 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
513 (reference_to_unused): Likewise.
514 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
515 * gimplify.c (unvisit_body): Likewise.
516 (unshare_body): Likewise.
517 * ipa-cp.c (ipcp_generate_summary): Likewise.
518 (ipcp_decision_stage): Likewise.
519 (identify_dead_nodes): Likewise.
520 (decide_whether_version_node): Likewise.
521 (decide_about_value): Likewise.
522 (perhaps_add_new_callers): Likewise.
523 (create_specialized_node): Likewise.
524 (update_profiling_info): Likewise.
525 (ipcp_propagate_stage): Likewise.
526 (estimate_local_effects): Likewise.
527 (good_cloning_opportunity_p): Likewise.
528 (devirtualization_time_bonus): Likewise.
529 (propagate_constants_accross_call): Likewise.
530 (initialize_node_lattices): Likewise.
531 (ipcp_cloning_candidate_p): Likewise.
532 (determine_versionability): Likewise.
533 (print_all_lattices): Likewise.
534 (print_lattice): Likewise.
535 (ipcp_discover_new_direct_edges): Likewise.
536 * ipa-devirt.c (ipa_devirt): Likewise.
537 (likely_target_p): Likewise.
538 (update_type_inheritance_graph): Likewise.
539 (possible_polymorphic_call_target_p): Likewise.
540 (dump_possible_polymorphic_call_targets): Likewise.
541 (devirt_variable_node_removal_hook): Likewise.
542 (record_binfo): Likewise.
543 (maybe_record_node): Likewise.
544 (build_type_inheritance_graph): Likewise.
545 * ipa-inline-analysis.c (inline_write_summary): Likewise.
546 (inline_generate_summary): Likewise.
547 (inline_analyze_function): Likewise.
548 (do_estimate_growth): Likewise.
549 (simple_edge_hints): Likewise.
550 (estimate_node_size_and_time): Likewise.
551 (estimate_edge_devirt_benefit): Likewise.
552 (compute_inline_parameters): Likewise.
553 (estimate_function_body_sizes): Likewise.
554 (compute_bb_predicates): Likewise.
555 (initialize_inline_failed): Likewise.
556 (dump_inline_summary): Likewise.
557 (dump_inline_edge_summary): Likewise.
558 * ipa-inline-transform.c (inline_transform): Likewise.
559 (preserve_function_body_p): Likewise.
560 (save_inline_function_body): Likewise.
561 (inline_call): Likewise.
562 (clone_inlined_nodes): Likewise.
563 (can_remove_node_now_p): Likewise.
564 (can_remove_node_now_p_1): Likewise.
565 * ipa-inline.c (early_inliner): Likewise.
566 (early_inline_small_functions): Likewise.
567 (inline_always_inline_functions): Likewise.
568 (ipa_inline): Likewise.
569 (flatten_function): Likewise.
570 (inline_small_functions): Likewise.
571 (speculation_useful_p): Likewise.
572 (recursive_inlining): Likewise.
573 (update_caller_keys): Likewise.
574 (reset_edge_caches): Likewise.
575 (update_edge_key): Likewise.
576 (edge_badness): Likewise.
577 (relative_time_benefit): Likewise.
578 (want_inline_self_recursive_call_p): Likewise.
579 (want_inline_small_function_p): Likewise.
580 (want_early_inline_function_p): Likewise.
581 (num_calls): Likewise.
582 (can_early_inline_edge_p): Likewise.
583 (can_inline_edge_p): Likewise.
584 (report_inline_failed_reason): Likewise.
585 * ipa-profile.c (ipa_profile): Likewise.
586 (ipa_propagate_frequency): Likewise.
587 (ipa_propagate_frequency_1): Likewise.
588 (ipa_profile_generate_summary): Likewise.
589 * ipa-prop.c (ipcp_transform_function): Likewise.
590 (read_replacements_section): Likewise.
591 (ipa_prop_read_section): Likewise.
592 (ipa_modify_call_arguments): Likewise.
593 (ipa_print_node_params): Likewise.
594 (propagate_controlled_uses): Likewise.
595 (update_indirect_edges_after_inlining): Likewise.
596 (remove_described_reference): Likewise.
597 (ipa_make_edge_direct_to_target): Likewise.
598 (ipa_analyze_node): Likewise.
599 (ipa_analyze_params_uses): Likewise.
600 (ipa_compute_jump_functions): Likewise.
601 (ipa_get_callee_param_type): Likewise.
602 (ipa_print_node_jump_functions): Likewise.
603 (ipa_initialize_node_params): Likewise.
604 (ipa_populate_param_decls): Likewise.
605 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
606 (write_agg_replacement_chain): Likewise.
607 (ipa_write_node_info): Likewise.
608 (ipa_edge_duplication_hook): Likewise.
609 (try_decrement_rdesc_refcount): Likewise.
610 * ipa-pure-const.c (propagate_nothrow): Likewise.
611 (propagate_pure_const): Likewise.
612 (pure_const_read_summary): Likewise.
613 (pure_const_write_summary): Likewise.
614 (analyze_function): Likewise.
615 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
616 (ipa_ref_referring_ref_list): Likewise.
617 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
618 (ipa_remove_stmt_references): Likewise.
619 (ipa_find_reference): Likewise.
620 (ipa_dump_referring): Likewise.
621 (ipa_dump_references): Likewise.
622 (ipa_record_reference): Likewise.
623 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
624 (ipa_reference_write_optimization_summary): Likewise.
625 (write_node_summary_p): Likewise.
626 (propagate): Likewise.
627 (read_write_all_from_decl): Likewise.
628 (generate_summary): Likewise.
629 (analyze_function): Likewise.
630 (propagate_bits): Likewise.
631 (ipa_reference_get_not_written_global): Likewise.
632 (ipa_reference_get_not_read_global): Likewise.
633 * ipa-split.c (execute_split_functions): Likewise.
634 (split_function): Likewise.
635 * ipa-utils.c (ipa_merge_profiles): Likewise.
636 (dump_cgraph_node_set): Likewise.
637 (ipa_reverse_postorder): Likewise.
638 (ipa_edge_within_scc): Likewise.
639 (ipa_get_nodes_in_cycle): Likewise.
640 (ipa_free_postorder_info): Likewise.
641 (ipa_reduced_postorder): Likewise.
643 (recursive_call_p): Likewise.
644 * ipa.c (ipa_cdtor_merge): Likewise.
645 (record_cdtor_fn): Likewise.
646 (function_and_variable_visibility): Likewise.
647 (varpool_externally_visible_p): Likewise.
648 (cgraph_externally_visible_p): Likewise.
649 (comdat_can_be_unshared_p): Likewise.
650 (comdat_can_be_unshared_p_1): Likewise.
651 (address_taken_from_non_vtable_p): Likewise.
652 (ipa_discover_readonly_nonaddressable_vars): Likewise.
653 (symtab_remove_unreachable_nodes): Likewise.
654 (walk_polymorphic_call_targets): Likewise.
655 (process_references): Likewise.
656 (enqueue_node): Likewise.
657 (has_addr_references_p): Likewise.
658 (cgraph_non_local_node_p_1): Likewise.
659 * is-a.h (varpool_analyze_node): Likewise.
660 * lto-cgraph.c (input_symtab): Likewise.
661 (merge_profile_summaries): Likewise.
662 (input_cgraph_1): Likewise.
663 (input_edge): Likewise.
664 (input_varpool_node): Likewise.
665 (input_node): Likewise.
666 (input_overwrite_node): Likewise.
667 (compute_ltrans_boundary): Likewise.
668 (output_refs): Likewise.
669 (lto_output_varpool_node): Likewise.
670 (lto_output_node): Likewise.
671 (reachable_from_other_partition_p): Likewise.
672 (referenced_from_other_partition_p): Likewise.
673 (lto_output_edge): Likewise.
674 (output_node_opt_summary): Likewise.
675 (add_node_to): Likewise.
676 (reachable_from_this_partition_p): Likewise.
677 (lto_set_symtab_encoder_in_partition): Likewise.
678 (lto_symtab_encoder_in_partition_p): Likewise.
679 (lto_set_symtab_encoder_encode_initializer): Likewise.
680 (lto_symtab_encoder_encode_initializer_p): Likewise.
681 (lto_set_symtab_encoder_encode_body): Likewise.
682 (lto_symtab_encoder_encode_body_p): Likewise.
683 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
685 * lto-streamer-in.c (lto_read_body): Likewise.
686 (fixup_call_stmt_edges): Likewise.
687 (fixup_call_stmt_edges_1): Likewise.
688 * lto-streamer-out.c (produce_symtab): Likewise.
689 (output_symbol_p): Likewise.
690 (write_symbol): Likewise.
691 (lto_output): Likewise.
692 (copy_function): Likewise.
693 (output_function): Likewise.
694 * passes.c (function_called_by_processed_nodes_p): Likewise.
695 (ipa_write_optimization_summaries): Likewise.
696 (ipa_write_summaries): Likewise.
697 (do_per_function_toporder): Likewise.
698 (do_per_function): Likewise.
699 (dump_passes): Likewise.
700 * symtab.c (symtab_semantically_equivalent_p): Likewise.
701 (symtab_nonoverwritable_alias): Likewise.
702 (symtab_nonoverwritable_alias_1): Likewise.
703 (symtab_for_node_and_aliases): Likewise.
704 (symtab_resolve_alias): Likewise.
705 (fixup_same_cpp_alias_visibility): Likewise.
706 (symtab_alias_ultimate_target): Likewise.
707 (symtab_used_from_object_file_p): Likewise.
708 (verify_symtab_base): Likewise.
709 (dump_symtab_base): Likewise.
710 (symtab_node_name): Likewise.
711 (symtab_node_asm_name): Likewise.
712 (symtab_dissolve_same_comdat_group_list): Likewise.
713 (symtab_add_to_same_comdat_group): Likewise.
714 (symtab_unregister_node): Likewise.
715 (symtab_insert_node_to_hashtable): Likewise.
716 (symtab_register_node): Likewise.
717 (unlink_from_assembler_name_hash): Likewise.
718 (insert_to_assembler_name_hash): Likewise.
719 (eq_assembler_name): Likewise.
720 (hash_node_by_assembler_name): Likewise.
722 (hash_node): Likewise.
723 * toplev.c (wrapup_global_declaration_2): Likewise.
724 * trans-mem.c (ipa_tm_execute): Likewise.
725 (ipa_tm_transform_clone): Likewise.
726 (ipa_tm_transform_transaction): Likewise.
727 (ipa_tm_transform_calls_redirect): Likewise.
728 (ipa_tm_insert_gettmclone_call): Likewise.
729 (ipa_tm_insert_irr_call): Likewise.
730 (ipa_tm_create_version): Likewise.
731 (ipa_tm_create_version_alias): Likewise.
732 (ipa_tm_mark_forced_by_abi_node): Likewise.
733 (ipa_tm_mark_force_output_node): Likewise.
734 (ipa_tm_diagnose_tm_safe): Likewise.
735 (ipa_tm_mayenterirr_function): Likewise.
736 (ipa_tm_scan_irr_function): Likewise.
737 (ipa_tm_note_irrevocable): Likewise.
738 (ipa_tm_scan_calls_clone): Likewise.
739 (get_cg_data): Likewise.
740 * tree-eh.c (tree_could_trap_p): Likewise.
741 * tree-emutls.c (ipa_lower_emutls): Likewise.
742 (create_emultls_var): Likewise.
743 (lower_emutls_function_body): Likewise.
744 (gen_emutls_addr): Likewise.
745 (emutls_decl): Likewise.
746 (new_emutls_decl): Likewise.
747 * tree-inline.c (tree_function_versioning): Likewise.
748 (optimize_inline_calls): Likewise.
749 (expand_call_inline): Likewise.
750 (estimate_num_insns): Likewise.
752 (delete_unreachable_blocks_update_callgraph): Likewise.
753 * tree-nested.c (gimplify_all_functions): Likewise.
754 (create_nesting_tree): Likewise.
755 (check_for_nested_with_variably_modified): Likewise.
756 * tree-pretty-print.c (dump_function_header): Likewise.
757 * tree-profile.c (tree_profiling): Likewise.
758 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
759 (modify_function): Likewise.
760 (convert_callers): Likewise.
761 (convert_callers_for_node): Likewise.
762 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
763 (associate_varinfo_to_alias): Likewise.
764 (create_variable_info_for): Likewise.
765 (get_constraint_for_ssa_var): Likewise.
766 * tree-vectorizer.c (increase_alignment): Likewise.
767 * tree.c (find_decls_types_in_var): Likewise.
768 (find_decls_types_in_node): Likewise.
769 (free_lang_data_in_decl): Likewise.
770 * value-prof.c (gimple_ic_transform): Likewise.
771 (gimple_ic): Likewise.
772 (check_ic_target): Likewise.
773 (init_node_map): Likewise.
774 * varasm.c (decl_binds_to_current_def_p): Likewise.
775 (default_binds_local_p_1): Likewise.
776 (dump_tm_clone_pairs): Likewise.
777 (assemble_alias): Likewise.
778 (find_decl): Likewise.
779 (mark_decl_referenced): Likewise.
780 * varpool.c (varpool_for_node_and_aliases): Likewise.
781 (varpool_extra_name_alias): Likewise.
782 (varpool_create_variable_alias): Likewise.
783 (add_new_static_var): Likewise.
784 (varpool_finalize_named_section_flags): Likewise.
785 (varpool_remove_unreferenced_decls): Likewise.
786 (enqueue_node): Likewise.
787 (varpool_assemble_decl): Likewise.
788 (assemble_aliases): Likewise.
789 (varpool_analyze_node): Likewise.
790 (cgraph_variable_initializer_availability): Likewise.
791 (varpool_add_new_variable): Likewise.
792 (ctor_for_folding): Likewise.
793 (dump_varpool_node): Likewise.
794 (varpool_remove_initializer): Likewise.
795 (varpool_remove_node): Likewise.
796 (varpool_node_for_decl): Likewise.
797 (varpool_create_empty_node): Likewise.
798 * config/i386/i386.c (ix86_generate_version_dispatcher_body):
800 (ix86_get_function_versions_dispatcher): Likewise.
802 2013-10-29 David Malcolm <dmalcolm@redhat.com>
804 * cgraph.h (symtab_node_base): Convert to a class;
805 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
806 chain_next/prev from symtab_node_def.
807 (cgraph_node): Inherit from symtab_node; add GTY option
808 tag ("SYMTAB_FUNCTION").
809 (varpool_node): Inherit from symtab_node; add GTY option
810 tag ("SYMTAB_VARIABLE").
811 (symtab_node_def): Remove.
812 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
813 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
814 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
815 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
817 * ipa-ref.h (symtab_node_def): Drop.
818 (symtab_node): Change underlying type from symtab_node_def to
820 (const_symtab_node): Likwise.
822 * is-a.h: Update examples in comment.
824 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
825 (assembler_name_hash): Likewise.
827 2013-10-29 Martin Liska <marxin.liska@gmail.com>
829 * doc/tree-ssa.texi (gimple_phi_result): Document.
830 (gimple_phi_num_args, gimple_phi_arg): Likewise.
831 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
832 (PHI_RESULT, PHI_NUM_ARGS): Remove.
833 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
835 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
837 * expr.h: Revert change and include tree-core.h.
838 * rtl.h: Revert change and don't include tree-core.h.
840 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
842 * config/darwin.c: Include gimple.h.
843 * config/i386/winnt.c: Likewise.
845 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
847 PR tree-optimization/19831
848 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
850 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
852 * tree-outof-ssa.h: Remove include files.
853 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
855 * tree-ssa-coalesce.c: Likewise.
856 * cfgexpand.c: Likewise.
857 * tree-ssa-ter.c: Likewise.
858 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
859 * lto-streamer.h: Likewise.
860 * cgraphbuild.c: Add gimple.h to include list.
861 * data-streamer-in.c: Likewise.
862 * ipa-cp.c: Likewise.
863 * tree-streamer.c: Likewise.
864 * lto-compress.c: Likewise.
865 * ipa-reference.c: Likewise.
866 * data-streamer-out.c: Likewise.
867 * lto-cgraph.c: Likewise.
868 * cgraphclones.c: Likewise.
869 * ipa-utils.c: Likewise.
870 * data-streamer.c: Likewise.
871 * ipa-split.c: Likewise.
872 * lto-section-in.c: Likewise.
873 * tree-streamer-out.c: Likewise.
874 * ipa-prop.c: Likewise.
875 * tree-streamer-in.c: Likewise.
876 * symtab.c: Likewise.
877 * opts-global.c: Likewise.
878 * lto-opts.c: Likewise.
879 * lto-section-out.c: Likewise.
880 * lto-streamer.c: Likewise.
881 * rtl.h: Add tree-core.h to include list.
882 * expr.h: Remove tree-core.h from include list.
883 * gimple.h: Likewise.
884 * ipa-utils.h: Likewise.
885 * streamer-hooks.h: Likewise.
886 * streamer-hooks.c: Include input.h.
888 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
890 * config/arm/arm.c (cortexa7_extra_costs): New table.
891 (arm_cortex_a7_tune): New.
892 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
894 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
896 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
898 2013-10-29 David Malcolm <dmalcolm@redhat.com>
900 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
901 to use autogenerated markers for a class-hierarchy, every class
902 must have a GTY marker.
903 * gengtype.h (struct type): Add linked list of subclasses to
904 the "s" member of the union.
905 (add_subclass): New decl.
906 * gengtype-state.c (read_state_struct_type): Set up subclass
908 * gengtype.c (get_ultimate_base_class): New.
910 (new_structure): Set up subclass linked list.
911 (set_gc_used_type): Propagate usage information to subclasses.
912 (output_mangled_typename): Use get_ultimate_base_class.
913 (walk_subclasses): Use the subclass linked list, avoiding an
914 O(N^2) when writing out all types.
915 (walk_type): Issue an error if the base class is missing a tag,
916 rather than generating bogus C code. Add a gcc_unreachable
917 default case, in case people omit tags from concrete subclasses,
918 or get the values wrong.
919 (write_func_for_structure): Issue an error for subclasses for
920 which the base doesn't have a "desc", since otherwise the
921 autogenerated routines for the base would silently fail to visit
923 (write_root): Use get_ultimate_base_class, tweaking constness of
924 tp to match that function's signature.
926 2013-10-29 David Malcolm <dmalcolm@redhat.com>
928 * doc/gty.texi (GTY Options): Add note about inheritance to
929 description of desc and tag.
930 (Inheritance and GTY): New.
932 2013-10-29 David Malcolm <dmalcolm@redhat.com>
934 * gengtype-parse.c (opts_have): Drop "static" so that
935 we can use this from gengtype.c.
936 * gengtype.c (set_gc_used_type): Mark any base class as used;
937 update field traversal to visit inherited fields.
938 (output_mangled_typename): Convert references to classes within
939 an inheritance hierarchy to reference the ultimate base class,
940 since only it will have gt_ functions.
941 (get_string_option): New.
942 (walk_subclasses): New.
943 (walk_type): Treat GTY structs that have a "desc" as being the
944 root of an inheritance hierarchy. Generate a switch on it
945 within the marking function which walks all subclasses, adding
946 cases for them via walk_subclasses. For subclasses, visit all
947 fields of the type (including inherited ones).
948 (write_func_for_structure): Don't write fns for subclasses, only
949 for the ultimate base class within an inheritance hierarchy.
950 Subclasses-marking will be handled by the base class marking
952 (write_types): Likewise.
953 (write_local_func_for_structure): Likewise.
954 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
955 a "tag" option (and are thus concrete subclasses).
956 (write_root): Use the marker function for the ultimate base class.
957 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
958 (opts_have): Add declaration.
960 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
962 * lra-spills.c (lra_final_code_change): Remove useless move insns
963 originated from moves of pseudos.
965 2013-10-28 Jeff Law <law@redhat.com>
967 * lower-subreg.c (resolve_simple_move): Fix comment typo.
969 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
971 * df-scan.c (df_collection_rec): Adjust.
972 (copy_defs): New constant.
973 (copy_uses): Likewise.
974 (copy_eq_uses): Likewise.
976 (copy_all): Likewise.
977 (df_insn_rescan): Adjust.
978 (df_notes_rescan): Likewise.
979 (df_swap_refs): Likewise.
980 (df_sort_and_compress_refs): Likewise.
981 (df_sort_and_compress_mws): Likewise.
982 (df_install_refs): Likewise.
983 (df_install_mws): Likewise.
984 (df_refs_add_to_chains): Add flags parameter controlling which vectors
986 (df_bb_refs_record): Adjust.
987 (df_record_entry_block_defs): Likewise.
988 (df_record_exit_block_defs): Likewise.
989 (df_refs_verify): Likewise.
990 (df_mws_verify): Likewise.
991 (df_insn_refs_verify): Likewise.
992 (df_bb_verify): Likewise.
993 * ipa-pure-const.c (finish_state): Remove.
995 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
996 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
997 var-tracking.c: Adjust.
998 * vec.c (stack_vecs): Remove.
999 (register_stack_vec): Likewise.
1000 (stack_vec_register_index): Likewise.
1001 (unregister_stack_vec): Likewise.
1002 * vec.h (struct va_stack): Remove.
1003 (struct vec<T, A, vl_ptr>): Specialize as
1004 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1005 allocation strategy compatable with the vl_ptr layout.
1006 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1007 specialization anyway.
1008 (class stack_vec): New class.
1009 (vec_stack_alloc): Remove.
1010 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1012 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1013 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1014 Sergey Lega <sergey.s.lega@intel.com>
1015 Anna Tikhonova <anna.tikhonova@intel.com>
1016 Ilya Tocar <ilya.tocar@intel.com>
1017 Andrey Turetskiy <andrey.turetskiy@intel.com>
1018 Ilya Verbin <ilya.verbin@intel.com>
1019 Kirill Yukhin <kirill.yukhin@intel.com>
1020 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1022 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1023 (*prefetch_avx512pf_<mode>): New.
1024 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1025 (avx512f_maskcmp<mode>3): Ditto.
1026 (vashrv16si3): Ditto.
1028 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1029 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1030 Sergey Lega <sergey.s.lega@intel.com>
1031 Anna Tikhonova <anna.tikhonova@intel.com>
1032 Ilya Tocar <ilya.tocar@intel.com>
1033 Andrey Turetskiy <andrey.turetskiy@intel.com>
1034 Ilya Verbin <ilya.verbin@intel.com>
1035 Kirill Yukhin <kirill.yukhin@intel.com>
1036 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1038 * config/i386/i386.md (any_truncate): New.
1039 (trunsuffix): Ditto.
1040 * config/i386/predicates.md (const_8_to_9_operand): New.
1041 (const_10_to_11_operand): Ditto.
1042 (const_12_to_13_operand): Ditto.
1043 (const_14_to_15_operand): Ditto.
1044 (const_16_to_19_operand): Ditto.
1045 (const_20_to_23_operand): Ditto.
1046 (const_24_to_27_operand): Ditto.
1047 (const_28_to_31_operand): Ditto.
1048 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1049 (cvtusi2<ssescalarmodesuffix>32): New.
1050 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1051 (ufloatv16siv16sf2): Ditto.
1052 (avx512f_fix_notruncv16sfv16si): Ditto.
1053 (avx512f_ufix_notruncv16sfv16si): Ditto.
1054 (avx512f_vcvtss2usi): Ditto.
1055 (avx512f_vcvtss2usiq): Ditto.
1056 (avx512f_vcvttss2usi): Ditto.
1057 (avx512f_vcvttss2usiq): Ditto.
1058 (avx512f_vcvtsd2usi): Ditto.
1059 (avx512f_vcvtsd2usiq): Ditto.
1060 (avx512f_vcvttsd2usi): Ditto.
1061 (avx512f_vcvttsd2usiq): Ditto.
1062 (ufloatv8siv8df): Ditto.
1063 (avx512f_cvtdq2pd512_2): Ditto.
1064 (avx512f_cvtpd2dq512): Ditto.
1065 (avx512f_ufix_notruncv8dfv8si): Ditto.
1066 (avx512f_cvtpd2ps512): Ditto.
1067 (vec_unpacks_lo_v16sf): Ditto.
1068 (vec_unpacks_hi_v16sf): Ditto.
1069 (vec_unpacks_float_hi_v16si): Ditto.
1070 (vec_unpacks_float_lo_v16si): Ditto.
1071 (avx512f_unpckhps512): Ditto.
1072 (avx512f_unpcklps512): Ditto.
1073 (avx512f_movshdup512): Ditto.
1074 (avx512f_movsldup512): Ditto.
1075 (vec_extract_lo_v32hi): Ditto.
1076 (vec_extract_hi_v32hi): Ditto.
1077 (vec_extract_lo_v64qi): Ditto.
1078 (vec_extract_hi_v64qi): Ditto.
1079 (avx512f_unpckhpd512): Ditto.
1080 (avx512f_movddup512): Ditto.
1081 (avx512f_unpcklpd512): Ditto.
1082 (*avx512f_unpcklpd512): Ditto.
1083 (avx512f_shufps512_1): Ditto.
1084 (avx512f_shufpd512_1): Ditto.
1085 (avx512f_interleave_highv8di): Ditto.
1086 (avx512f_interleave_lowv8di): Ditto.
1087 (PMOV_DST_MODE): Ditto.
1088 (pmov_src_mode): Ditto.
1089 (pmov_src_lower): Ditto.
1091 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1092 (*avx512f_<code>v8div16qi2): Ditto.
1093 (*avx512f_<code>v8div16qi2_store): Ditto.
1094 (vec_widen_umult_even_v16si): Ditto.
1095 (*vec_widen_umult_even_v16si): Ditto.
1096 (vec_widen_smult_even_v16si): Ditto.
1097 (*vec_widen_smult_even_v16si): Ditto.
1098 (avx512f_interleave_highv16si): Ditto.
1099 (avx512f_interleave_lowv16si): Ditto.
1100 (avx512f_<code>v16qiv16si2): Ditto.
1101 (avx512f_<code>v16hiv16si2): Ditto.
1102 (avx512f_<code>v8qiv8di2): Ditto.
1103 (avx512f_<code>v8hiv8di2): Ditto.
1104 (avx512f_<code>v8siv8di2): Ditto.
1105 (avx512cd_maskb_vec_dupv8di): Ditto.
1106 (avx512cd_maskw_vec_dupv16si): Ditto.
1107 (avx512f_vcvtph2ps512): Ditto.
1108 (avx512f_vcvtps2ph512): Ditto.
1109 (VEC_EXTRACT_MODE): Extened with wider modes.
1110 (VEC_PERM_AVX2): Ditto.
1111 (VEC_PERM_CONST): Ditto.
1114 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1116 * config/arc/arc.c (arc_ccfsm_post_advance):
1117 Add comment about TYPE_RETURN.
1119 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1121 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1122 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1123 (strip_offset): Convert offset to unsigned number.
1125 2013-10-27 Tom de Vries <tom@codesourcery.com>
1127 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1128 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1130 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1132 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1133 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1134 (TARGET_BIG_ENDIAN): New macro.
1135 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1136 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1137 * config/sh/sh.c: Likewise.
1138 * config/sh/sh.h: Likewise.
1140 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1142 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1143 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1145 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1147 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1149 (cmpstr_t, cmpstrsi): Fix formatting.
1151 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1154 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1155 addressing, do not use general_operand for memory operands.
1157 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1160 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1161 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1163 2013-10-26 Jeff Law <law@redhat.com>
1165 * predict.c (PRED_MUDFLAP): Remove.
1166 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1168 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1169 (OBJS): Remove tree-nomudflap.o
1170 (GTFILES): Remove tree-mudflap.c
1171 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1172 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1173 (mfwrap_spec, mflib_spec): Likewise.
1174 (cpp_unique_options, cc1_options, static_specs): Likewise.
1175 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1176 * passes.def: Likewise.
1177 * toplev.c (compile_file, process_options): Likewise.
1178 * tree-inline.c (copy_tree_r): Likewise.
1179 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
1180 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
1181 (build_constant_desc, output_constant_def_contents): Likewise.
1182 (categorize_decl_for_section): Likewise.
1183 * tree-mudflap.c: Removed.
1184 * tree-mudflap.h: Removed.
1185 * tree-nomudflap.c: Removed.
1186 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
1187 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
1188 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1189 * config/sol2.h (MFLIB_SPEC): Likewise.
1190 * doc/install.texi: Remove mudflap references.
1191 * doc/passes.texi: Similarly.
1192 * doc/sourcebuild.texi: Similarly.
1193 * doc/invoke.texi: Remove mudlfap related options.
1195 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1197 PR rtl-optimization/58759
1198 * lra-constraints.c (lra_constraints): Remove wrong condition to
1199 remove insn setting up an equivalent pseudo.
1201 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1203 * config/rs6000/rs6000-protos.h
1204 (rs6000_secondary_memory_needed_mode): New prototype.
1205 * config/rs6000/rs6000.c: Include ira.h.
1206 (TARGET_LRA_P): Redefine.
1207 (rs6000_legitimate_offset_address_p): Call
1208 legitimate_constant_pool_address_p in strict mode for LRA.
1209 (rs6000_legitimate_address_p): Ditto.
1210 (legitimate_lo_sum_address_p): Add code for LRA.
1211 Use lra_in_progress.
1212 (rs6000_emit_move): Add LRA version of code to generate load/store
1214 (rs6000_secondary_memory_needed_mode): New.
1215 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
1216 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
1217 constants, memory, and FP registers.
1218 (rs6000_lra_p): New.
1219 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New
1221 * config/rs6000/rs6000.opt (mlra): New option.
1222 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1224 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
1226 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
1227 has_single_use () and not do the conversion if has_single_use ()
1228 returns false for the multiplication result.
1230 2013-10-25 David Malcolm <dmalcolm@redhat.com>
1232 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
1233 to "NODE", since this works on a "tree", not an
1235 (CONSTANT_CLASS_P): Likewise.
1238 (INDIRECT_REF_P): Likewise.
1239 (REFERENCE_CLASS_P): Likewise.
1240 (COMPARISON_CLASS_P): Likewise.
1241 (UNARY_CLASS_P): Likewise.
1242 (BINARY_CLASS_P): Likewise.
1243 (STATEMENT_CLASS_P): Likewise.
1244 (VL_EXP_CLASS_P): Likewise.
1245 (EXPRESSION_CLASS_P): Likewise.
1246 (IS_TYPE_OR_DECL_P): Likewise.
1248 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
1250 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
1251 ADDR_EXPR in the defining statement.
1253 2013-10-25 Richard Biener <rguenther@suse.de>
1255 PR tree-optimization/58626
1256 * tree-loop-distribution.c (enum rdg_dep_type): Remove
1257 anti_dd, output_dd and input_dd.
1258 (struct rdg_edge): Remove level and relation members.
1259 (RDGE_LEVEL, RDGE_RELATION): Remove.
1260 (dot_rdg_1): Adjust.
1261 (create_rdg_edge_for_ddr): Remove.
1262 (create_rdg_edges_for_scalar): Adjust.
1263 (create_edge_for_control_dependence): Likewise.
1264 (create_rdg_edges): Split into ...
1265 (create_rdg_flow_edges): ... this
1266 (create_rdg_cd_edges): ... and this.
1268 (build_rdg): Likewise, do not compute data dependences or
1270 (pg_add_dependence_edges): New function.
1272 (distribute_loop): First apply all non-dependence based
1273 partition mergings. Then compute dependences between partitions
1274 and merge and order partitions according to them.
1276 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1278 PR rtl-optimization/58831
1279 * alias.c (init_alias_analysis): At the beginning of each iteration, set
1280 the reg_seen[N] bit if static_reg_base_value[N] is non-null.
1282 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1284 * recog.c (search_ofs): New static variable moved from...
1285 (peep2_find_free_register): ...here.
1286 (peephole2_optimize): Initialize it.
1288 2013-10-25 Tobias Burnus <burnus@net-b.de>
1290 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
1292 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
1294 * config/i386/i386.h (TARGET_MPX): New define.
1295 (TARGET_MPX_P): Ditto.
1297 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1299 * config/i386/constraints.md (B): New.
1302 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
1303 * config/i386/i386-modes.def (BND32): New.
1305 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
1306 * config/i386/i386.c (isa_opts): Add mmpx.
1307 (regclass_map): Add bound registers.
1308 (dbx_register_map): Likewise.
1309 (dbx64_register_map): Likewise.
1310 (svr4_dbx_register_map): Likewise.
1312 (ix86_option_override_internal): Support MPX ISA.
1313 (ix86_conditional_register_usage): Support bound registers.
1314 (print_reg): Likewise.
1315 (ix86_code_end): Add MPX bnd prefix.
1316 (output_set_got): Likewise.
1317 (ix86_output_call_insn): Likewise.
1318 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
1319 (ix86_print_operand_punct_valid_p): Likewise.
1320 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
1322 (ix86_class_likely_spilled_p): Add bound regs support.
1323 (ix86_hard_regno_mode_ok): Likewise.
1324 (x86_order_regs_for_local_alloc): Likewise.
1325 (ix86_bnd_prefixed_insn_p): New.
1326 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
1327 (FIXED_REGISTERS): Add bound registers.
1328 (CALL_USED_REGISTERS): Likewise.
1329 (REG_ALLOC_ORDER): Likewise.
1330 (HARD_REGNO_NREGS): Likewise.
1332 (VALID_BND_REG_MODE): New.
1333 (FIRST_BND_REG): New.
1334 (LAST_BND_REG): New.
1335 (reg_class): Add BND_REGS.
1336 (REG_CLASS_NAMES): Likewise.
1337 (REG_CLASS_CONTENTS): Likewise.
1339 (ANY_BND_REG_P): New.
1341 (HI_REGISTER_NAMES): Add bound registers.
1342 * config/i386/i386.md (UNSPEC_BNDMK): New.
1343 (UNSPEC_BNDMK_ADDR): New.
1344 (UNSPEC_BNDSTX): New.
1345 (UNSPEC_BNDLDX): New.
1346 (UNSPEC_BNDLDX_ADDR): New.
1347 (UNSPEC_BNDCL): New.
1348 (UNSPEC_BNDCU): New.
1349 (UNSPEC_BNDCN): New.
1350 (UNSPEC_MPX_FENCE): New.
1353 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
1354 (length_immediate): Likewise.
1355 (prefix_0f): Likewise.
1357 (prefix_rep): Check for bnd prefix.
1358 (length_nobnd): New.
1359 (length): Use length_nobnd if specified.
1364 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
1367 (simple_return_internal): Likewise.
1368 (simple_return_pop_internal): Likewise.
1369 (*indirect_jump): Add MPX bnd prefix.
1370 (*tablejump_1): Likewise.
1371 (simple_return_internal_long): Likewise.
1372 (simple_return_indirect_internal): Likewise.
1376 (*mov<mode>_internal_mpx): New.
1377 (<mode>_<bndcheck>): New.
1378 (*<mode>_<bndcheck>): New.
1383 * config/i386/predicates.md (lea_address_operand): Rename to...
1384 (address_no_seg_operand): ... this.
1385 (address_mpx_no_base_operand): New.
1386 (address_mpx_no_index_operand): New.
1387 (bnd_mem_operator): New.
1388 * config/i386/i386.opt (mmpx): New.
1389 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
1390 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
1392 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1394 * mode-classes.def (MODE_POINTER_BOUNDS): New.
1395 * tree.def (POINTER_BOUNDS_TYPE): New.
1396 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
1397 (POINTER_BOUNDS_MODE): New.
1398 (make_pointer_bounds_mode): New.
1399 * machmode.h (POINTER_BOUNDS_MODE_P): New.
1400 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
1401 (layout_type): Support POINTER_BOUNDS_TYPE.
1402 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
1403 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
1404 (type_contains_placeholder_1): Likewise.
1405 * tree.h (POINTER_BOUNDS_TYPE_P): New.
1406 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
1407 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
1409 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
1411 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
1412 mode of address computation when calling memory_address_addr_space.
1414 2013-08-24 Richard Henderson <rth@twiddle.net>
1417 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
1418 instead of create_convert_operand_to.
1419 (maybe_emit_sync_lock_test_and_set): Likewise.
1420 (expand_atomic_compare_and_swap): Likewise.
1421 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
1423 2013-08-24 Sriraman Tallam <tmsriram@google.com>
1425 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
1426 Change param_is to use the struct and not the pointer to the struct.
1428 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
1430 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
1431 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
1432 Move to targhooks.c.
1433 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
1434 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
1435 Relocate from gimplify.c.
1436 * targhooks.h: Add 2 prototypes.
1437 * tree.h: Delete 2 prototypes.
1439 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
1441 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
1442 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
1444 2013-10-24 Cong Hou <congh@google.com>
1446 * convert.c (convert_to_real): Guard those unsafe math function
1447 convertions with flag_unsafe_math_optimizations. Handle sqrt()
1450 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
1453 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
1455 (cgraph_create_edge): Use the new argument.
1456 (cgraph_create_indirect_edge): Likewise.
1458 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
1460 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
1462 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
1463 changes statep->state.
1465 2013-10-24 Tobias Burnus <burnus@net-b.de>
1468 * tree-cfg.c (replace_loop_annotate): New function.
1469 (execute_build_cfg): Call it.
1470 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
1471 * internal-fn.c (expand_ANNOTATE): New function.
1472 * internal-fn.def (ANNOTATE): Define as new internal function.
1473 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
1474 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
1475 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
1476 * doc/extend.texi (Pragmas): Document #pragma ivdep.
1477 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
1479 2013-10-17 Ian Bolton <ian.bolton@arm.com>
1480 Marcus Shawcroft <marcus.shawcroft@arm.com>
1482 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
1483 Special case reload SP+C into none GENERAL_REGS.
1485 2013-10-24 Michael Matz <matz@suse.de>
1487 * gengtype.c (is_file_equal): Check that files will be same length.
1489 2013-10-25 Christian Bruel <christian.bruel@st.com>
1491 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
1492 * gcc/config/sh/t-sh (sh-mem.o): New rule.
1493 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
1494 (sh_expand_cmpstr): New function.
1495 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
1496 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
1497 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
1498 (rotlhi3_8): Rename.
1500 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1502 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
1503 * Makefile.in (CXXDEPMODE): Assign and change users.
1504 (CCDEPMODE): Delete.
1505 * configure: Regenerate.
1507 2013-10-23 David Malcolm <dmalcolm@redhat.com>
1509 * gengtype-parse.c (require_without_advance): New.
1510 (type): For GTY-marked types that are not GTY((user)), parse any
1511 base classes, requiring them to be single-inheritance, and not
1512 be templates. For non-GTY-marked types and GTY((user)),
1513 continue to skip over any C++ inheritance specification.
1514 * gengtype-state.c (state_writer::write_state_struct_type):
1515 Write base class of type (if any).
1516 (read_state_struct_type): Read base class of type (if any).
1517 * gengtype.c (new_structure): Add a "base_class" parameter.
1518 (create_optional_field_): Update for new parameter to new_structure.
1519 (adjust_field_rtx_def): Likewise.
1520 (adjust_field_tree_exp): Likewise.
1521 * gengtype.h (struct type): Add "base_class" field to the s
1523 (new_structure): Add "base" parameter.
1525 2013-10-23 Sriraman Tallam <tmsriram@google.com>
1528 * config/i386/i386.c (ix86_option_override_internal):
1529 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
1530 (ix86_valid_target_attribute_tree):
1531 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
1532 Change TARGET_SSE to TARGET_SSE_P (opts->...)
1534 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
1536 * tree-ssa-loop.h: Remove include files.
1537 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
1538 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
1539 * graphite-clast-to-gimple.c: Likewise.
1540 * graphite-scop-detection.c: Likewise.
1541 * graphite-sese-to-poly.c: Likewise.
1542 * ipa-inline-analysis.c: Likewise.
1543 * ipa-pure-const.c: Likewise.
1544 * loop-init.c: Likewise.
1545 * passes.c: Likewise.
1546 * predict.c: Likewise.
1547 * tree-cfg.c: Likewise.
1548 * tree-cfgcleanup.c: Likewise.
1549 * tree-chrec.c: Likewise.
1550 * tree-data-ref.c: Likewise.
1551 * tree-loop-distribution.c: Likewise.
1552 * tree-parloops.c: Likewise.
1553 * tree-predcom.c: Likewise.
1554 * tree-scalar-evolution.c: Likewise.
1555 * tree-ssa-address.c: Likewise.
1556 * tree-ssa.c: Likewise.
1557 * tree-ssa-dce.c: Likewise.
1558 * tree-ssa-loop.c: Likewise.
1559 * tree-ssa-loop-im.c: Likewise.
1560 * tree-ssa-loop-ivcanon.c: Likewise.
1561 * tree-ssa-loop-ivopts.c: Likewise.
1562 * tree-ssa-loop-manip.c: Likewise.
1563 * tree-ssa-loop-niter.c: Likewise.
1564 * tree-ssa-loop-prefetch.c: Likewise.
1565 * tree-ssa-loop-unswitch.c: Likewise.
1566 * tree-ssa-reassoc.c: Likewise.
1567 * tree-vect-data-refs.c: Likewise.
1568 * tree-vect-loop.c: Likewise.
1569 * tree-vect-loop-manip.c: Likewise.
1570 * tree-vectorizer.c: Likewise.
1571 * tree-vect-stmts.c: Likewise.
1572 * tree-vrp.c: Likewise.
1574 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1576 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
1578 2013-10-23 Jakub Jelinek <jakub@redhat.com>
1580 PR tree-optimization/58775
1581 PR tree-optimization/58791
1582 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
1583 (insert_stmt_after): Rewritten, don't move the stmt, but really
1585 (get_stmt_uid_with_default): Remove.
1586 (build_and_add_sum): Use insert_stmt_after and
1587 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
1588 (update_range_test): Set uid on stmts added by
1589 force_gimple_operand_gsi. Don't immediately modify statements
1590 in inter-bb optimization, just update oe->op values.
1591 (optimize_range_tests): Return bool whether any changed have been made.
1592 (update_ops): New function.
1593 (struct inter_bb_range_test_entry): New type.
1594 (maybe_optimize_range_tests): Perform statement changes here.
1595 (not_dominated_by, appears_later_in_bb, get_def_stmt,
1596 ensure_ops_are_available): Remove.
1597 (find_insert_point): Rewritten.
1598 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
1599 return LHS of the (new resp. old) stmt. Don't call
1600 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
1601 instead of last, move new stmt at the right place.
1602 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
1603 (negate_value): Likewise. Set uids.
1604 (break_up_subtract_bb): Initialize uids.
1605 (reassociate_bb): Adjust rewrite_expr_tree caller.
1606 (do_reassoc): Don't call renumber_gimple_stmt_uids.
1608 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
1611 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
1612 TARGET_32BIT final condition.
1613 (mulsi3_internal2 and splitter): Same.
1615 2013-10-23 Jeff Law <law@redhat.com>
1617 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
1618 through joiner blocks with abnormal outgoing edges.
1620 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
1621 Add parameter JOINERS, to allow/disallow threading through joiner
1623 (thread_block): New. Call thread_block_1.
1624 (mark_threaded_blocks): Remove code to filter out certain cases
1625 of threading through joiner blocks.
1626 (thread_through_all_blocks): Document how we can have a dangling
1627 edge AUX field and clear it.
1629 2013-10-23 Ian Lance Taylor <iant@google.com>
1631 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
1632 (C++ Dialect Options): Likewise.
1633 (Optimize Options): Likewise.
1635 2013-10-23 Tom de Vries <tom@codesourcery.com>
1637 PR tree-optimization/58805
1638 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
1640 2013-10-23 Jakub Jelinek <jakub@redhat.com>
1642 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
1643 sequence based on get_range_info returned range.
1645 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
1647 * tree-ssa.h: Remove all #include's
1648 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
1649 * alias.c: Move required includes from tree-ssa.h.
1651 * builtins.c: Likewise.
1652 * calls.c: Likewise.
1653 * cfgexpand.c: Likewise.
1654 * cfghooks.c: Likewise.
1655 * cfgloop.c: Likewise.
1656 * cfgloopmanip.c: Likewise.
1657 * cgraph.c: Likewise.
1658 * cgraphbuild.c: Likewise.
1659 * cgraphclones.c: Likewise.
1660 * cgraphunit.c: Likewise.
1662 * except.c: Likewise.
1664 * final.c: Likewise.
1665 * fold-const.c: Likewise.
1666 * ggc-page.c: Likewise.
1667 * gimple-builder.c: Likewise.
1668 * gimple-fold.c: Likewise.
1669 * gimple-iterator.c: Likewise.
1670 * gimple-low.c: Likewise.
1671 * gimple-pretty-print.c: Likewise.
1672 * gimple-ssa-strength-reduction.c: Likewise.
1673 * gimple-streamer-in.c: Likewise.
1674 * gimple-streamer-out.c: Likewise.
1675 * gimplify.c: Likewise.
1676 * graphite-blocking.c: Likewise.
1677 * graphite-clast-to-gimple.c: Likewise.
1678 * graphite-dependences.c: Likewise.
1679 * graphite-interchange.c: Likewise.
1680 * graphite-optimize-isl.c: Likewise.
1681 * graphite-poly.c: Likewise.
1682 * graphite-scop-detection.c: Likewise.
1683 * graphite-sese-to-poly.c: Likewise.
1684 * graphite.c: Likewise.
1685 * ipa-cp.c: Likewise.
1686 * ipa-inline-analysis.c: Likewise.
1687 * ipa-inline-transform.c: Likewise.
1688 * ipa-inline.c: Likewise.
1689 * ipa-prop.c: Likewise.
1690 * ipa-pure-const.c: Likewise.
1691 * ipa-reference.c: Likewise.
1692 * ipa-split.c: Likewise.
1693 * ipa-utils.c: Likewise.
1694 * loop-init.c: Likewise.
1695 * lto-cgraph.c: Likewise.
1696 * lto-section-in.c: Likewise.
1697 * lto-section-out.c: Likewise.
1698 * lto-streamer-in.c: Likewise.
1699 * lto-streamer-out.c: Likewise.
1700 * lto-streamer.c: Likewise.
1701 * omp-low.c: Likewise.
1702 * passes.c: Likewise.
1703 * predict.c: Likewise.
1704 * print-tree.c: Likewise.
1705 * profile.c: Likewise.
1707 * targhooks.c: Likewise.
1708 * tracer.c: Likewise.
1709 * trans-mem.c: Likewise.
1710 * tree-call-cdce.c: Likewise.
1711 * tree-cfg.c: Likewise.
1712 * tree-cfgcleanup.c: Likewise.
1713 * tree-chrec.c: Likewise.
1714 * tree-complex.c: Likewise.
1715 * tree-data-ref.c: Likewise.
1716 * tree-dfa.c: Likewise.
1717 * tree-eh.c: Likewise.
1718 * tree-emutls.c: Likewise.
1719 * tree-if-conv.c: Likewise.
1720 * tree-inline.c: Likewise.
1721 * tree-into-ssa.c: Likewise.
1722 * tree-loop-distribution.c: Likewise.
1723 * tree-mudflap.c: Likewise.
1724 * tree-nested.c: Likewise.
1725 * tree-nrv.c: Likewise.
1726 * tree-object-size.c: Likewise.
1727 * tree-outof-ssa.c: Likewise.
1728 * tree-parloops.c: Likewise.
1729 * tree-phinodes.c: Likewise.
1730 * tree-predcom.c: Likewise.
1731 * tree-pretty-print.c: Likewise.
1732 * tree-profile.c: Likewise.
1733 * tree-scalar-evolution.c: Likewise.
1734 * tree-sra.c: Likewise.
1735 * tree-ssa-address.c: Likewise.
1736 * tree-ssa-alias.c: Likewise.
1737 * tree-ssa-ccp.c: Likewise.
1738 * tree-ssa-coalesce.c: Likewise.
1739 * tree-ssa-copy.c: Likewise.
1740 * tree-ssa-copyrename.c: Likewise.
1741 * tree-ssa-dce.c: Likewise.
1742 * tree-ssa-dom.c: Likewise.
1743 * tree-ssa-dse.c: Likewise.
1744 * tree-ssa-forwprop.c: Likewise.
1745 * tree-ssa-ifcombine.c: Likewise.
1746 * tree-ssa-live.c: Likewise.
1747 * tree-ssa-loop-ch.c: Likewise.
1748 * tree-ssa-loop-im.c: Likewise.
1749 * tree-ssa-loop-ivcanon.c: Likewise.
1750 * tree-ssa-loop-ivopts.c: Likewise.
1751 * tree-ssa-loop-manip.c: Likewise.
1752 * tree-ssa-loop-niter.c: Likewise.
1753 * tree-ssa-loop-prefetch.c: Likewise.
1754 * tree-ssa-loop-unswitch.c: Likewise.
1755 * tree-ssa-loop.c: Likewise.
1756 * tree-ssa-math-opts.c: Likewise.
1757 * tree-ssa-operands.c: Likewise.
1758 * tree-ssa-phiopt.c: Likewise.
1759 * tree-ssa-phiprop.c: Likewise.
1760 * tree-ssa-pre.c: Likewise.
1761 * tree-ssa-propagate.c: Likewise.
1762 * tree-ssa-reassoc.c: Likewise.
1763 * tree-ssa-sccvn.c: Likewise.
1764 * tree-ssa-sink.c: Likewise.
1765 * tree-ssa-strlen.c: Likewise.
1766 * tree-ssa-structalias.c: Likewise.
1767 * tree-ssa-tail-merge.c: Likewise.
1768 * tree-ssa-ter.c: Likewise.
1769 * tree-ssa-threadedge.c: Likewise.
1770 * tree-ssa-threadupdate.c: Likewise.
1771 * tree-ssa-uncprop.c: Likewise.
1772 * tree-ssa-uninit.c: Likewise.
1773 * tree-ssa.c: Likewise.
1774 * tree-ssanames.c: Likewise.
1775 * tree-stdarg.c: Likewise.
1776 * tree-streamer-in.c: Likewise.
1777 * tree-switch-conversion.c: Likewise.
1778 * tree-tailcall.c: Likewise.
1779 * tree-vect-data-refs.c: Likewise.
1780 * tree-vect-generic.c: Likewise.
1781 * tree-vect-loop-manip.c: Likewise.
1782 * tree-vect-loop.c: Likewise.
1783 * tree-vect-patterns.c: Likewise.
1784 * tree-vect-slp.c: Likewise.
1785 * tree-vect-stmts.c: Likewise.
1786 * tree-vectorizer.c: Likewise.
1787 * tree-vrp.c: Likewise.
1790 * value-prof.c: Likewise.
1791 * var-tracking.c: Likewise.
1792 * varpool.c: Likewise.
1793 * vtable-verify.c: Likewise.
1795 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1797 * config/tilegx/tilegx.c: Include "tree.h".
1799 2013-10-23 Jakub Jelinek <jakub@redhat.com>
1801 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
1802 the info, not after it.
1803 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
1804 dump_ssaname_info call.
1805 (pp_gimple_stmt_1): Adjust caller.
1806 (dump_phi_nodes): Likewise. Don't print "# " here.
1808 2013-10-22 Jan Hubicka <jh@suse.cz>
1810 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
1812 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
1813 * i386.c (expand_small_movmem_or_setmem): New function.
1814 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
1816 (alg_usable_p): Add support for value ranges; cleanup.
1817 (ix86_expand_set_or_movmem): Add support for misaligned moves.
1819 2013-10-22 Sterling Augustine <saugustine@google.com>
1821 * doc/invoke.texi: Document -ggnu-pubnames.
1822 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
1824 * dwarf2out.c: Include gdb/gdb-index.h.
1825 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
1826 debug_generate_pub_sections.
1827 (is_java, output_pubtables, output_pubname): New functions.
1828 (include_pubname_in_output): Handle debug_generate_pub_sections at
1830 (size_of_pubnames): Use new local space_for_flags based on
1831 debug_generate_pub_sections.
1832 (output_pubnames): Unify pubnames and pubtypes output logic.
1833 Genericize comments. Call output_pubname.
1834 (dwarf2out_finish): Move logic to output_pubnames and call it.
1836 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
1839 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
1840 Remove CCCmode handling.
1841 <case LTU>: Return 'c' suffix for CCCmode.
1842 <case GEU>: Return 'nc' suffix for CCCmode.
1843 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
1844 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
1845 (*sub<mode>3_cc_overflow): Ditto.
1846 (*subsi3_zext_cc_overflow): Ditto.
1848 2013-10-22 Steve Ellcey <sellcey@mips.com>
1850 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
1851 (AND (NOT OP1) (NOT OP2)).
1853 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
1855 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
1856 meaning of merge-high and merge-low masks for little endian; avoid
1857 use of vector-pack masks for little endian for mismatched modes.
1859 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1861 * config/tilepro/tilepro.c: Include "tree.h".
1863 2013-10-22 Andreas Schwab <schwab@suse.de>
1865 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
1868 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
1870 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
1871 [Ed Smith-Rowland]): New entries.
1872 ([Stephen M. Webb]): Update.
1874 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
1876 * tree-ssa-ter.h: Remove duplicate copy of file contents.
1878 2013-10-21 Marek Polacek <polacek@redhat.com>
1881 * fold-const.c (fold_range_test): Return 0 if the type is not
1884 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
1886 * system.h: Move hwint.h include further down.
1887 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
1888 gcc_checking_asserts.
1889 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
1891 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
1893 Fix volatile issues in optimize_bit_field_compare.
1894 * fold-const.c (optimize_bit_field_compare): Bail out if
1895 lvolatilep or rvolatilep.
1897 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
1899 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
1900 and get_inner_reference returning different pmode for non-volatile
1901 bit-field members dependent on flag_strict_volatile_bitfields.
1902 * stor-layout.c (layout_decl): Remove special handling of
1903 flag_strict_volatile_bitfields.
1904 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
1905 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
1907 2013-10-21 Paulo Matos <pmatos@broadcom.com>
1909 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
1912 2013-10-21 Jeff Law <law@redhat.com>
1914 * tree-ssa-threadedge.c (thread_through_normal_block): New
1915 argument VISITED. Remove VISISTED as a local variable. When we
1916 have a threadable jump, verify the destination of the jump has not
1918 (thread_across_edge): Allocate VISITED bitmap once at function
1919 scope and use it throughout. Make sure to set appropriate bits in
1920 VISITED for E (start of jump thread path).
1921 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
1922 through a joiner if any edge on the path has a recorded jump thread.
1924 2013-10-21 Ian Lance Taylor <iant@google.com>
1926 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
1927 don't imply that attributes can solve all problems.
1928 (Directory Options): Fix typo.
1930 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1932 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
1933 extend and extend_add.
1935 2013-10-21 Richard Biener <rguenther@suse.de>
1937 PR tree-optimization/58794
1938 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
1939 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
1941 2013-10-21 Richard Biener <rguenther@suse.de>
1944 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
1945 to (T) X for sign-changing conversions (or no conversion).
1947 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
1949 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
1951 2013-10-20 Jan Hubicka <jh@suse.cz>
1953 * config/i386/i386-tune.def: Add comment; organize into categories
1955 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
1957 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
1958 argument. Update function comment.
1959 (expand_set_or_movmem_via_rep): New function combining
1960 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
1961 (expand_movmem_via_rep_mov): Remove.
1962 expand_setmem_via_rep_stos): Remove.
1963 (expand_movmem_epilogue): Update calls correspondingly.
1964 (expand_setmem_epilogue_via_loop): Likewise.
1966 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
1967 (expand_set_or_movmem_prologue): New function combining
1968 expand_movmem_prologue and expand_setmem_prologue.
1969 (expand_movmem_prologue): Remove.
1970 (expand_setmem_prologue): Remove.
1971 (expand_set_or_movmem_constant_prologue): New function combining
1972 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
1973 (expand_constant_movmem_prologue): Remove.
1974 (expand_constant_setmem_prologue): Remove.
1975 (promote_duplicated_reg): Allow vector-const0 value.
1976 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
1977 and ix86_expand_setmem.
1978 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
1979 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
1981 2013-10-21 Diego Novillo <dnovillo@google.com>
1983 * asan.c: Include tree.h
1984 * bb-reorder.c: Likewise.
1985 * cfgcleanup.c: Likewise.
1986 * cfgloopmanip.c: Likewise.
1987 * data-streamer-in.c: Likewise.
1988 * data-streamer-out.c: Likewise.
1989 * data-streamer.c: Likewise.
1990 * dwarf2cfi.c: Likewise.
1991 * graphite-blocking.c: Likewise.
1992 * graphite-clast-to-gimple.c: Likewise.
1993 * graphite-dependences.c: Likewise.
1994 * graphite-interchange.c: Likewise.
1995 * graphite-optimize-isl.c: Likewise.
1996 * graphite-poly.c: Likewise.
1997 * graphite-scop-detection.c: Likewise.
1998 * graphite-sese-to-poly.c: Likewise.
1999 * graphite.c: Likewise.
2000 * ipa-devirt.c: Likewise.
2001 * ipa-profile.c: Likewise.
2004 * loop-init.c: Likewise.
2005 * loop-unroll.c: Likewise.
2006 * lower-subreg.c: Likewise.
2007 * lto/lto-object.c: Likewise.
2008 * recog.c: Likewise.
2009 * reginfo.c: Likewise.
2010 * tree-loop-distribution.c: Likewise.
2011 * tree-parloops.c: Likewise.
2012 * tree-ssa-strlen.c: Likewise.
2013 * tree-streamer.c: Likewise.
2014 * value-prof.c: Likewise.
2015 * target-globals.c: Likewise.
2016 * expr.h: Include tree-core.h instead of tree.h.
2017 * gimple.h: Likewise.
2018 * ipa-prop.h: Likewise.
2019 * ipa-utils.h: Likewise.
2020 * lto-streamer.h: Likewise.
2021 * streamer-hooks.h: Likewise.
2022 * ipa-reference.h: Include cgraph.h instead of tree.h.
2023 * cgraph.h: Include basic-block.h instead of tree.h.
2024 * tree-streamer.h: Do not include tree.h.
2025 * genattrtab.c (write_header): Generate inclusion of
2027 * genautomata.c (main): Likewise.
2028 * genemit.c: Likewise.
2029 * genopinit.c: Likewise.
2030 * genoutput.c (output_prologue): Likewise.
2031 * genpeep.c: Likewise.
2033 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2035 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2037 (vec_unpacku_hi_v8hi): Likewise.
2038 (vec_unpacku_lo_v16qi): Likewise.
2039 (vec_unpacku_lo_v8hi): Likewise.
2041 2013-10-20 Jan Hubicka <jh@suse.cz>
2043 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2044 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2045 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2047 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2049 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2050 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2052 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2054 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2057 * system.h: Undef m_slot.
2059 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2061 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2062 all elements for both endian flavors.
2064 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2067 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2068 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2069 and SI_REG for 64bit SYSV ABI targets.
2071 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2073 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2074 to multi_reg_return. Clarify that we are skipping USEs of multiple
2075 return registers. Use bool type where appropriate.
2077 2013-10-18 Jan Hubicka <jh@suse.cz>
2079 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2081 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2082 (X86_TUNE_PUSH_MEMORY): Likewise.
2083 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2084 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2085 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2087 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2088 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2090 (ix86_option_override_internal): Update to use tune features instead
2093 2013-10-18 Cong Hou <congh@google.com>
2095 PR tree-optimization/58508
2096 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2097 statement that contains data refs with zero-step.
2099 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2101 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2102 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2103 * gimple-low.c (gimple_check_call_arg,
2104 gimple_check_call_matching_types): Move to cgraph.c.
2105 * gimple-low.h: Remove prototype.
2106 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2107 Relocate from gimple-low.c.
2108 * cgraph.h: Add prototype. Don't include basic-block.h.
2109 * gimplify.c: Add gimple-low to include list.
2110 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2111 * tree-eh.c: Add gimple-low to include list.
2112 * tree-nested.c: Likewise.
2113 * cfgexpand.c: Add tree-ssa-address.h to include list.
2115 * gimple-fold.c: Likewise.
2116 * gimple-ssa-strength-reduction.c: Likewise.
2117 * trans-mem.c: Likewise.
2118 * tree-mudflap.c: Likewise.
2119 * tree-ssa-loop-ivopts.c: Likewise.
2120 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2121 (degenerate_phi_result): Move to tree-phinodes.c.
2122 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2123 * tree-ssa-threadedge.c: Likewise.
2124 * tree-vrp.c: Likewise.
2125 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2126 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2127 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2128 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2129 * tree-ssa-forwprop.c: Likewise.
2130 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2131 pass_data_cleanup_cfg_post_optimizing,
2132 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2133 * tree-optimize.c: Delete File.
2134 * graphite.c: Include tree-cfgcleanup.h.
2135 * passes.c: Likewise.
2136 * tree-cfg.c: Likewise.
2137 * tree-profile.c: Likewise.
2138 * tree-ssa-dse.c: Likewise.
2139 * tree-ssa-loop-ivcanon.c: Likewise.
2140 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2141 * tree-outof-ssa.c: Include sbitmap.h.
2142 * tree-ssa-live.c: Likewise.
2143 * tree-ssa-propagate.c: Likewise.
2144 * tree-ssa-structalias.c: Likewise.
2145 * tree-stdarg.c: Likewise.
2146 * Makefile.in (OBJS): Delete tree-optimize.o.
2147 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2148 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2149 * varasm.c: Include basic-block.h.
2150 * cfgloop.h: Include function.h instead of basic-block.h
2151 (bb_loop_depth): Move to cfgloop.c.
2152 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2154 2013-10-18 Teresa Johnson <tejohnson@google.com>
2156 * predict.c (probably_never_executed): Compare frequency-based
2157 count to number of training runs.
2158 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2160 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2162 * config/arm/arm.c (cortexa9_extra_costs): New table.
2163 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2165 2013-10-18 Jeff Law <law@redhat.com>
2167 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2169 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2170 (register_jump_thread): Add "registered_jump_thread" debug
2172 * dbgcnt.def (registered_jump_thread): New debug counter.
2174 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2176 * config/rs6000/rs6000.c: Include cgraph.h.
2178 2013-10-18 Teresa Johnson <tejohnson@google.com>
2180 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2181 weights during merging.
2183 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2185 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
2186 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
2187 * tree-ssa-operands.h: Remove prototype.
2188 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
2189 * gimple.c (virtual_operand_p): Relocate from gimple.c.
2190 * gimple.h: Add prototype.
2191 * gimple-ssa.h: Include tree-ssa-operands.h.
2192 * tree-dump.c: Add tree-cfg.h to include list.
2193 * tree-ssa-alias.c: Add ipa-reference.h to include list.
2194 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
2195 * config/i386/i386.c: Don't include tree-flow.h.
2196 * config/rs6000/rs6000.c: Likewise.
2198 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2200 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
2202 2013-10-18 Richard Biener <rguenther@suse.de>
2204 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
2205 or TYPE_UNSIGNED of integral types.
2206 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
2207 NULL_TREE for zero-precision integral types.
2209 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
2211 * config/aarch64/arm_neon.h
2212 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
2214 2013-10-17 Sriraman Tallam <tmsriram@google.com>
2217 * opth-gen.awk: Define target_flags_explicit.
2219 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2221 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
2222 fields to the reg_addr array that describes the valid addressing
2223 mode for any register, general purpose registers, floating point
2224 registers, and Altivec registers.
2225 (FIRST_RELOAD_REG_CLASS): Likewise.
2226 (LAST_RELOAD_REG_CLASS): Likewise.
2227 (struct reload_reg_map_type): Likewise.
2228 (reload_reg_map_type): Likewise.
2229 (RELOAD_REG_VALID): Likewise.
2230 (RELOAD_REG_MULTIPLE): Likewise.
2231 (RELOAD_REG_INDEXED): Likewise.
2232 (RELOAD_REG_OFFSET): Likewise.
2233 (RELOAD_REG_PRE_INCDEC): Likewise.
2234 (RELOAD_REG_PRE_MODIFY): Likewise.
2235 (reg_addr): Likewise.
2236 (mode_supports_pre_incdec_p): New helper functions to say whether
2237 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
2238 (mode_supports_pre_modify_p): Likewise.
2239 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
2240 print the valid address mode bits for each mode.
2241 (rs6000_debug_print_mode): Likewise.
2242 (rs6000_debug_reg_global): Likewise.
2243 (rs6000_setup_reg_addr_masks): New function to set up the address
2244 mask bits for each type.
2245 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
2246 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
2247 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
2248 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
2249 PRE_MODIFY are supported.
2250 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
2251 registers, instead of {src,dest}_av_p.
2252 (rs6000_print_options_internal): Tweak the debug output slightly.
2254 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2256 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
2259 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
2261 * tree-flow.h (struct omp_region): Move to omp-low.c.
2262 Remove omp_ prototypes and variables.
2263 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
2264 (copy_var_decl): Relocate prototype from tree-flow.h.
2265 * gimple.c (copy_var_decl): Relocate from omp-low.c.
2266 * tree.h: Move prototype to omp-low.h.
2267 * omp-low.h: New File. Relocate prototypes here.
2268 * omp-low.c (struct omp_region): Make local here.
2269 (root_omp_region): Make static.
2270 (copy_var_decl) Move to gimple.c.
2271 (new_omp_region): Make static.
2272 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
2273 * tree-cfg.c: Include omp-low.h.
2274 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
2275 * gimplify.c: Include omp-low.h.
2276 * tree-parloops.c: Likewise.
2278 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2280 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
2281 operands are in memory, prefer to force non-matched operand 1 to
2284 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2287 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
2288 restrict TImode addresses to single indirect registers if both
2289 -mquad-memory and -mvsx-timode are used.
2290 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
2291 we should emit load/store quad. Remove using %y for quad memory
2294 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
2295 constraints to allow load/store quad on machines where TImode is
2296 not allowed in VSX registers. Use 'n' instead of 'F' constraint
2297 for TImode to load integer constants.
2299 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2301 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
2303 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
2305 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
2306 handling of STACK_REG.
2308 2013-10-17 Richard Biener <rguenther@suse.de>
2310 PR tree-optimization/58143
2311 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
2313 (rewrite_to_defined_overflow): Likewise.
2314 (move_computations_dom_walker::before_dom): Rewrite stmts
2315 with undefined signed overflow that are not always executed
2316 into unsigned arithmetic.
2318 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2321 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
2322 explicit isa flag to be an options variable, instead of using
2323 global_options_set. Remove define from rs6000.h.
2324 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
2326 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2327 Initialize rs6000_isa_flags_explicit.
2328 (rs6000_function_specific_save): Add gcc_options* parameter, so
2329 that the powerpc builds after the 2013-10-15 changes.
2330 (rs6000_function_specific_restore): Likewise.
2332 2013-10-16 DJ Delorie <dj@redhat.com>
2334 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
2335 op is a REG before checking REGNO.
2336 (rl78_alloc_physical_registers): Verify pattern is a SET before
2339 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2341 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
2343 (vec_unpacks_lo_v4sf): Likewise.
2344 (vec_unpacks_float_hi_v4si): Likewise.
2345 (vec_unpacks_float_lo_v4si): Likewise.
2346 (vec_unpacku_float_hi_v4si): Likewise.
2347 (vec_unpacku_float_lo_v4si): Likewise.
2349 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2351 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
2352 (vsx_concat_v2sf): Likewise.
2354 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
2356 * config/aarch64/aarch64.md
2357 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
2359 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
2361 PR tree-optimization/58697
2362 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
2363 estimated_loop_iterations_int.
2364 (max_stmt_executions_int): Call get_max_loop_iterations_int.
2365 (get_max_loop_iterations_int): New. HWINT version of
2366 get_max_loop_iterations.
2367 * cfgloop.h: Add prototypes.
2368 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
2369 * loop-unroll.c (decide_peel_once_rolling): Call
2370 get_estimated_loop_iterations_int.
2371 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
2372 * tree-ssa-loop-niter.h: Tweak prototypes.
2374 2013-10-16 David Malcolm <dmalcolm@redhat.com>
2376 * gengtype-parse.c (struct_field_seq): Ignore access-control
2377 keywords ("public:" etc).
2379 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
2381 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
2382 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
2384 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
2386 * config/arm/arm.opt (mlra): New option.
2387 * config/arm/arm.c (arm_lra_p): New function.
2388 (TARGET_LRA_P): Define.
2390 2013-10-16 Paulo Matos <pmatos@broadcom.com>
2392 * tree-core.h (tree_code_name): Remove.
2393 * tree.h (get_tree_code_name): New prototype.
2394 * tree.c (tree_code_name): Make static.
2395 (get_tree_code_name): New function.
2396 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
2397 tree_class_check_failed, tree_range_check_failed,
2398 tree_not_class_check_failed, omp_clause_check_failed,
2399 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
2400 wrapper get_tree_code_name instead of calling tree_code_name directly.
2401 * tree-vrp.c (dump_asserts_for): Likewise.
2402 * tree-dump.c (dequeue_and_dump): Likewise.
2403 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
2404 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
2405 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
2406 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
2407 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
2408 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
2409 dump_gimple_omp_for): Likewise.
2410 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
2411 * tree-ssa-pre.c (print_pre_expr): Likewise.
2412 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
2413 * print-tree.c (print_node_brief, print_node): Likewise.
2414 * gimple.c (gimple_check_failed): Likewise.
2415 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
2416 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
2417 * config/mep/mep.c (mep_validate_vliw): Likewise.
2418 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
2419 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
2421 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
2423 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
2426 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
2428 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
2429 (MULTILIB_MATCHES): Add multilib for -march=v8.
2431 2013-10-15 Sriraman Tallam <tmsriram@google.com>
2434 * optc-save-gen.awk: Add extra parameter to the save and restore
2436 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
2437 * tree.c (build_optimization_node): New parameter. Add extra parameter
2438 to call to cl_optimization_save.
2439 (build_target_option_node): New parameter. Add extra parameter
2440 to call to cl_target_option_save.
2441 * tree.h (build_optimization_node): New parameter.
2442 (build_target_option_node): New parameter.
2443 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
2444 build_optimization_node and build_target_option_node.
2445 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
2446 (handle_pragma_push_options): Ditto.
2447 * toplev.c (process_options): Ditto.
2448 * opts.c (init_options_struct): Check for opts_set non-null.
2449 * target.def (target_option.save): New parameter.
2450 (target_option.restore): New parameter.
2451 * tm.texi: Generate.
2452 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
2453 (ix86_pragma_target_parse): Ditto.
2454 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
2456 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
2457 to build_optimization_node and build_target_option_node.
2458 (rs6000_valid_attribute_p): Ditto.
2459 (rs6000_pragma_target_parse): Ditto.
2460 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
2462 * config/i386/i386.h:
2463 TARGET_64BIT_P: New Macro
2464 TARGET_MMX_P: New Macro.
2465 TARGET_3DNOW_P: New Macro.
2466 TARGET_3DNOW_A_P: New Macro.
2467 TARGET_SSE_P: New Macro.
2468 TARGET_SSE2_P: New Macro.
2469 TARGET_SSE3_P: New Macro.
2470 TARGET_SSSE3_P: New Macro.
2471 TARGET_SSE4_1_P: New Macro.
2472 TARGET_SSE4_2_P: New Macro.
2473 TARGET_AVX_P: New Macro.
2474 TARGET_AVX2_P: New Macro.
2475 TARGET_AVX512F_P: New Macro.
2476 TARGET_AVX512PF_P: New Macro.
2477 TARGET_AVX512ER_P: New Macro.
2478 TARGET_AVX512CD_P: New Macro.
2479 TARGET_FMA_P: New Macro.
2480 TARGET_SSE4A_P: New Macro.
2481 TARGET_FMA4_P: New Macro.
2482 TARGET_XOP_P: New Macro.
2483 TARGET_LWP_P: New Macro.
2484 TARGET_ABM_P: New Macro.
2485 TARGET_BMI_P: New Macro.
2486 TARGET_BMI2_P: New Macro.
2487 TARGET_LZCNT_P: New Macro.
2488 TARGET_TBM_P: New Macro.
2489 TARGET_POPCNT_P: New Macro.
2490 TARGET_SAHF_P: New Macro.
2491 TARGET_MOVBE_P: New Macro.
2492 TARGET_CRC32_P: New Macro.
2493 TARGET_AES_P: New Macro.
2494 TARGET_PCLMUL_P: New Macro.
2495 TARGET_CMPXCHG16B_P: New Macro.
2496 TARGET_FSGSBASE_P: New Macro.
2497 TARGET_RDRND_P: New Macro.
2498 TARGET_F16C_P: New Macro.
2499 TARGET_RTM_P: New Macro.
2500 TARGET_HLE_P: New Macro.
2501 TARGET_RDSEED_P: New Macro.
2502 TARGET_PRFCHW_P: New Macro.
2503 TARGET_ADX_P: New Macro.
2504 TARGET_FXSR_P: New Macro.
2505 TARGET_XSAVE_P: New Macro.
2506 TARGET_XSAVEOPT_P: New Macro.
2507 TARGET_LP64_P: New Macro.
2508 TARGET_X32_P: New Macro.
2509 TARGET_FPMATH_DEFAULT_P: New Macro.
2510 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
2511 * config/i386/i386.c (ix86_option_override_internal): New parameters.
2513 Change ix86_tune_string to access opts->x_ix86_tune_string.
2514 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
2515 Change ix86_arch_string to access opts->x_ix86_arch_string.
2516 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
2517 Change ix86_pmode to access opts->x_ix86_pmode.
2518 Change ix86_abi to access opts->x_ix86_abi.
2519 Change ix86_cmodel to access opts->x_ix86_cmodel.
2520 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
2521 Change ix86_isa_flags_explicit to access
2522 opts->x_ix86_isa_flags_explicit.
2523 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
2524 Change ix86_regparm to access opts->x_ix86_regparm.
2525 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
2526 Change ix86_preferred_stack_boundary_arg to access
2527 opts->x_ix86_preferred_stack_boundary_arg.
2528 Change ix86_force_align_arg_pointer to access
2529 opts->x_ix86_force_align_arg_pointer.
2530 Change ix86_incoming_stack_boundar_arg to access
2531 opts->x_ix86_incoming_stack_boundar_arg.
2532 Change ix86_fpmath to access opts->x_ix86_fpmath.
2533 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
2534 Change ix86_recip_name to access opts->x_ix86_recip_name.
2535 Change ix86_stack_protector_guard to access
2536 opts->x_ix86_stack_protector_guard.
2537 Change ix86_tune_memcpy_strategy to access
2538 opts->x_ix86_tune_memcpy_strategy.
2539 Change ix86_tune_memset_strategy to access
2540 opts->x_ix86_tune_memset_strategy.
2541 Change global_options to access opts.
2542 Change global_options_set to access opts_set.
2543 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
2544 Change TARGET_MMX to TARGET_MMX_P (opts->...).
2545 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
2546 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
2547 Change TARGET_SSE to TARGET_SSE_P (opts->...).
2548 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
2549 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
2550 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
2551 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
2552 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
2553 Change TARGET_AVX to TARGET_AVX_P (opts->...).
2554 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
2555 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
2556 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
2557 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
2558 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
2559 Change TARGET_FMA to TARGET_FMA_P (opts->...).
2560 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
2561 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
2562 Change TARGET_XOP to TARGET_XOP_P (opts->...).
2563 Change TARGET_LWP to TARGET_LWP_P (opts->...).
2564 Change TARGET_ABM to TARGET_ABM_P (opts->...).
2565 Change TARGET_BMI to TARGET_BMI_P (opts->...).
2566 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
2567 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
2568 Change TARGET_TBM to TARGET_TBM_P (opts->...).
2569 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
2570 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
2571 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
2572 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
2573 Change TARGET_AES to TARGET_AES_P (opts->...).
2574 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
2575 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
2576 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
2577 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
2578 Change TARGET_F16C to TARGET_F16C_P (opts->...).
2579 Change TARGET_RTM to TARGET_RTM_P (opts->...).
2580 Change TARGET_HLE to TARGET_HLE_P (opts->...).
2581 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
2582 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
2583 Change TARGET_ADX to TARGET_ADX_P (opts->...).
2584 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
2585 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
2586 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
2587 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
2588 Change TARGET_X32 to TARGET_X32_P (opts->...).
2589 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
2590 Change TARGET_FLOAT_RETURNS_IN_80387 to
2591 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
2592 (ix86_function_specific_save): New parameter. Use opts-> fields
2593 to replace global fields.
2594 (ix86_function_specific_restore): Ditto.
2595 (ix86_valid_target_attribute_inner_p): New parameters.
2597 Fix call to ix86_handle_option and set_option.
2598 (ix86_valid_target_attribute_tree): New parameters.
2599 Change global_options to access opts.
2600 Change global_options_set to access opts_set.
2601 Fix call to ix86_valid_target_attribute_inner_p.
2602 Change ix86_tune_string to access opts->x_ix86_tune_string.
2603 Change ix86_arch_string to access opts->x_ix86_arch_string.
2604 Change ix86_fpmath to access opts->x_ix86_fpmath
2605 Fix call to ix86_option_override_internal.
2606 Fix call to ix86_add_new_builtins.
2607 Fix calls to build_optimization_node and build_target_option_node.
2608 (ix86_valid_target_attribute_p): Remove access to global_options.
2609 Use new gcc_options structure func_options.
2610 Fix call to ix86_valid_target_attribute_tree.
2611 Fix call to build_optimization_node.
2612 (get_builtin_code_for_version): Fix call to
2613 ix86_valid_target_attribute_tree.
2615 2013-10-15 David Malcolm <dmalcolm@redhat.com>
2617 * Makefile.in (PICFLAG): New.
2618 (enable_host_shared): New.
2619 (INTERNAL_CFLAGS): Use PICFLAG.
2620 (LIBIBERTY): Use pic build of libiberty.a if configured with
2621 --enable-host-shared.
2622 * configure.ac: Add --enable-host-shared, setting up new
2624 * configure: Regenerate.
2625 * doc/install.texi (--enable-shared): Add note contrasting it with ...
2626 (--enable-host-shared): New option.
2628 2013-10-15 Richard Biener <rguenther@suse.de>
2630 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
2631 for built-in functions.
2633 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
2635 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
2636 (optimize_range_tests_1): New function,
2637 extracted from optimize_range_tests.
2638 (optimize_range_tests_xor): Similarly.
2639 (optimize_range_tests_diff): New function.
2640 (optimize_range_tests): Use optimize_range_tests_1.
2642 2013-10-15 Cong Hou <congh@google.com>
2644 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
2645 requirement of the reduction pattern so that one operand of the
2646 reduction operation can come from outside of the loop.
2648 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2650 * config/arm/neon-schedgen.ml: Remove.
2651 * config/arm/cortex-a9-neon.md: Remove comment regarding
2654 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2656 * config/arm/types: Remove old neon types.
2658 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2660 * config/arm/cortex-a7.md
2661 (cortex_a7_neon_type): New.
2662 (cortex_a7_neon_mul): Update for new types.
2663 (cortex_a7_neon_mla): Likewise.
2664 (cortex_a7_neon): Likewise.
2666 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2668 * config/arm/cortex-a15-neon.md
2669 (cortex_a15_neon_type): New,
2671 (cortex_a15_neon_int_1): Remove.
2672 (cortex_a15_neon_int_2): Likewise.
2673 (cortex_a15_neon_int_3): Likewise.
2674 (cortex_a15_neon_int_4): Likewise.
2675 (cortex_a15_neon_int_5): Likewise.
2676 (cortex_a15_neon_vqneg_vqabs): Likewise.
2677 (cortex_a15_neon_vmov): Likewise.
2678 (cortex_a15_neon_vaba): Likewise.
2679 (cortex_a15_neon_vaba_qqq): Likewise.
2680 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2681 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2682 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
2684 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2685 (cortex_a15_neon_mla_qqq_8_16): Likewise.
2686 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
2687 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2688 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2689 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
2690 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2691 (cortex_a15_neon_shift_1): Likewise.
2692 (cortex_a15_neon_shift_2): Likewise.
2693 (cortex_a15_neon_shift_3): Likewise.
2694 (cortex_a15_neon_vshl_ddd): Likewise.
2695 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2696 (cortex_a15_neon_vsra_vrsra): Likewise.
2697 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
2698 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
2699 (cortex_a15_neon_bp_3cycle): Likewise.
2700 (cortex_a15_neon_ldm_2): Likewise.
2701 (cortex_a15_neon_stm_2): Likewise.
2702 (cortex_a15_neon_mcr): Likewise.
2703 (cortex_a15_neon_mrc): Likewise.
2704 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
2705 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
2706 (cortex_a15_neon_fp_vmul_ddd): Likewise.
2707 (cortex_a15_neon_fp_vmul_qqd): Likewise.
2708 (cortex_a15_neon_fp_vmla_ddd): Likewise.
2709 (cortex_a15_neon_fp_vmla_qqq): Likewise.
2710 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
2711 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
2712 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2713 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2714 (cortex_a15_neon_bp_simple): Likewise.
2715 (cortex_a15_neon_bp_2cycle): Likewise.
2716 (cortex_a15_neon_bp_3cycle): Likewise.
2717 (cortex_a15_neon_vld1_1_2_regs): Likewise.
2718 (cortex_a15_neon_vld1_3_4_regs): Likewise.
2719 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2720 (cortex_a15_neon_vld2_4_regs): Likewise.
2721 (cortex_a15_neon_vld3_vld4): Likewise.
2722 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2723 (cortex_a15_neon_vst1_3_4_regs): Likewise.
2724 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
2725 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
2726 (cortex_a15_neon_vst3_vst4): Rename to...
2727 (cortex_a15_neon_vst4): This, update for new attributes.
2728 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
2729 (cortex_a15_neon_vld3_vld4_lane): Likewise.
2730 (cortex_a15_neon_vst1_vst2_lane): Likewise.
2731 (cortex_a15_neon_vst3_vst4_lane): Likewise.
2732 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
2733 (cortex_a15_neon_ldm_2): Likewise.
2734 (cortex_a15_neon_stm_2): Likewise.
2735 (cortex_a15_neon_mcr): Likewise.
2736 (cortex_a15_neon_mcr_2_mcrr): Likewise.
2737 (cortex_a15_neon_mrc): Likewise.
2738 (cortex_a15_neon_mrrc): Likewise.
2740 (cortex_a15_neon_abd): New.
2741 (cortex_a15_neon_abd_q): Likewise.
2742 (cortex_a15_neon_aba): Likewise.
2743 (cortex_a15_neon_aba_q): Likewise.
2744 (cortex_a15_neon_acc): Likewise.
2745 (cortex_a15_neon_acc_q): Likewise.
2746 (cortex_a15_neon_arith_basic): Likewise.
2747 (cortex_a15_neon_arith_complex): Likewise.
2748 (cortex_a15_neon_multiply): Likewise.
2749 (cortex_a15_neon_multiply_q): Likewise.
2750 (cortex_a15_neon_mla): Likewise.
2751 (cortex_a15_neon_mla_q): Likewise.
2752 (cortex_a15_neon_sat_mla_long): Likewise.
2753 (cortex_a15_neon_shift_acc): Likewise.
2754 (cortex_a15_neon_shift_imm_basic): Likewise.
2755 (cortex_a15_neon_shift_imm_complex): Likewise.
2756 (cortex_a15_neon_shift_reg_basic): Likewise.
2757 (cortex_a15_neon_shift_reg_basic_q): Likewise.
2758 (cortex_a15_neon_shift_reg_complex): Likewise.
2759 (cortex_a15_neon_shift_reg_complex_q): Likewise.
2760 (cortex_a15_neon_fp_negabs): Likewise
2761 (cortex_a15_neon_fp_arith): Likewise
2762 (cortex_a15_neon_fp_arith_q): Likewise
2763 (cortex_a15_neon_fp_cvt_int): Likewise
2764 (cortex_a15_neon_fp_cvt_int_q): Likewise
2765 (cortex_a15_neon_fp_cvt_16): Likewise
2766 (cortex_a15_neon_fp_mul): Likewise
2767 (cortex_a15_neon_fp_mul_q): Likewise
2768 (cortex_a15_neon_fp_mla): Likewise
2769 (cortex_a15_neon_fp_mla_q): Likewise
2770 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
2771 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
2772 (cortex_a15_neon_bitops): Likewise.
2773 (cortex_a15_neon_bitops_q): Likewise.
2774 (cortex_a15_neon_from_gp): Likewise.
2775 (cortex_a15_neon_from_gp_q): Likewise.
2776 (cortex_a15_neon_tbl3_tbl4): Likewise.
2777 (cortex_a15_neon_zip_q): Likewise.
2778 (cortex_a15_neon_to_gp): Likewise.
2779 (cortex_a15_neon_load_a): Likewise.
2780 (cortex_a15_neon_load_b): Likewise.
2781 (cortex_a15_neon_load_c): Likewise.
2782 (cortex_a15_neon_load_d): Likewise.
2783 (cortex_a15_neon_load_e): Likewise.
2784 (cortex_a15_neon_load_f): Likewise.
2785 (cortex_a15_neon_store_a): Likewise.
2786 (cortex_a15_neon_store_b): Likewise.
2787 (cortex_a15_neon_store_c): Likewise.
2788 (cortex_a15_neon_store_d): Likewise.
2789 (cortex_a15_neon_store_e): Likewise.
2790 (cortex_a15_neon_store_f): Likewise.
2791 (cortex_a15_neon_store_g): Likewise.
2792 (cortex_a15_neon_store_h): Likewise.
2793 (cortex_a15_vfp_to_from_gp): Likewise.
2795 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2797 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
2799 (cortex_a9_neon_vshl_ddd): Remove.
2800 (cortex_a9_neon_vst3_vst4): Likewise.
2801 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
2803 (cortex_a9_neon_bit_ops_q): New.
2805 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
2806 (cortex_a9_neon_int_2): Likewise.
2807 (cortex_a9_neon_int_3): Likewise.
2808 (cortex_a9_neon_int_4): Likewise.
2809 (cortex_a9_neon_int_5): Likewise.
2810 (cortex_a9_neon_vqneg_vqabs): Likewise.
2811 (cortex_a9_neon_vmov): Likewise.
2812 (cortex_a9_neon_vaba): Likewise.
2813 (cortex_a9_neon_vaba_qqq): Likewise.
2814 (cortex_a9_neon_shift_1): Likewise.
2815 (cortex_a9_neon_shift_2): Likewise.
2816 (cortex_a9_neon_shift_3): Likewise.
2817 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2818 (cortex_a9_neon_vsra_vrsra): Likewise.
2819 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2820 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2821 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
2823 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2824 (cortex_a9_neon_mla_qqq_8_16): Likewise.
2825 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
2827 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2828 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2829 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
2830 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2831 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
2832 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
2833 (cortex_a9_neon_fp_vsum): Likewise.
2834 (cortex_a9_neon_fp_vmul_ddd): Likewise.
2835 (cortex_a9_neon_fp_vmul_qqd): Likewise.
2836 (cortex_a9_neon_fp_vmla_ddd): Likewise.
2837 (cortex_a9_neon_fp_vmla_qqq): Likewise.
2838 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
2839 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
2840 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2841 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2842 (cortex_a9_neon_bp_simple): Likewise.
2843 (cortex_a9_neon_bp_2cycle): Likewise.
2844 (cortex_a9_neon_bp_3cycle): Likewise.
2845 (cortex_a9_neon_ldr): Likewise.
2846 (cortex_a9_neon_str): Likewise.
2847 (cortex_a9_neon_vld1_1_2_regs): Likewise.
2848 (cortex_a9_neon_vld1_3_4_regs): Likewise.
2849 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2850 (cortex_a9_neon_vld2_4_regs): Likewise.
2851 (cortex_a9_neon_vld3_vld4): Likewise.
2852 (cortex_a9_neon_vld1_vld2_lane): Likewise.
2853 (cortex_a9_neon_vld3_vld4_lane): Likewise.
2854 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
2855 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2856 (cortex_a9_neon_vst1_3_4_regs): Likewise.
2857 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
2858 (cortex_a9_neon_vst1_vst2_lane): Likewise.
2859 (cortex_a9_neon_vst3_vst4_lane): Likewise.
2860 (cortex_a9_neon_mcr): Likewise.
2861 (cortex_a9_neon_mcr_2_mcrr): Likewise.
2862 (cortex_a9_neon_mrc): Likewise.
2863 (cortex_a9_neon_mrrc): Likewise.
2865 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2867 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
2869 (cortex_a8_neon_vshl_ddd): Remove.
2870 (cortex_a8_neon_vst3_vst4): Likewise.
2871 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
2873 (cortex_a8_neon_bit_ops_q): New.
2875 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
2876 (cortex_a8_neon_int_2): Likewise..
2877 (cortex_a8_neon_int_3): Likewise.
2878 (cortex_a8_neon_int_5): Likewise.
2879 (cortex_a8_neon_vqneg_vqabs): Likewise.
2880 (cortex_a8_neon_int_4): Likewise.
2881 (cortex_a8_neon_vaba): Likewise.
2882 (cortex_a8_neon_vaba_qqq): Likewise.
2883 (cortex_a8_neon_shift_1): Likewise.
2884 (cortex_a8_neon_shift_2): Likewise.
2885 (cortex_a8_neon_shift_3): Likewise.
2886 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2887 (cortex_a8_neon_vsra_vrsra): Likewise.
2888 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2889 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2890 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
2892 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2893 (cortex_a8_neon_mla_qqq_8_16): Likewise.
2894 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
2896 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2897 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2898 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
2899 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2900 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
2901 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
2902 (cortex_a8_neon_fp_vsum): Likewise.
2903 (cortex_a8_neon_fp_vmul_ddd): Likewise.
2904 (cortex_a8_neon_fp_vmul_qqd): Likewise.
2905 (cortex_a8_neon_fp_vmla_ddd): Likewise.
2906 (cortex_a8_neon_fp_vmla_qqq): Likewise.
2907 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
2908 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
2909 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2910 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2911 (cortex_a8_neon_bp_simple): Likewise.
2912 (cortex_a8_neon_bp_2cycle): Likewise.
2913 (cortex_a8_neon_bp_3cycle): Likewise.
2914 (cortex_a8_neon_ldr): Likewise.
2915 (cortex_a8_neon_str): Likewise.
2916 (cortex_a8_neon_vld1_1_2_regs): Likewise.
2917 (cortex_a8_neon_vld1_3_4_regs): Likewise.
2918 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2919 (cortex_a8_neon_vld2_4_regs): Likewise.
2920 (cortex_a8_neon_vld3_vld4): Likewise.
2921 (cortex_a8_neon_vld1_vld2_lane): Likewise.
2922 (cortex_a8_neon_vld3_vld4_lane): Likewise.
2923 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2924 (cortex_a8_neon_vst1_3_4_regs): Likewise.
2925 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
2926 (cortex_a8_neon_vst1_vst2_lane): Likewise.
2927 (cortex_a8_neon_vst3_vst4_lane): Likewise.
2928 (cortex_a8_neon_mcr): Likewise.
2929 (cortex_a8_neon_mcr_2_mcrr): Likewise.
2930 (cortex_a8_neon_mrc): Likewise.
2931 (cortex_a8_neon_mrrc): Likewise.
2933 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2935 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
2937 * config/aarch64/aarch64-simd.md (neon_type): Remove.
2938 (aarch64_simd_dup<mode>): Add "type" attribute.
2939 (aarch64_dup_lane<mode>): Likewise.
2940 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
2941 (*aarch64_simd_mov<mode>): Likewise.
2942 (aarch64_simd_mov_from_<mode>low): Likewise.
2943 (aarch64_simd_mov_from_<mode>high): Likewise.
2944 (orn<mode>3): Likewise.
2945 (bic<mode>3): Likewise.
2946 (add<mode>3): Likewise.
2947 (sub<mode>3): Likewise.
2948 (mul<mode>3): Likewise.
2949 (*aarch64_mul3_elt<mode>): Likewise.
2950 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2951 (*aarch64_mul3_elt_to_128df): Likewise.
2952 (*aarch64_mul3_elt_to_64v2df): Likewise.
2953 (neg<mode>2): Likewise.
2954 (abs<mode>2): Likewise.
2955 (abd<mode>_3): Likewise.
2956 (aba<mode>_3): Likewise.
2957 (fabd<mode>_3): Likewise.
2958 (*fabd_scalar<mode>3): Likewise.
2959 (and<mode>3): Likewise.
2960 (ior<mode>3): Likewise.
2961 (xor<mode>3): Likewise.
2962 (one_cmpl<mode>2): Likewise.
2963 (aarch64_simd_vec_set<mode>): Likewise.
2964 (aarch64_simd_lshr<mode>): Likewise.
2965 (aarch64_simd_ashr<mode>): Likewise.
2966 (aarch64_simd_imm_shl<mode>): Likewise.
2967 (aarch64_simd_reg_sshl<mode): Likewise.
2968 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
2969 (aarch64_simd_reg_shl<mode>_signed): Likewise.
2970 (aarch64_simd_vec_setv2di): Likewise.
2971 (aarch64_simd_vec_set<mode>): Likewise.
2972 (aarch64_mla<mode>): Likewise.
2973 (*aarch64_mla_elt<mode>): Likewise.
2974 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2975 (aarch64_mls<mode>): Likewise.
2976 (*aarch64_mls_elt<mode>): Likewise.
2977 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2978 (<su><maxmin><mode>3): Likewise.
2979 (move_lo_quad_<mode>): Likewise.
2980 (aarch64_simd_move_hi_quad_<mode>): Likewise.
2981 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
2982 (vec_pack_trunc_<mode>): Likewise.
2983 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
2984 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
2985 (*aarch64_<su>mlal_lo<mode>): Likewise.
2986 (*aarch64_<su>mlal_hi<mode>): Likewise.
2987 (*aarch64_<su>mlsl_lo<mode>): Likewise.
2988 (*aarch64_<su>mlsl_hi<mode>): Likewise.
2989 (*aarch64_<su>mlal<mode>): Likewise.
2990 (*aarch64_<su>mlsl<mode>): Likewise.
2991 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
2992 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
2993 (add<mode>3): Likewise.
2994 (sub<mode>3): Likewise.
2995 (mul<mode>3): Likewise.
2996 (div<mode>3): Likewise.
2997 (neg<mode>2): Likewise.
2998 (abs<mode>2): Likewise.
2999 (fma<mode>4): Likewise.
3000 (*aarch64_fma4_elt<mode>): Likewise.
3001 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3002 (*aarch64_fma4_elt_to_128df): Likewise.
3003 (*aarch64_fma4_elt_to_64v2df): Likewise.
3004 (fnma<mode>4): Likewise.
3005 (*aarch64_fnma4_elt<mode>): Likewise.
3006 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3007 (*aarch64_fnma4_elt_to_128df): Likewise.
3008 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3009 (<frint_pattern><mode>2): Likewise.
3010 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3011 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3012 (vec_unpacks_lo_v4sf): Likewise.
3013 (aarch64_float_extend_lo_v2df): Likewise.
3014 (vec_unpacks_hi_v4sf): Likewise.
3015 (aarch64_float_truncate_lo_v2sf): Likewise.
3016 (aarch64_float_truncate_hi_v4sf): Likewise.
3017 (aarch64_vmls<mode>): Likewise.
3018 (<su><maxmin><mode>3): Likewise.
3019 (<maxmin_uns><mode>3): Likewise.
3020 (reduc_<sur>plus_<mode>): Likewise.
3021 (reduc_<sur>plus_v2di): Likewise.
3022 (reduc_<sur>plus_v2si): Likewise.
3023 (reduc_<sur>plus_<mode>): Likewise.
3024 (aarch64_addpv4sf): Likewise.
3025 (clz<mode>2): Likewise.
3026 (reduc_<maxmin_uns>_<mode>): Likewise.
3027 (reduc_<maxmin_uns>_v2di): Likewise.
3028 (reduc_<maxmin_uns>_v2si): Likewise.
3029 (reduc_<maxmin_uns>_<mode>): Likewise.
3030 (reduc_<maxmin_uns>_v4sf): Likewise.
3031 (aarch64_simd_bsl<mode>_internal): Likewise.
3032 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3033 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3034 (aarch64_get_lane<mode>): Likewise.
3035 (*aarch64_combinez<mode>): Likewise.
3036 (aarch64_combine<mode>): Likewise.
3037 (aarch64_simd_combine<mode>): Likewise.
3038 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3039 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3040 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3041 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3042 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3043 (aarch64_<sur>h<addsub><mode>): Likewise.
3044 (aarch64_<sur><addsub>hn<mode>): Likewise.
3045 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3046 (aarch64_pmul<mode>): Likewise.
3047 (aarch64_<su_optab><optab><mode>): Likewise.
3048 (aarch64_<sur>qadd<mode>): Likewise.
3049 (aarch64_sqmovun<mode>): Likewise.
3050 (aarch64_<sur>qmovn<mode>): Likewise.
3051 (aarch64_s<optab><mode>): Likewise.
3052 (aarch64_sq<r>dmulh<mode>): Likewise.
3053 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3054 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3055 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3056 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3057 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3058 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3059 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3060 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3061 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3062 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3063 (aarch64_sqdmull<mode>): Likewise.
3064 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3065 (aarch64_sqdmull_n<mode>): Likewise.
3066 (aarch64_sqdmull2<mode>_internal): Likewise.
3067 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3068 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3069 (aarch64_<sur>shl<mode>): Likewise.
3070 (aarch64_<sur>q<r>shl<mode>
3071 (aarch64_<sur>shll_n<mode>): Likewise.
3072 (aarch64_<sur>shll2_n<mode>): Likewise.
3073 (aarch64_<sur>shr_n<mode>): Likewise.
3074 (aarch64_<sur>sra_n<mode>): Likewise.
3075 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3076 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3077 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3078 (aarch64_cm<optab><mode>): Likewise.
3079 (aarch64_cm<optab>di): Likewise.
3080 (aarch64_cm<optab><mode>): Likewise.
3081 (aarch64_cm<optab>di): Likewise.
3082 (aarch64_cmtst<mode>): Likewise.
3083 (aarch64_cmtstdi): Likewise.
3084 (aarch64_cm<optab><mode>): Likewise.
3085 (*aarch64_fac<optab><mode>): Likewise.
3086 (aarch64_addp<mode>): Likewise.
3087 (aarch64_addpdi): Likewise.
3088 (sqrt<mode>2): Likewise.
3089 (vec_load_lanesoi<mode>): Likewise.
3090 (vec_store_lanesoi<mode>): Likewise.
3091 (vec_load_lanesci<mode>): Likewise.
3092 (vec_store_lanesci<mode>): Likewise.
3093 (vec_load_lanesxi<mode>): Likewise.
3094 (vec_store_lanesxi<mode>): Likewise.
3095 (*aarch64_mov<mode>): Likewise.
3096 (aarch64_ld2<mode>_dreg): Likewise.
3097 (aarch64_ld2<mode>_dreg): Likewise.
3098 (aarch64_ld3<mode>_dreg): Likewise.
3099 (aarch64_ld3<mode>_dreg): Likewise.
3100 (aarch64_ld4<mode>_dreg): Likewise.
3101 (aarch64_ld4<mode>_dreg): Likewise.
3102 (aarch64_tbl1<mode>): Likewise.
3103 (aarch64_tbl2v16qi): Likewise.
3104 (aarch64_combinev16qi): Likewise.
3105 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3106 (aarch64_st2<mode>_dreg): Likewise.
3107 (aarch64_st2<mode>_dreg): Likewise.
3108 (aarch64_st3<mode>_dreg): Likewise.
3109 (aarch64_st3<mode>_dreg): Likewise.
3110 (aarch64_st4<mode>_dreg): Likewise.
3111 (aarch64_st4<mode>_dreg): Likewise.
3112 (*aarch64_simd_ld1r<mode>): Likewise.
3113 (aarch64_frecpe<mode>): Likewise.
3114 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3115 (aarch64_frecps<mode>): Likewise.
3117 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3119 * config/arm/iterators.md (V_elem_ch): New.
3121 (VQH_type): Likewise.
3122 * config/arm/arm.md (is_neon_type): New.
3123 (conds): Use is_neon_type.
3124 (anddi3_insn): Update type attribute.
3125 (xordi3_insn): Likewise.
3126 (one_cmpldi2): Likewise.
3127 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3128 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3129 (*movmisalign<mode>_neon_store): Likewise.
3130 (*movmisalign<mode>_neon_load): Likewise.
3131 (vec_set<mode>_internal): Likewise.
3132 (vec_set<mode>_internal): Likewise.
3133 (vec_setv2di_internal): Likewise.
3134 (vec_extract<mode>): Likewise.
3135 (vec_extract<mode>): Likewise.
3136 (vec_extractv2di): Likewise.
3137 (*add<mode>3_neon): Likewise.
3138 (adddi3_neon): Likewise.
3139 (*sub<mode>3_neon): Likewise.
3140 (subdi3_neon): Likewise.
3141 (fma<VCVTF:mode>4): Likewise.
3142 (fma<VCVTF:mode>4_intrinsic): Likewise.
3143 (*fmsub<VCVTF:mode>4): Likewise.
3144 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3145 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3146 (ior<mode>3): Likewise.
3147 (and<mode>3): Likewise.
3148 (orn<mode>3_neon): Likewise.
3149 (orndi3_neon): Likewise.
3150 (bic<mode>3_neon): Likewise.
3151 (bicdi3_neon): Likewise.
3152 (xor<mode>3): Likewise.
3153 (one_cmpl<mode>2): Likewise.
3154 (abs<mode>2): Likewise.
3155 (neg<mode>2): Likewise.
3156 (negdi2_neon): Likewise.
3157 (*umin<mode>3_neon): Likewise.
3158 (*umax<mode>3_neon): Likewise.
3159 (*smin<mode>3_neon): Likewise.
3160 (*smax<mode>3_neon): Likewise.
3161 (vashl<mode>3): Likewise.
3162 (vashr<mode>3_imm): Likewise.
3163 (vlshr<mode>3_imm): Likewise.
3164 (ashl<mode>3_signed): Likewise.
3165 (ashl<mode>3_unsigned): Likewise.
3166 (neon_load_count): Likewise.
3167 (ashldi3_neon_noclobber): Likewise.
3168 (ashldi3_neon): Likewise.
3169 (signed_shift_di3_neon): Likewise.
3170 (unsigned_shift_di3_neon): Likewise.
3171 (ashrdi3_neon_imm_noclobber): Likewise.
3172 (lshrdi3_neon_imm_noclobber): Likewise.
3173 (<shift>di3_neon): Likewise.
3174 (widen_ssum<mode>3): Likewise.
3175 (widen_usum<mode>3): Likewise.
3176 (quad_halves_<code>v4si): Likewise.
3177 (quad_halves_<code>v4sf): Likewise.
3178 (quad_halves_<code>v8hi): Likewise.
3179 (quad_halves_<code>v16qi): Likewise.
3180 (reduc_splus_v2di): Likewise.
3181 (neon_vpadd_internal<mode>): Likewise.
3182 (neon_vpsmin<mode>): Likewise.
3183 (neon_vpsmax<mode>): Likewise.
3184 (neon_vpumin<mode>): Likewise.
3185 (neon_vpumax<mode>): Likewise.
3186 (*ss_add<mode>_neon): Likewise.
3187 (*us_add<mode>_neon): Likewise.
3188 (*ss_sub<mode>_neon): Likewise.
3189 (*us_sub<mode>_neon): Likewise.
3190 (neon_vadd<mode>_unspec): Likewise.
3191 (neon_vaddl<mode>): Likewise.
3192 (neon_vaddw<mode>): Likewise.
3193 (neon_vhadd<mode>): Likewise.
3194 (neon_vqadd<mode>): Likewise.
3195 (neon_vaddhn<mode>): Likewise.
3196 (neon_vmul<mode>): Likewise.
3197 (neon_vfms<VCVTF:mode>): Likewise.
3198 (neon_vmlal<mode>): Likewise.
3199 (neon_vmls<mode>): Likewise.
3200 (neon_vmlsl<mode>): Likewise.
3201 (neon_vqdmulh<mode>): Likewise.
3202 (neon_vqdmlal<mode>): Likewise.
3203 (neon_vqdmlsl<mode>): Likewise.
3204 (neon_vmull<mode>): Likewise.
3205 (neon_vqdmull<mode>): Likewise.
3206 (neon_vsub<mode>_unspec): Likewise.
3207 (neon_vsubl<mode>): Likewise.
3208 (neon_vsubw<mode>): Likewise.
3209 (neon_vqsub<mode>): Likewise.
3210 (neon_vhsub<mode>): Likewise.
3211 (neon_vsubhn<mode>): Likewise.
3212 (neon_vceq<mode>): Likewise.
3213 (neon_vcge<mode>): Likewise.
3214 (neon_vcgeu<mode>): Likewise.
3215 (neon_vcgt<mode>): Likewise.
3216 (neon_vcgtu<mode>): Likewise.
3217 (neon_vcle<mode>): Likewise.
3218 (neon_vclt<mode>): Likewise.
3219 (neon_vcage<mode>): Likewise.
3220 (neon_vcagt<mode>): Likewise.
3221 (neon_vtst<mode>): Likewise.
3222 (neon_vabd<mode>): Likewise.
3223 (neon_vabdl<mode>): Likewise.
3224 (neon_vaba<mode>): Likewise.
3225 (neon_vabal<mode>): Likewise.
3226 (neon_vmax<mode>): Likewise.
3227 (neon_vmin<mode>): Likewise.
3228 (neon_vpaddl<mode>): Likewise.
3229 (neon_vpadal<mode>): Likewise.
3230 (neon_vpmax<mode>): Likewise.
3231 (neon_vpmin<mode>): Likewise.
3232 (neon_vrecps<mode>): Likewise.
3233 (neon_vrsqrts<mode>): Likewise.
3234 (neon_vqabs<mode>): Likewise.
3235 (neon_vqneg<mode>): Likewise.
3236 (neon_vcls<mode>): Likewise.
3237 (clz<mode>2): Likewise.
3238 (popcount<mode>2): Likewise.
3239 (neon_vrecpe<mode>): Likewise.
3240 (neon_vrsqrte<mode>): Likewise.
3241 (neon_vget_lane<mode>_sext_internal): Likewise.
3242 (neon_vget_lane<mode>_zext_internal): Likewise.
3243 (neon_vdup_n<mode>): Likewise.
3244 (neon_vdup_n<mode>): Likewise.
3245 (neon_vdup_nv2di): Likewise.
3246 (neon_vdup_lane<mode>_interal): Likewise.
3247 (*neon_vswp<mode>): Likewise.
3248 (neon_vcombine<mode>): Likewise.
3249 (float<mode><V_cvtto>2): Likewise.
3250 (floatuns<mode><V_cvtto>2): Likewise.
3251 (fix_trunc<mode><V_cvtto>2): Likewise.
3252 (fixuns_trunc<mode><V_cvtto>2
3253 (neon_vcvt<mode>): Likewise.
3254 (neon_vcvt<mode>): Likewise.
3255 (neon_vcvtv4sfv4hf): Likewise.
3256 (neon_vcvtv4hfv4sf): Likewise.
3257 (neon_vcvt_n<mode>): Likewise.
3258 (neon_vcvt_n<mode>): Likewise.
3259 (neon_vmovn<mode>): Likewise.
3260 (neon_vqmovn<mode>): Likewise.
3261 (neon_vqmovun<mode>): Likewise.
3262 (neon_vmovl<mode>): Likewise.
3263 (neon_vmul_lane<mode>): Likewise.
3264 (neon_vmul_lane<mode>): Likewise.
3265 (neon_vmull_lane<mode>): Likewise.
3266 (neon_vqdmull_lane<mode>): Likewise.
3267 (neon_vqdmulh_lane<mode>): Likewise.
3268 (neon_vqdmulh_lane<mode>): Likewise.
3269 (neon_vmla_lane<mode>): Likewise.
3270 (neon_vmla_lane<mode>): Likewise.
3271 (neon_vmlal_lane<mode>): Likewise.
3272 (neon_vqdmlal_lane<mode>): Likewise.
3273 (neon_vmls_lane<mode>): Likewise.
3274 (neon_vmls_lane<mode>): Likewise.
3275 (neon_vmlsl_lane<mode>): Likewise.
3276 (neon_vqdmlsl_lane<mode>): Likewise.
3277 (neon_vext<mode>): Likewise.
3278 (neon_vrev64<mode>): Likewise.
3279 (neon_vrev32<mode>): Likewise.
3280 (neon_vrev16<mode>): Likewise.
3281 (neon_vbsl<mode>_internal): Likewise.
3282 (neon_vshl<mode>): Likewise.
3283 (neon_vqshl<mode>): Likewise.
3284 (neon_vshr_n<mode>): Likewise.
3285 (neon_vshrn_n<mode>): Likewise.
3286 (neon_vqshrn_n<mode>): Likewise.
3287 (neon_vqshrun_n<mode>): Likewise.
3288 (neon_vshl_n<mode>): Likewise.
3289 (neon_vqshl_n<mode>): Likewise.
3290 (neon_vqshlu_n<mode>): Likewise.
3291 (neon_vshll_n<mode>): Likewise.
3292 (neon_vsra_n<mode>): Likewise.
3293 (neon_vsri_n<mode>): Likewise.
3294 (neon_vsli_n<mode>): Likewise.
3295 (neon_vtbl1v8qi): Likewise.
3296 (neon_vtbl2v8qi): Likewise.
3297 (neon_vtbl3v8qi): Likewise.
3298 (neon_vtbl4v8qi): Likewise.
3299 (neon_vtbl1v16qi): Likewise.
3300 (neon_vtbl2v16qi): Likewise.
3301 (neon_vcombinev16qi): Likewise.
3302 (neon_vtbx1v8qi): Likewise.
3303 (neon_vtbx2v8qi): Likewise.
3304 (neon_vtbx3v8qi): Likewise.
3305 (neon_vtbx4v8qi): Likewise.
3306 (*neon_vtrn<mode>_insn): Likewise.
3307 (*neon_vzip<mode>_insn): Likewise.
3308 (*neon_vuzp<mode>_insn): Likewise.
3309 (neon_vld1<mode>): Likewise.
3310 (neon_vld1_lane<mode>): Likewise.
3311 (neon_vld1_lane<mode>): Likewise.
3312 (neon_vld1_dup<mode>): Likewise.
3313 (neon_vld1_dup<mode>): Likewise.
3314 (neon_vld1_dupv2di): Likewise.
3315 (neon_vst1<mode>): Likewise.
3316 (neon_vst1_lane<mode>): Likewise.
3317 (neon_vst1_lane<mode>): Likewise.
3318 (neon_vld2<mode>): Likewise.
3319 (neon_vld2<mode>): Likewise.
3320 (neon_vld2_lane<mode>): Likewise.
3321 (neon_vld2_lane<mode>): Likewise.
3322 (neon_vld2_dup<mode>): Likewise.
3323 (neon_vst2<mode>): Likewise.
3324 (neon_vst2<mode>): Likewise.
3325 (neon_vst2_lane<mode>): Likewise.
3326 (neon_vst2_lane<mode>): Likewise.
3327 (neon_vld3<mode>): Likewise.
3328 (neon_vld3qa<mode>): Likewise.
3329 (neon_vld3qb<mode>): Likewise.
3330 (neon_vld3_lane<mode>): Likewise.
3331 (neon_vld3_lane<mode>): Likewise.
3332 (neon_vld3_dup<mode>): Likewise.
3333 (neon_vst3<mode>): Likewise.
3334 (neon_vst3qa<mode>): Likewise.
3335 (neon_vst3qb<mode>): Likewise.
3336 (neon_vst3_lane<mode>): Likewise.
3337 (neon_vst3_lane<mode>): Likewise.
3338 (neon_vld4<mode>): Likewise.
3339 (neon_vld4qa<mode>): Likewise.
3340 (neon_vld4qb<mode>): Likewise.
3341 (neon_vld4_lane<mode>): Likewise.
3342 (neon_vld4_lane<mode>): Likewise.
3343 (neon_vld4_dup<mode>): Likewise.
3344 (neon_vst4<mode>): Likewise.
3345 (neon_vst4qa<mode>): Likewise.
3346 (neon_vst4qb<mode>): Likewise.
3347 (neon_vst4_lane<mode>): Likewise.
3348 (neon_vst4_lane<mode>): Likewise.
3349 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3350 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3351 (neon_vec_<US>mult_lo_<mode>): Likewise.
3352 (neon_vec_<US>mult_hi_<mode>): Likewise.
3353 (neon_vec_<US>shiftl_<mode>): Likewise.
3354 (neon_unpack<US>_<mode>): Likewise.
3355 (neon_vec_<US>mult_<mode>): Likewise.
3356 (vec_pack_trunc_<mode>): Likewise.
3357 (neon_vec_pack_trunc_<mode>): Likewise.
3358 (neon_vabd<mode>_2): Likewise.
3359 (neon_vabd<mode>_3): Likewise.
3361 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3363 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
3364 (load_pair): Update type attribute.
3365 (store_pair): Update type attribute.
3366 * config/aarch64/iterators.md (q): New.
3368 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3370 * config/arm/types.md: Add new types for Neon insns.
3372 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3373 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3374 Sergey Lega <sergey.s.lega@intel.com>
3375 Anna Tikhonova <anna.tikhonova@intel.com>
3376 Ilya Tocar <ilya.tocar@intel.com>
3377 Andrey Turetskiy <andrey.turetskiy@intel.com>
3378 Ilya Verbin <ilya.verbin@intel.com>
3379 Kirill Yukhin <kirill.yukhin@intel.com>
3380 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3382 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
3383 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
3384 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
3386 (srcp14<mode>): Ditto.
3387 (rsqrt14<mode>): Ditto.
3388 (rsqrt14<mode>): Ditto.
3389 (avx512f_vmscalef<mode>): Ditto.
3390 (avx512f_scalef<mode>): Ditto.
3391 (avx512f_getexp<mode>): Ditto.
3392 (avx512f_sgetexp<mode>): Ditto.
3393 (avx512f_fixupimm<mode>): Ditto.
3394 (avx512f_sfixupimm<mode>): Ditto.
3395 (avx512f_rndscale<mode>): Ditto.
3396 (*avx512er_exp2<mode>): Ditto.
3397 (*avx512er_rcp28<mode>): Ditto.
3398 (avx512er_rsqrt28<mode>): Ditto.
3399 (avx512f_getmant<mode>): Ditto.
3400 (avx512f_getmant<mode>): Ditto.
3401 (avx512f_rndscale<mode>): Fix formatting.
3403 2013-10-15 Martin Jambor <mjambor@suse.cz>
3405 * ipa-utils.h (ipa_edge_within_scc): Declare.
3406 * ipa-cp.c (edge_within_scc): Moved...
3407 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
3409 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3410 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3411 Sergey Lega <sergey.s.lega@intel.com>
3412 Anna Tikhonova <anna.tikhonova@intel.com>
3413 Ilya Tocar <ilya.tocar@intel.com>
3414 Andrey Turetskiy <andrey.turetskiy@intel.com>
3415 Ilya Verbin <ilya.verbin@intel.com>
3416 Kirill Yukhin <kirill.yukhin@intel.com>
3417 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3419 * config/i386/predicates.md (const_8_to_15_operand): New.
3420 (const_16_to_31_operand): Ditto.
3421 * config/i386/sse.md (V8FI): New.
3423 (reduc_splus_v8df): Ditto.
3424 (reduc_splus_v16sf): Ditto.
3425 (avx512f_vextract<shuffletype>32x4_1): Ditto.
3426 (vec_extract_hi_<mode>): Ditto.
3427 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
3428 (vec_set_lo_<mode>): Ditto.
3429 (vec_set_hi_<mode>): Ditto.
3430 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
3431 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
3432 (avx512f_pshufd_1): Ditto.
3433 (avx512f_broadcast<mode>): Ditto.
3434 (avx512f_broadcast<mode>): Ditto.
3435 (define_split): Split vec_extract_lo into move.
3436 (ssequartermode): Ditto.
3437 (ssedoublemode): Extened with wider modes.
3438 (vec_extract_lo_<mode>): Ditto.
3440 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3441 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3442 Sergey Lega <sergey.s.lega@intel.com>
3443 Anna Tikhonova <anna.tikhonova@intel.com>
3444 Ilya Tocar <ilya.tocar@intel.com>
3445 Andrey Turetskiy <andrey.turetskiy@intel.com>
3446 Ilya Verbin <ilya.verbin@intel.com>
3447 Kirill Yukhin <kirill.yukhin@intel.com>
3448 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3450 * config/i386/predicates.md (register_or_constm1_operand): New.
3451 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
3452 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
3453 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
3454 UNSPEC_SCATTER_PREFETCH
3456 (avx512f_ucmp<mode>3): Ditto.
3457 (avx512f_vternlog<mode>): Ditto.
3458 (avx512f_align<mode>): Ditto.
3459 (<shift_insn><mode>3): Ditto.
3460 (avx512f_<rotate>v<mode>): Ditto.
3461 (avx512f_<rotate><mode>): Ditto.
3462 (avx512f_eq<mode>3): Ditto.
3463 (avx512f_eq<mode>3_1): Ditto.
3464 (avx512f_gt<mode>3): Ditto.
3465 (avx512f_testm<mode>3): Ditto.
3466 (avx512f_testnm<mode>3): Ditto.
3467 (avx512pf_gatherpf<mode>): Ditto.
3468 (*avx512pf_gatherpf<mode>_mask): Ditto.
3469 (*avx512pf_gatherpf<mode>): Ditto.
3470 (avx512pf_scatterpf<mode>): Ditto.
3471 (*avx512pf_scatterpf<mode>_mask): Ditto.
3472 (*avx512pf_scatterpf<mode>): Ditto.
3473 (avx512f_vec_dup_gpr<mode>): Ditto.
3474 (clz<mode>2): Ditto.
3475 (conflict<mode>): Ditto.
3476 (REDUC_SMINMAX_MODE): Extened with wider modes.
3477 (reduc_<code>_<mode>): Ditto.
3478 (vlshr<mode>3): Ditto.
3479 (vashl<mode>3): Ditto.
3481 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3482 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3483 Sergey Lega <sergey.s.lega@intel.com>
3484 Anna Tikhonova <anna.tikhonova@intel.com>
3485 Ilya Tocar <ilya.tocar@intel.com>
3486 Andrey Turetskiy <andrey.turetskiy@intel.com>
3487 Ilya Verbin <ilya.verbin@intel.com>
3488 Kirill Yukhin <kirill.yukhin@intel.com>
3489 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3491 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
3494 (avx512fmaskmode): Ditto.
3495 (bcstscalarsuff): Ditto.
3496 (avx512f_blendm<mode>): Ditto.
3497 (cmp_imm_predicate): Ditto.
3498 (avx512f_cmp<mode>3): Ditto.
3499 (avx512f_vec_dup<mode>): Ditto.
3500 (avx512f_vec_dup_mem<mode>): Ditto.
3501 (avx512f_vpermi2var<mode>3): Ditto.
3502 (avx512f_vpermt2var<mode>3): Ditto.
3503 (vec_init<mode>): Ditto.
3504 (avx512f_gathersi<mode>): Ditto.
3505 (*avx512f_gathersi<mode>): Ditto.
3506 (*avx512f_gathersi<mode>_2): Ditto.
3507 (avx512f_gatherdi<mode>): Ditto.
3508 (*avx512f_gatherdi<mode>): Ditto.
3509 (*avx512f_gatherdi<mode>_2): Ditto.
3510 (avx512f_scattersi<mode>): Ditto.
3511 (*avx512f_scattersi<mode>): Ditto.
3512 (avx512f_scatterdi<mode>): Ditto.
3513 (*avx512f_scatterdi<mode>): Ditto.
3514 (sseintprefix): Extened with wider modes.
3515 (VEC_GATHER_IDXSI): Ditto.
3516 (VEC_GATHER_IDXDI): Ditto.
3517 (VEC_GATHER_SRCDI): Ditto.
3519 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
3520 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3522 * config/arm/t-aprofile: New file.
3523 * config.gcc: Handle --with-multilib-list option.
3525 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
3527 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
3530 2013-10-15 Richard Biener <rguenther@suse.de>
3532 * tree-loop-distribution.c (build_empty_rdg): Inline into
3534 (rdg_flag_vertex): Inline into single user.
3535 (rdg_flag_vertex_and_dependent): Likewise.
3536 (build_rdg_partition_for_vertex): Remove processed bitmap.
3537 (rdg_build_partitions): Simplify.
3539 2013-10-15 Richard Biener <rguenther@suse.de>
3541 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
3542 Restructure forwarding through conversions and copies to
3543 avoid performing copy-propagation the wrong way. Adjust
3544 recursion invocations.
3545 (forward_propagate_addr_expr): Add argument stating if we
3546 are recursing from a single-use.
3547 (ssa_forward_propagate_and_combine): Adjust.
3549 2013-10-14 David Malcolm <dmalcolm@redhat.com>
3551 * dumpfile.h (gcc::dump_manager): New class, to hold state
3552 relating to dumpfile management.
3553 (get_dump_file_name): Remove in favor of method of dump_manager.
3554 (dump_initialized_p): Likewise.
3555 (dump_start): Likewise.
3556 (dump_finish): Likewise.
3557 (dump_switch_p): Likewise.
3558 (dump_register): Likewise.
3559 (get_dump_file_info): Likewise.
3560 * context.c (gcc::context::context): Construct the dump_manager
3562 * context.h (gcc::context::get_dumps): New.
3563 (gcc::context::m_dumps): New.
3564 * coverage.c (coverage_init): Port to dump_manager API.
3565 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
3566 (extra_dump_files_in_use): Likewise.
3567 (extra_dump_files_alloced): Likewise.
3568 (gcc::dump_manager::dump_manager): New.
3569 (dump_register): Convert to...
3570 (gcc::dump_manager::dump_register): ...method, replacing
3571 function-static next_dump with m_next_dump field.
3572 (get_dump_file_info): Convert to...
3573 (gcc::dump_manager::get_dump_file_info): ...method.
3574 (get_dump_file_name): Convert to...
3575 (gcc::dump_manager::get_dump_file_name): ...method.
3576 (dump_start): Convert to...
3577 (gcc::dump_manager::dump_start): ...method.
3578 (dump_finish): Convert to...
3579 (gcc::dump_manager::dump_finish): ...method.
3580 (dump_begin): Replace body with...
3581 (gcc::dump_manager::dump_begin): ...new method.
3582 (dump_phase_enabled_p): Convert to...
3583 (gcc::dump_manager::dump_phase_enabled_p): ...method.
3584 (dump_phase_enabled_p): Convert to...
3585 (gcc::dump_manager::dump_phase_enabled_p): ...method.
3586 (dump_initialized_p): Convert to...
3587 (gcc::dump_manager::dump_initialized_p): ...method.
3588 (dump_flag_name): Replace body with...
3589 (gcc::dump_manager::dump_flag_name): ...new method.
3590 (dump_enable_all): Convert to...
3591 (gcc::dump_manager::dump_enable_all): ...new method.
3592 (opt_info_enable_passes): Convert to...
3593 (gcc::dump_manager::opt_info_enable_passes): ...new method.
3594 (dump_switch_p_1): Convert to...
3595 (gcc::dump_manager::dump_switch_p_1): ...new method.
3596 (dump_switch_p): Convert to...
3597 (gcc::dump_manager::dump_switch_p): ...new method.
3598 (opt_info_switch_p): Port to dump_manager API.
3599 (enable_rtl_dump_file): Likewise.
3600 * opts-global.c (handle_common_deferred_options): Port to new
3602 * passes.c (pass_manager::finish_optimization_passes): Likewise.
3603 (pass_manager::register_one_dump_file): Likewise.
3604 (pass_manager::register_pass): Likewise.
3605 (pass_init_dump_file): Likewise.
3606 (pass_fini_dump_file): Likewise.
3607 * statistics.c (statistics_early_init): Likewise.
3609 2013-10-14 Richard Biener <rguenther@suse.de>
3611 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
3612 iterative_hash_canonical_type, gimple_canonical_type_hash,
3613 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
3614 gimple_register_canonical_type, print_gimple_types_stats,
3615 free_gimple_type_tables): Move to lto/lto.c
3616 (gt-gimple.h): Do not include.
3617 * gimple.h (gimple_register_canonical_type,
3618 print_gimple_types_stats, free_gimple_type_tables): Remove.
3619 * Makefile.in (GTFILES): Remove gimple.c.
3621 2013-10-14 Travis Snoozy <quandary@remstate.com>
3624 * config/msp430/msp430.c (msp430_option_override): Correct thinko
3625 scanning for msp430x targets.
3627 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
3630 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
3631 (registers_ok_for_ldd_peep): Move around.
3632 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
3633 * config/sparc/sparc.md (widening peepholes): Use it.
3635 2013-10-14 Richard Biener <rguenther@suse.de>
3639 * gimple.c (iterative_hash_canonical_type): Make sure to
3640 record the hash into the correct hashtable slot.
3642 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
3644 PR rtl-optimization/58662
3645 * combine.c (try_combine): Take into account death nodes on I2 when
3646 splitting a PARALLEL of two independent SETs. Fix dump message.
3648 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
3651 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
3652 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
3653 * config/sh/t-sh (sh_treg_combine.o): New entry.
3654 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
3655 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
3656 (register_sh_passes): New function. Register sh_treg_combine pass.
3657 (sh_option_override): Invoke it.
3658 (sh_canonicalize_comparison): Handle op0_preserve_value.
3659 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
3660 opportunities. Canonicalize branch condition.
3661 (nott): Allow only if pseudos can be created for non-SH2A.
3663 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
3666 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
3667 (ix86_expand_movmem): Replace copy_addr_to_reg with
3668 ix86_copy_addr_to_reg.
3669 (ix86_expand_setmem): Likewise.
3671 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
3673 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
3674 provided by ix86_fp_compare_mode instead of CCFPUmode.
3676 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
3678 * config/aarch64/arm_neon.h
3679 (vtbx<1,3>_<psu>8): Fix register constriants.
3681 2013-10-11 Jeff Law <law@redhat.com>
3683 PR tree-optimization/58640
3684 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
3685 threading paths that cross over two loop entry points.
3687 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3689 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
3690 handle vector float as well.
3691 (*vsx_le_perm_load_v4si): Likewise.
3692 (*vsx_le_perm_store_v2di): Likewise.
3693 (*vsx_le_perm_store_v4si): Likewise.
3695 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3697 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
3698 directly to circumvent subtract from splat{31} workaround.
3699 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
3701 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
3702 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
3703 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
3704 (altivec_vperm_<mode>): Convert to define_insn_and_split to
3705 separate big and little endian logic.
3706 (*altivec_vperm_<mode>_internal): New define_insn.
3707 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
3708 separate big and little endian logic.
3709 (*altivec_vperm_<mode>_uns_internal): New define_insn.
3710 (vec_permv16qi): Add little endian logic.
3712 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
3714 * doc/extend.texi (returns_nonnull): Remove arguments.
3716 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3717 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3718 Sergey Lega <sergey.s.lega@intel.com>
3719 Anna Tikhonova <anna.tikhonova@intel.com>
3720 Ilya Tocar <ilya.tocar@intel.com>
3721 Andrey Turetskiy <andrey.turetskiy@intel.com>
3722 Ilya Verbin <ilya.verbin@intel.com>
3723 Kirill Yukhin <kirill.yukhin@intel.com>
3724 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3726 * config/i386/sse.md (VI48F_256_512): New.
3727 (avx2_permvar<mode>): Change to ...
3728 (<avx2_avx512f>_permvar<mode>): This.
3730 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3731 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3732 Sergey Lega <sergey.s.lega@intel.com>
3733 Anna Tikhonova <anna.tikhonova@intel.com>
3734 Ilya Tocar <ilya.tocar@intel.com>
3735 Andrey Turetskiy <andrey.turetskiy@intel.com>
3736 Ilya Verbin <ilya.verbin@intel.com>
3737 Kirill Yukhin <kirill.yukhin@intel.com>
3738 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3740 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
3741 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
3742 * config/i386/sse.md (VI4_AVX): New.
3743 (sf2simodelower): Ditto.
3744 (sse2_cvtps2dq): Change to ...
3745 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
3747 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3748 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3749 Sergey Lega <sergey.s.lega@intel.com>
3750 Anna Tikhonova <anna.tikhonova@intel.com>
3751 Ilya Tocar <ilya.tocar@intel.com>
3752 Andrey Turetskiy <andrey.turetskiy@intel.com>
3753 Ilya Verbin <ilya.verbin@intel.com>
3754 Kirill Yukhin <kirill.yukhin@intel.com>
3755 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3757 * config/i386/sse.md (V_512): New.
3759 (vcond<V_512:mode><VF_512:mode>): Ditto.
3760 (vcond<V_512:mode><VI_512:mode>): Ditto.
3761 (vcondu<V_512:mode><VI_512:mode>): Ditto.
3763 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3764 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3765 Sergey Lega <sergey.s.lega@intel.com>
3766 Anna Tikhonova <anna.tikhonova@intel.com>
3767 Ilya Tocar <ilya.tocar@intel.com>
3768 Andrey Turetskiy <andrey.turetskiy@intel.com>
3769 Ilya Verbin <ilya.verbin@intel.com>
3770 Kirill Yukhin <kirill.yukhin@intel.com>
3771 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3773 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
3774 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
3776 (fma<mode>4): Removed condition.
3777 (fms<mode>4): Ditto.
3778 (fnma<mode>4): Ditto.
3779 (fnms<mode>4): Ditto.
3780 (fma4i_fmadd_<mode>): Ditto.
3781 (*fma_fmadd_<mode>): Ditto.
3782 (*fma_fmsub_<mode>): Ditto.
3783 (*fma_fnmadd_<mode>): Ditto.
3784 (*fma_fnmsub_<mode>): Ditto.
3785 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
3786 (*fma_fmaddsub_<mode>): Ditto.
3787 (*fma_fmsubadd_<mode>): Ditto.
3788 (*fmai_fmadd_<mode>): Ditto.
3789 (*fmai_fmsub_<mode>): Ditto.
3790 (*fmai_fnmadd_<mode>): Ditto.
3791 (*fmai_fnmsub_<mode>): Ditto.
3793 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3794 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3795 Sergey Lega <sergey.s.lega@intel.com>
3796 Anna Tikhonova <anna.tikhonova@intel.com>
3797 Ilya Tocar <ilya.tocar@intel.com>
3798 Andrey Turetskiy <andrey.turetskiy@intel.com>
3799 Ilya Verbin <ilya.verbin@intel.com>
3800 Kirill Yukhin <kirill.yukhin@intel.com>
3801 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3803 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
3804 (VI124_256): Changed to ...
3805 (VI124_256_48_512): This.
3806 (ssepackmode): Extended with wider modes.
3807 (<code><mode>3): Changed iterator.
3808 (*avx2_<code><mode>3): Ditto.
3809 (vec_pack_trunc_<mode>): Ditto.
3811 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3812 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3813 Sergey Lega <sergey.s.lega@intel.com>
3814 Anna Tikhonova <anna.tikhonova@intel.com>
3815 Ilya Tocar <ilya.tocar@intel.com>
3816 Andrey Turetskiy <andrey.turetskiy@intel.com>
3817 Ilya Verbin <ilya.verbin@intel.com>
3818 Kirill Yukhin <kirill.yukhin@intel.com>
3819 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3821 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
3822 (VI8F_256_512): Ditto.
3823 (abs<mode>2): Changed iterator.
3824 (avx2_perm<mode>): Changed to ...
3825 (<avx2_avx512f>_perm<mode>): This.
3826 (avx2_perm<mode>_1): Changed to ...
3827 (<avx2_avx512f>_perm<mode>_1): This.
3829 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3830 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3831 Sergey Lega <sergey.s.lega@intel.com>
3832 Anna Tikhonova <anna.tikhonova@intel.com>
3833 Ilya Tocar <ilya.tocar@intel.com>
3834 Andrey Turetskiy <andrey.turetskiy@intel.com>
3835 Ilya Verbin <ilya.verbin@intel.com>
3836 Kirill Yukhin <kirill.yukhin@intel.com>
3837 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3839 * config/i386/sse.md (VI48_AVX512F): New.
3840 (VI48_AVX2): Changed to ...
3841 (VI48_AVX2_48_AVX512F): This.
3842 (avx2_ashrv<mode>): Changed to ...
3843 (<avx2_avx512f>_ashrv<mode>): This.
3844 (avx2_<shift_insn>v<mode>): Changed to ...
3845 (<avx2_avx512f>_<shift_insn>v<mode>): This.
3847 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3848 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3849 Sergey Lega <sergey.s.lega@intel.com>
3850 Anna Tikhonova <anna.tikhonova@intel.com>
3851 Ilya Tocar <ilya.tocar@intel.com>
3852 Andrey Turetskiy <andrey.turetskiy@intel.com>
3853 Ilya Verbin <ilya.verbin@intel.com>
3854 Kirill Yukhin <kirill.yukhin@intel.com>
3855 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3857 * config/i386/sse.md (VI4_AVX512F): New.
3858 (VI8_AVX2_AVX512F): Ditto.
3859 (mul<mode>3): Extended with wider modes.
3860 (*<sse4_1_avx2>_mul<mode>3): Ditto.
3861 (mul<mode>3): Ditto.
3862 (vec_widen_<s>mult_odd_<mode>): Ditto.
3864 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3865 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3866 Sergey Lega <sergey.s.lega@intel.com>
3867 Anna Tikhonova <anna.tikhonova@intel.com>
3868 Ilya Tocar <ilya.tocar@intel.com>
3869 Andrey Turetskiy <andrey.turetskiy@intel.com>
3870 Ilya Verbin <ilya.verbin@intel.com>
3871 Kirill Yukhin <kirill.yukhin@intel.com>
3872 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3874 * config/i386/sse.md (VI2_AVX512F): New.
3875 (VI124_AVX512F): Ditto.
3876 (sseunpackmode): Extended with wider modes.
3877 (sseunpackfltmode): Ditto.
3878 (vec_unpacks_float_hi_<mode>): Ditto.
3879 (vec_unpacks_float_lo_<mode>): Ditto.
3880 (vec_unpacku_float_hi_<mode>): Ditto.
3881 (vec_unpacku_float_lo_<mode>): Ditto.
3882 (vec_unpacks_lo_<mode>): Ditto.
3883 (vec_unpacks_hi_<mode>): Ditto.
3884 (vec_unpacku_lo_<mode>): Ditto.
3885 (vec_unpacku_hi_<mode>): Ditto.
3887 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3888 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3889 Sergey Lega <sergey.s.lega@intel.com>
3890 Anna Tikhonova <anna.tikhonova@intel.com>
3891 Ilya Tocar <ilya.tocar@intel.com>
3892 Andrey Turetskiy <andrey.turetskiy@intel.com>
3893 Ilya Verbin <ilya.verbin@intel.com>
3894 Kirill Yukhin <kirill.yukhin@intel.com>
3895 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3897 * config/i386/i386.md (multdiv): New.
3898 (multdiv_mnemonic): Ditto.
3899 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
3900 (<sse>_vm<multdiv_mnemonic><mode>3): This.
3901 (<sse>_vmdiv<mode>3): Removed.
3903 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3904 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3905 Sergey Lega <sergey.s.lega@intel.com>
3906 Anna Tikhonova <anna.tikhonova@intel.com>
3907 Ilya Tocar <ilya.tocar@intel.com>
3908 Andrey Turetskiy <andrey.turetskiy@intel.com>
3909 Ilya Verbin <ilya.verbin@intel.com>
3910 Kirill Yukhin <kirill.yukhin@intel.com>
3911 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3913 * config/i386/sse.md (V): Extended with wider modes.
3915 (ssehalfvecmode): Ditto.
3917 (ssepackfltmode): Ditto.
3918 (avx_vec_concat<mode>): Ditto.
3919 (V_256_512): New iterator.
3920 (VF2_512_256): Ditto.
3921 (si2dfmode): New attribute.
3922 (si2dfmodelower): Ditto.
3924 (concat_tg_mode): Ditto.
3925 (floatv4siv4df2): Changed to ...
3926 (float<si2dfmodelower><mode>2): This.
3927 (avx_cvtps2pd256): Changed to ...
3928 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
3929 (vec_pack_trunc_v4df): Changed to ...
3930 (vec_pack_trunc_<mode>): This.
3931 (avx_vpermil<mode>): Changed to ...
3932 (<sse2_avx_avx512f>_vpermil<mode>): This.
3933 (<fixsuffix>fix_truncv8dfv8si2): New.
3934 (vec_pack_sfix_trunc_v8df): Ditto.
3935 (avx512f_rndscale<mode>): Ditto.
3936 (avx512f_roundpd512): Ditto.
3937 (vec_pack_ufix_trunc_<mode>): Updated iterator.
3939 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3940 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3941 Sergey Lega <sergey.s.lega@intel.com>
3942 Anna Tikhonova <anna.tikhonova@intel.com>
3943 Ilya Tocar <ilya.tocar@intel.com>
3944 Andrey Turetskiy <andrey.turetskiy@intel.com>
3945 Ilya Verbin <ilya.verbin@intel.com>
3946 Kirill Yukhin <kirill.yukhin@intel.com>
3947 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3949 * config/i386/i386.md (any_fix): New iterator.
3950 (fixsuffix): New attribute.
3951 * config/i386/sse.md (VF1): Extened with wider modes.
3955 (sseintvecmodelower): Ditto.
3956 (ssescalarmode): Ditto.
3957 (ssescalarnum): Ditto.
3959 (ssexmmmode): Ditto.
3960 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
3961 (<sse>_rcp<mode>2): Change iterator.
3962 (rsqrt<mode>2): Ditto.
3963 (<sse>_rsqrt<mode>2): Ditto.
3964 (avx2_vec_dup<mode>): Ditto.
3965 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
3966 (round<mode>2_sfix): Ditto.
3967 (avx2_pbroadcast<mode>): Ditto.
3968 (*andnot<mode>3): Handle XI mode.
3969 (*<code><mode>3): Ditto.
3970 (AVXTOSSEMODE): Removed.
3971 (avx_vpermil<mode>): Changed to ...
3972 (<sse2_avx_avx512f>_vpermil<mode>): This.
3974 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3975 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3976 Sergey Lega <sergey.s.lega@intel.com>
3977 Anna Tikhonova <anna.tikhonova@intel.com>
3978 Ilya Tocar <ilya.tocar@intel.com>
3979 Andrey Turetskiy <andrey.turetskiy@intel.com>
3980 Ilya Verbin <ilya.verbin@intel.com>
3981 Kirill Yukhin <kirill.yukhin@intel.com>
3982 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3984 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
3985 (<sse>_comi): Ditto.
3986 (<sse>_ucomi): Ditto.
3987 (sse_cvtss2siq_2): Ditto.
3988 (sse2_cvtsd2si): Ditto.
3989 (sse2_cvtsd2siq): Ditto.
3990 (sse2_cvttsd2si): Ditto.
3991 (sse2_cvttsd2siq): Ditto.
3992 (<shift_insn><mode>3): Ditto.
3993 (sse2_cvtsi2sdq): Update constraint and prefix.
3994 (sse_cvtsi2ss): Update prefix.
3995 (sse_cvtsi2ssq): Ditto.
3997 2013-10-11 Jakub Jelinek <jakub@redhat.com>
3999 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4000 ignore internal calls.
4002 2013-10-11 Richard Biener <rguenther@suse.de>
4004 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4005 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4006 TREE_OVERFLOW is set.
4008 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4010 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4012 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4013 * gimple.h: Likewise.
4015 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4017 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4020 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4022 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4023 * configure: Regenerate.
4025 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4027 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4028 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4029 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4030 similarly pp_right_paren.
4031 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4032 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4033 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4034 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4036 (omp_declare_simd_clauses_equal,
4037 omp_remove_redundant_declare_simd_attrs): New functions.
4038 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4039 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4040 * tree.h (OMP_LOOP_CHECK): Define.
4041 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4042 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4043 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4044 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4045 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4046 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4047 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4048 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4049 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4050 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4051 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4052 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4053 (omp_remove_redundant_declare_simd_attrs): New prototype.
4054 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4055 GIMPLE_OMP_TEAMS): New codes.
4056 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4057 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4059 (target_nesting_level): New variable.
4060 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4061 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4062 for collapse > 1 static schedule unless ordered.
4063 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4064 (determine_parallel_type): Adjust get_ws_args_for caller.
4065 (install_var_field): Handle mask & 4 for double indirection.
4066 (scan_sharing_clauses): Ignore shared clause on teams construct.
4067 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4068 (create_omp_child_function): If inside target or declare target
4069 constructs, set "omp declare target" attribute on the child function.
4070 (find_combined_for): New function.
4071 (scan_omp_parallel): Handle combined loops.
4072 (scan_omp_target, scan_omp_teams): New functions.
4073 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4074 restrictions and set ctx->cancellable for cancellable constructs.
4075 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4076 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4077 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4078 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4079 (omp_clause_aligned_alignment): New function.
4080 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4081 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4082 on teams constructs. Handle user defined reductions and new
4084 (lower_reduction_clauses): Don't set placeholder to address of ref
4085 if it has already the right type.
4086 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4087 (expand_parallel_call): Use the new non-_start suffixed builtins,
4088 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4089 and GOMP_parallel_end after the call.
4090 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4091 (expand_omp_for_init_counts): Handle combined loops.
4092 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4094 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4095 end of cancellable loops.
4096 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4097 Likewise. Handle collapse > 1 loops.
4098 (expand_omp_simd): Handle combined loops.
4099 (expand_omp_for): Add inner_stmt argument, adjust callers of
4100 expand_omp_for* functions, use expand_omp_for_static*chunk even
4101 for collapse > 1 unless ordered.
4102 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4103 of cancellable sections.
4104 (expand_omp_single): Remove need_barrier variable, just rely on
4105 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4106 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4107 (expand_omp_atomic_load, expand_omp_atomic_store,
4108 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4109 (expand_omp_target): New function.
4110 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4111 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4112 (build_omp_regions_1): Immediately close region for
4113 GF_OMP_TARGET_KIND_UPDATE.
4114 (maybe_add_implicit_barrier_cancel): New function.
4115 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4117 (lower_omp_single): Likewise. Add clobber after the barrier.
4118 (lower_omp_taskgroup): New function.
4119 (lower_omp_for): Handle combined loops. Adjust
4120 lower_rec_input_clauses caller. Handle cancellation.
4121 (lower_depend_clauses): New function.
4122 (lower_omp_taskreg): Lower depend clauses. Adjust
4123 lower_rec_input_clauses caller. Add clobber after the call. Handle
4125 (lower_omp_target, lower_omp_teams): New functions.
4126 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4127 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4128 and GOMP_cancellation_point calls.
4129 (lower_omp): Fold stmts inside of target region.
4130 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4131 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4132 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4133 (BT_FN_VOID_OMPFN_PTR_UINT,
4134 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4135 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4136 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4137 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4138 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4139 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4140 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4141 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4142 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4143 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4144 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4145 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4146 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4147 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4148 BUILT_IN_GOMP_PARALLEL_END.
4149 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4150 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4151 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4152 GF_OMP_FOR_KIND_DISTRIBUTE.
4153 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4154 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4155 (dump_gimple_omp_return): Print lhs if it has any.
4156 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4157 gimple_omp_atomic_seq_cst_p.
4158 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4159 and GIMPLE_OMP_TEAMS.
4160 * langhooks.c (lhd_omp_mappable_type): New function.
4161 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4162 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4164 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4165 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4166 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4167 (struct gimplify_omp_ctx): Add combined_loop field.
4168 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4169 on stmts inside of target region.
4170 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4171 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4172 ORT_TARGET and ORT_TARGET_DATA.
4173 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4175 (omp_notice_threadprivate_variable): Complain about threadprivate
4176 variables in target region.
4177 (omp_notice_variable): Complain about vars with non-mappable type
4178 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4179 (omp_check_private): Ignore ORT_TARGET* regions.
4180 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4181 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
4182 (find_combined_omp_for): New function.
4183 (gimplify_omp_for): Handle gimplification of combined loops.
4184 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
4186 (gimplify_omp_target_update): New function.
4187 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
4188 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
4189 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
4190 (gimplify_body): If fndecl has "omp declare target" attribute, add
4191 implicit ORT_TARGET context around it.
4192 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
4193 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
4194 * tree-nested.c (convert_nonlocal_reference_stmt,
4195 convert_local_reference_stmt, convert_gimple_call): Handle
4196 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4197 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
4198 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
4199 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
4200 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
4201 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
4202 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
4203 BUILT_IN_GOMP_LOOP_END_CANCEL,
4204 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
4205 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
4206 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
4207 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
4208 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
4209 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
4210 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
4211 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
4212 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
4213 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
4214 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
4215 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
4216 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
4217 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
4218 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
4219 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4220 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
4221 gimple_build_omp_teams): New functions.
4222 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
4223 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
4224 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
4225 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4226 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
4227 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
4228 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
4229 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
4230 GF_OMP_ATOMIC_SEQ_CST): New.
4231 (gimple_build_omp_taskgroup, gimple_build_omp_target,
4232 gimple_build_omp_teams): New prototypes.
4233 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
4234 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4235 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
4236 GIMPLE_OMP_SINGLE as end of range.
4237 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
4238 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
4239 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
4240 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
4241 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
4242 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
4243 gimple_omp_target_kind, gimple_omp_target_set_kind,
4244 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
4245 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
4246 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
4247 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
4248 gimple_omp_teams_set_clauses): New inlines.
4249 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
4250 and GIMPLE_OMP_TASKGROUP.
4251 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
4252 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
4253 enum omp_clause_proc_bind_kind): New.
4254 (union omp_clause_subcode): Add depend_kind, map_kind and
4255 proc_bind_kind fields.
4256 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
4257 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4258 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
4259 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
4260 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
4262 2013-10-10 Teresa Johnson <tejohnson@google.com>
4264 * predict.c (tree_estimate_probability): Add new parameter
4265 for estimate_bb_frequencies.
4266 (estimate_bb_frequencies): Add new parameter to force estimation.
4267 (rebuild_frequencies): When max frequency in function is small,
4268 recompute counts from frequencies.
4269 * predict.h (estimate_bb_frequencies): New parameter.
4271 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4273 * ipa-inline.c (ipa_inline): Fix leak of "order" when
4274 optimizations are disabled.
4276 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4278 * coverage.c (coverage_finish): Fix leak of da_file_name.
4280 2013-10-10 Jan Hubicka <jh@suse.cz>
4282 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
4283 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
4286 2013-10-10 Jakub Jelinek <jakub@redhat.com>
4289 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
4290 if any labels are in FALLTHRU_BB, use a special label emitted
4291 immediately after the asm goto insn rather than label_rtx
4293 (expand_asm_stmt): Adjust caller.
4294 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
4295 edge if the last insn in predecessor is a jump with single successor,
4296 but it isn't simplejump_p.
4298 2013-10-10 Richard Biener <rguenther@suse.de>
4300 PR tree-optimization/58656
4301 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
4303 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4305 * gimplify.c: Include expr.h and tm_p.h for targets with special
4306 va-arg padding requirements.
4308 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4310 * tree-flow.h: Move some prototypes to gimple.h.
4311 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
4312 * gimple.h: Relocate some prototypes from tree-flow.h
4313 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
4315 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
4316 (build_va_arg_indirect_ref): Relocate and make static.
4317 (std_gimplify_va_arg_expr): Relocate here.
4318 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
4319 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
4321 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4323 * doc/md.texi: Document the mnemonic attribute.
4325 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4328 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
4329 (set_attr_alternative x ...) when searching for user defined
4332 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4334 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
4336 2013-10-09 Easwaran Raman <eraman@google.com>
4338 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
4339 * cfgexpand.c (defer_stack_allocation): ...use here
4340 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
4342 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
4344 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
4345 (operand_equal_for_value_replacement): New function, extracted from
4346 value_replacement and enhanced to catch more cases.
4347 (value_replacement): Use operand_equal_for_value_replacement.
4349 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4351 * loop-doloop.c (doloop_modify, doloop_optimize): Use
4352 get_max_loop_iterations.
4354 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4356 * config/arm/aarch-common.c (arm_early_load_addr_dep):
4357 Place comment above function.
4359 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4361 * tree-flow.h: Remove all remaining prototypes, enums and structs that
4362 are not related to tree-cfg.c.
4363 * tree-ssa-address.h: New file. Relocate prototypes.
4364 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
4365 (addr_for_mem_ref): New. Combine call to get_address_description and
4366 return addr_for_mem_ref.
4367 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
4368 * tree-ssa-live.h: Adjust prototypes.
4369 * passes.c: Include tree-ssa-live.h.
4370 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
4371 * graphite.c (graphite_transform_loops): Make static.
4372 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
4373 make_pass_graphite, pass_data_graphite_transforms,
4374 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
4375 * ipa-pure-const.c (warn_function_noreturn): Make static.
4376 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4377 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4378 Relocate from tree-cfg.c
4379 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
4381 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4382 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4383 Move to ipa-pure-const.c.
4384 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
4385 Relocate from tree-optimize.c.
4386 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
4387 make_pass_fixup_cfg): Move to tree-cfg.c.
4388 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
4389 Relocate some prototypes.
4390 * tree-data-ref.h (tree_check_data_deps) Add prototype.
4391 * tree-dump.c (dump_function_to_file): Remove prototype.
4392 Add tree-flow.h to the include file.
4393 * tree-dump.h: Remove prototype.
4394 * tree-parloops.h: New File. Add prototypes.
4395 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
4396 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
4397 from tree-ssa-loop.c.
4398 * tree-predcom.c (run_tree_predictive_commoning,
4399 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
4400 Relocate here from tree-ssa-loop.c.
4401 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
4402 ssa_name_values.release ().
4403 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
4404 (ssa_name_values): Relocate from tree-flow.h.
4405 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
4406 * tree-ssa-loop.c (run_tree_predictive_commoning,
4407 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
4408 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
4409 make_pass_graphite, pass_data_graphite_transforms,
4410 make_pass_graphite_transforms, gate_tree_parallelize_loops,
4411 tree_parallelize_loops, pass_data_parallelize_loops,
4412 make_pass_parallelize_loops): Move to other files.
4413 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
4415 * tree.h: Remove prototypes from tree-address.c.
4417 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4419 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
4420 (struct int_tree_map): Move to tree-hasher.h
4421 (SCALE, LABEL, PERCENT): Move to gimple.h
4422 * tree-flow-inline.h: Delete. Move functions to other files.
4423 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
4424 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
4425 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
4426 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
4427 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
4428 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
4429 * tree-hasher.h: Don't include tree-flow.h.
4430 (struct int_tree_map): Relocate from tree-flow.h.
4431 * tree-sra.c (contains_view_convert_expr_p): Relocate from
4432 tree-flow-inline.h and make static.
4433 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
4435 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
4436 tree-flow-inline.h and make static.
4437 * tree.h (is_global_var, may_be_aliased): Relocate from
4439 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
4440 * value-prof.c: No longer include tree-flow-inline.h.
4441 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
4443 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4445 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
4446 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
4447 (enum move_pos): Move to tree-ssa-loop-im.h
4448 * cfgloop.h: Move some prototypes.
4449 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
4450 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
4451 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
4452 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
4453 (loop_containing_stmt): Relocate from tree-flow-inline.h.
4454 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
4455 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
4456 (enum move_pos): Relocate here.
4457 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
4459 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
4460 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
4461 make_pass_lim): Relocate here from tree-ssa-loop.c.
4462 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
4464 (loop_edge_to_cancel, unloop_loops): Make static.
4465 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
4466 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
4467 (tree_complete_unroll, gate_tree_complete_unroll,
4468 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
4469 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
4470 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
4471 * tree-ssa-loop-ivopts.c: Remove local prototypes.
4472 (stmt_invariant_in_loop_p): Remove unused function.
4473 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
4474 * tree-ssa-loop-manip.h: New file. Add prototypes.
4475 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
4476 (gcov_type_to_double_int): Move to cfgloop.h.
4477 (double_int_cmp, bound_index,
4478 estimate_numbers_of_iterations_loop): Make static.
4479 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
4480 (max_loop_iterations): Factor out get_max_loop_iterations.
4481 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
4483 * tree-ssa-loop-niter.h: New file. Add prototypes.
4484 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
4485 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
4486 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
4487 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
4488 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
4489 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
4490 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
4491 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
4492 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
4493 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
4494 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
4495 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
4496 pass_data_complete_unroll, make_pass_complete_unroll,
4497 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
4498 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
4499 tree-ssa-loop-ivcanon.c.
4500 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
4501 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
4502 tree-ssa-loop-prefetch.c.
4503 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
4505 (get_lsm_tmp_name): Relocate and add suffix parameter.
4506 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
4507 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
4508 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
4509 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
4510 (get_estimated_loop_iterations): Factor out accessor from
4511 estimated_loop_iterations in tree-ssa-loop-niter.c.
4512 (get_max_loop_iterations): Factor out accessor from
4513 _max_loop_iterations in tree-ssa-niter.c.
4514 * loop-unroll.c (decide_unroll_constant_iterations,
4515 decide_unroll_runtime_iterations, decide_peel_simple,
4516 decide_unroll_stupid): Use new get_* accessors.
4518 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
4520 PR tree-optimization/20318
4521 * doc/extend.texi (returns_nonnull): New function attribute.
4522 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
4524 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
4525 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
4527 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
4530 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
4531 false if both components are bitfields.
4533 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4535 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
4536 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
4537 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
4538 * config/aarch64/aarch64.h
4539 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
4540 * config/aarch64/aarch64-simd-builtins.def
4541 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
4543 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4545 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
4547 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4549 * config/aarch64/arm_neon.h (vdiv_f64): Added.
4551 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4553 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
4554 (vneg_f64): New intrinsic.
4555 (vneg_s8): Asm replaced with C.
4556 (vneg_s16): Likewise.
4557 (vneg_s32): Likewise.
4558 (vneg_s64): New intrinsic.
4559 (vnegq_f32): Asm replaced with C.
4560 (vnegq_f64): Likewise.
4561 (vnegq_s8): Likewise.
4562 (vnegq_s16): Likewise.
4563 (vnegq_s32): Likewise.
4564 (vnegq_s64): Likewise.
4566 2013-10-09 Renlin Li <Renlin.Li@arm.com>
4568 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
4570 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4572 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
4573 packed stack special handling.
4574 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
4575 back to fixed stack slots for FPRs saved due to stdarg.
4577 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4579 * config/s390/s390.c (s390_frame_info): Restructure function.
4581 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4583 * config/s390/s390.c (struct s390_frame_layout): New field
4585 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
4586 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
4587 regs_ever_clobbered to char*.
4588 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
4589 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
4590 (s390_register_info): Enable FPR save slots. Move/Copy some
4591 functionality into ...
4592 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
4593 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
4595 (s390_frame_info): Do gpr slot allocation here now. stdarg does
4596 not imply a stack frame.
4597 (s390_init_frame_layout): Remove variable clobbered_regs.
4598 (s390_update_register_info): Remove function.
4599 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
4600 cannot be used for register renaming.
4601 (s390_hard_regno_scratch_ok): New function.
4602 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
4603 (s390_initial_elimination_offset): Change offset calculation of
4604 the return address pointer.
4605 (save_gprs): Deal with only r6 being saved from the call-saved regs.
4606 (restore_gprs): Set frame related flag.
4607 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
4608 (s390_emit_prologue): Call s390_register_info instead of
4609 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
4610 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
4611 (s390_optimize_prologue): Call s390_optimize_register_info.
4612 Try to remove also FPR slot save/restore INSNs. Remove frame
4613 related flags from restore INSNs.
4615 2013-10-08 DJ Delorie <dj@redhat.com>
4617 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
4620 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
4621 avoid conflict with the MI use of %c.
4622 * config/rl78/rl78-real.md: change %c to %C throughout.
4623 * config/rl78/rl78-virt.md: Likewise.
4625 2013-10-08 Jan Hubicka <jh@suse.cz>
4627 * config/i386/i386.c (ix86_option_override_internal): Switch
4628 to SSE math for -ffast-math when target ISA supports SSE2.
4630 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
4632 * tree-flow.h: Remove some prototypes.
4633 * tree.h: Remove some protypes, add a couple.
4634 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
4635 using_eh_for_cleanups_p): Add interface routines for front ends.
4636 * tree-eh.h: New file. Add protoptyes.
4637 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
4638 (add_stmt_to_eh_lp_fn): Make static.
4639 (lower_try_finally): Use new using_eh_for_cleanups_p.
4640 * emit-rtl.c: Include tree-eh.h.
4641 * gimple.h: Include tree-eh.h.
4643 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
4645 PR tree-optimization/58480
4646 * tree-vrp.c (infer_nonnull_range): New function.
4647 (infer_value_range): Call infer_nonnull_range.
4649 2013-10-08 Dehao Chen <dehao@google.com>
4651 PR tree-optimization/58619
4652 * tree-inline.c (copy_phis_for_bb): Combine location data
4655 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4658 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
4659 RTX_FRAME_RELATED_P on INSN.
4661 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4663 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
4664 (altivec_expand_vec_perm_const): Call it.
4666 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4668 * config/rs6000/vector.md (mov<mode>): Emit permuted move
4669 sequences for LE VSX loads and stores at expand time.
4670 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
4672 * config/rs6000/rs6000.c (rs6000_const_vec): New.
4673 (rs6000_gen_le_vsx_permute): New.
4674 (rs6000_gen_le_vsx_load): New.
4675 (rs6000_gen_le_vsx_store): New.
4676 (rs6000_gen_le_vsx_move): New.
4677 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
4678 (*vsx_le_perm_load_v4si): New.
4679 (*vsx_le_perm_load_v8hi): New.
4680 (*vsx_le_perm_load_v16qi): New.
4681 (*vsx_le_perm_store_v2di): New.
4682 (*vsx_le_perm_store_v4si): New.
4683 (*vsx_le_perm_store_v8hi): New.
4684 (*vsx_le_perm_store_v16qi): New.
4685 (*vsx_xxpermdi2_le_<mode>): New.
4686 (*vsx_xxpermdi4_le_<mode>): New.
4687 (*vsx_xxpermdi8_le_V8HI): New.
4688 (*vsx_xxpermdi16_le_V16QI): New.
4689 (*vsx_lxvd2x2_le_<mode>): New.
4690 (*vsx_lxvd2x4_le_<mode>): New.
4691 (*vsx_lxvd2x8_le_V8HI): New.
4692 (*vsx_lxvd2x16_le_V16QI): New.
4693 (*vsx_stxvd2x2_le_<mode>): New.
4694 (*vsx_stxvd2x4_le_<mode>): New.
4695 (*vsx_stxvd2x8_le_V8HI): New.
4696 (*vsx_stxvd2x16_le_V16QI): New.
4698 2013-10-07 Renlin Li <Renlin.Li@arm.com>
4700 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
4702 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4704 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
4705 loop to work also for 31bit ABI.
4706 Save the stack pointer for frame_size > 0.
4708 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4710 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
4711 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
4712 constraint letters from expanders.
4713 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
4714 retry count to general_operand.
4715 ("tabort"): Give operand 0 a mode.
4716 ("tabort_1"): Add mode and constraint letter for operand 0.
4717 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
4719 2013-10-04 Jeff Law <law@redhat.com>
4721 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
4723 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
4725 (thread_across_edge): Here. Call it.
4727 2013-10-04 Cary Coutant <ccoutant@google.com>
4729 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
4730 discarding a location list expression (or a piece of one).
4732 2013-10-03 Jan Hubicka <jh@suse.cz>
4734 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
4735 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
4736 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
4738 2013-10-03 Jan Hubicka <jh@suse.cz>
4740 * config/i386/i386.c (ix86_option_override_internal): Do not enable
4741 accumulate-outgoing-args when producing unwind info.
4743 2013-10-03 Wei Mi <wmi@google.com>
4745 * lra-constraints.c (insert_move_for_subreg): New function
4746 extracted from simplify_operand_subreg.
4747 (simplify_operand_subreg): Add reload for paradoxical subreg.
4749 2013-10-03 Rong Xu <xur@google.com>
4751 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
4752 the candidate of builtin_expect such that we should fix the
4753 size/time estimation.
4754 (estimate_function_body_sizes): Do the acutally size/time fix-up
4757 2013-10-03 Rong Xu <xur@google.com>
4759 * predict.c (tree_predict_by_opcode): Get the probability
4760 for builtin_expect from param builtin_expect_probability.
4761 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
4762 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
4763 * doc/invoke.texi: Add documentation for builtin-expect-probability.
4765 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
4768 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
4769 * calls.c (alloca_call_p): Use get_callee_fndecl.
4770 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
4771 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
4773 (vrp_visit_stmt): Remove duplicated code.
4775 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
4777 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
4778 ceildf2, btruncdf2, instead of vsx_* name.
4780 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
4781 iterators to only do V2DF and V4SF here. Move the DF code to
4782 rs6000.md where it is combined with SF mode. Replace <VSv> with
4783 just 'v' since only vector operations are handled with these insns
4784 after moving the DF support to rs6000.md.
4785 (vsx_sub<mode>3): Likewise.
4786 (vsx_mul<mode>3): Likewise.
4787 (vsx_div<mode>3): Likewise.
4788 (vsx_fre<mode>2): Likewise.
4789 (vsx_neg<mode>2): Likewise.
4790 (vsx_abs<mode>2): Likewise.
4791 (vsx_nabs<mode>2): Likewise.
4792 (vsx_smax<mode>3): Likewise.
4793 (vsx_smin<mode>3): Likewise.
4794 (vsx_sqrt<mode>2): Likewise.
4795 (vsx_rsqrte<mode>2): Likewise.
4796 (vsx_fms<mode>4): Likewise.
4797 (vsx_nfma<mode>4): Likewise.
4798 (vsx_copysign<mode>3): Likewise.
4799 (vsx_btrunc<mode>2): Likewise.
4800 (vsx_floor<mode>2): Likewise.
4801 (vsx_ceil<mode>2): Likewise.
4802 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
4803 (vsx_sminsf3): Likewise.
4804 (vsx_fmadf4): Likewise.
4805 (vsx_fmsdf4): Likewise.
4806 (vsx_nfmadf4): Likewise.
4807 (vsx_nfmsdf4): Likewise.
4808 (vsx_cmpdf_internal1): Likewise.
4810 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
4811 simpler to select whether a target has SPE or traditional floating
4812 point support in iterators.
4813 (TARGET_DF_SPE): Likewise.
4814 (TARGET_SF_FPR): Likewise.
4815 (TARGET_DF_FPR): Likewise.
4816 (TARGET_SF_INSN): Macros to say whether floating point support
4817 exists for a given operation for expanders.
4818 (TARGET_DF_INSN): Likewise.
4820 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
4821 combining of SF/DF mode operations, using both traditional and VSX
4829 (abs<mode>2): Combine SF/DF modes using traditional floating point
4830 instructions. Add support for using the upper DF registers with
4831 VSX support, and SF registers with power8-vector support. Update
4832 expanders for operations supported by both the SPE and traditional
4833 floating point units.
4834 (abs<mode>2_fpr): Likewise.
4835 (nabs<mode>2): Likewise.
4836 (nabs<mode>2_fpr): Likewise.
4837 (neg<mode>2): Likewise.
4838 (neg<mode>2_fpr): Likewise.
4839 (add<mode>3): Likewise.
4840 (add<mode>3_fpr): Likewise.
4841 (sub<mode>3): Likewise.
4842 (sub<mode>3_fpr): Likewise.
4843 (mul<mode>3): Likewise.
4844 (mul<mode>3_fpr): Likewise.
4845 (div<mode>3): Likewise.
4846 (div<mode>3_fpr): Likewise.
4847 (sqrt<mode>3): Likewise.
4848 (sqrt<mode>3_fpr): Likewise.
4849 (fre<Fs>): Likewise.
4850 (rsqrt<mode>2): Likewise.
4851 (cmp<mode>_fpr): Likewise.
4852 (smax<mode>3): Likewise.
4853 (smin<mode>3): Likewise.
4854 (smax<mode>3_vsx): Likewise.
4855 (smin<mode>3_vsx): Likewise.
4856 (negsf2): Delete SF operations that are merged with DF.
4863 (fmasf4_fpr): Likewise.
4864 (fmssf4_fpr): Likewise.
4865 (nfmasf4_fpr): Likewise.
4866 (nfmssf4_fpr): Likewise.
4867 (sqrtsf2): Likewise.
4868 (rsqrtsf_internal1): Likewise.
4869 (smaxsf3): Likewise.
4870 (sminsf3): Likewise.
4871 (cmpsf_internal1): Likewise.
4872 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
4873 (negdf2): Delete DF operations that are merged with SF.
4875 (nabsdf2): Likewise.
4881 (rsqrtdf_internal1): Likewise.
4882 (fmadf4_fpr): Likewise.
4883 (fmsdf4_fpr): Likewise.
4884 (nfmadf4_fpr): Likewise.
4885 (nfmsdf4_fpr): Likewise.
4886 (sqrtdf2): Likewise.
4887 (smaxdf3): Likewise.
4888 (smindf3): Likewise.
4889 (cmpdf_internal1): Likewise.
4890 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
4891 (btrunc<mode>2): Delete separate expander, and combine with the
4892 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
4893 (btrunc<mode>2_fpr): Likewise.
4894 (ceil<mode>2): Likewise.
4895 (ceil<mode>2_fpr): Likewise.
4896 (floor<mode>2): Likewise.
4897 (floor<mode>2_fpr): Likewise.
4898 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
4899 Add support for using the upper registers with VSX and
4900 power8-vector. Move insns to be closer to the define_expands. On
4901 VSX systems, prefer the traditional form of FMA over the VSX
4902 version, since the traditional form allows the target not to
4903 overlap with the inputs.
4904 (fms<mode>4_fpr): Likewise.
4905 (nfma<mode>4_fpr): Likewise.
4906 (nfms<mode>4_fpr): Likewise.
4908 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4909 Richard Earnshaw <richard.earnshaw@arm.com>
4911 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
4912 (struct mult_cost_table): Likewise.
4913 (struct mem_cost_table): Likewise.
4914 (struct fp_cost_table): Likewise.
4915 (struct vector_cost_table): Likewise.
4916 (cpu_cost_table): Likewise.
4917 * config/arm/arm.opt (mold-rts-costs): New option.
4918 (mnew-generic-costs): Likewise.
4919 * config/arm/arm.c (generic_extra_costs): New table.
4920 (cortexa15_extra_costs): Likewise.
4921 (arm_slowmul_tune): Use NULL as new costs.
4922 (arm_fastmul_tune): Likewise.
4923 (arm_strongarm_tune): Likewise.
4924 (arm_xscale_tune): Likewise.
4925 (arm_9e_tune): Likewise.
4926 (arm_v6t2_tune): Likewise.
4927 (arm_cortex_a5_tune): Likewise.
4928 (arm_cortex_a9_tune): Likewise.
4929 (arm_v6m_tune): Likewise.
4930 (arm_fa726te_tune): Likewise.
4931 (arm_cortex_a15_tune): Use cortex15_extra_costs.
4932 (arm_cortex_tune): Use generict_extra_costs.
4933 (shifter_op_p): New function.
4934 (arm_unspec_cost): Likewise.
4935 (LIBCALL_COST): Define.
4936 (arm_new_rtx_costs): New function.
4937 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
4938 table is available. Use old costs otherwise unless mnew-generic-costs
4940 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
4941 (cpu_cost_table): Declare.
4943 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
4946 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
4947 (*subs_mul_imm_<mode>)
4948 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
4949 (*sub_<shift>_<mode>)
4950 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
4951 Remove k constraint.
4953 2013-10-03 Ian Bolton <ian.bolton@arm.com>
4955 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
4957 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
4959 2013-10-02 Teresa Johnson <tejohnson@google.com>
4961 * predict.c (probably_never_executed): New function.
4962 (probably_never_executed_bb_p): Invoke probably_never_executed.
4963 (probably_never_executed_edge_p): Ditto.
4964 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
4965 Treat profile insanities conservatively.
4967 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
4969 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
4971 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
4973 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
4974 use it. Use smaller increase for scratch. Don't increase reject
4975 for early clobber scratch.
4976 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
4977 setting eliminated regs except setting fp from hfp.
4978 (lra_eliminate): Check lra_insn_recog_data on NULL.
4980 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4983 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
4984 setting -mvsx-timode by default until the underlying problem is fixed.
4985 (RS6000_CPU, power7 defaults): Likewise.
4987 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
4989 * config/x-linux (host-linux.o): Remove header dependencies.
4990 Use $(COMPILE) and $(POSTCOMPILE).
4991 * config/t-linux-android (linux-android.o): Ditto.
4993 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
4995 * Makefile.in (expmed.o-warn): Remove.
4997 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
4999 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5000 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5002 2013-10-02 Teresa Johnson <tejohnson@google.com>
5004 * dojump.c (do_jump_1): Divide probability between
5005 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5007 2013-10-02 Tom Tromey <tromey@redhat.com>
5009 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5011 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5013 * tree-flow.h: Remove some prototypes.
5014 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5015 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5016 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5017 mark_virtual_phi_result_for_renaming): Relocate here.
5018 * tree-into-ssa.h: Add prototypes.
5019 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5020 single_pred_before_succ_order.
5021 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5022 (nonfreeing_call_p) Move to gimple.c.
5023 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5025 * basic-block.h (single_pred_before_succ_order): Add prototype.
5026 * gimple.c (nonfreeing_call_p): Relocate here.
5027 * gimple.h: Add prototype.
5028 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5029 * tree-ssa-dom.h: New file. Relocate prototypes here.
5030 * tree-ssa.h: Include tree-ssa-dom.h.
5032 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5034 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5035 Use $(COMPILE) and $(POSTCOMPILE).
5037 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5039 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5041 * tree-flow.h: Remove some prototypes.
5042 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5043 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5044 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5045 propagate_tree_value*): Move from here to...
5046 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5047 propagate_tree_value*): Relocate here.
5048 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5049 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5050 * gimple-fold.c: Remove gimple-fold.h from include list.
5051 * tree-vrp.c: Remove gimple-fold.h from include list.
5052 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5053 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5054 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5055 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5056 * sese.c: Add tree-ssa-propagate.h to include list.
5058 2013-10-02 Richard Biener <rguenther@suse.de>
5060 * tree-loop-distribution.c: Include tree-vectorizer.h for
5062 (enum partition_kind): Remove PKIND_REDUCTION.
5063 (struct partition_s): Remove has_writes member, add reduction_p member.
5064 (partition_alloc): Adjust.
5065 (partition_builtin_p): Likewise.
5066 (partition_has_writes): Remove.
5067 (partition_reduction_p): New function.
5068 (partition_merge_into): Likewise.
5069 (generate_code_for_partition): Commonize builtin partition
5071 (rdg_cannot_recompute_vertex_p): Remove.
5072 (already_processed_vertex_p): Likewise.
5073 (rdg_flag_vertex): Do not set has_writes.
5074 (classify_partition): Adjust.
5075 (rdg_build_partitions): Do not set has_writes, treat all
5076 partitions as useful.
5077 (distribute_loop): Record number of library calls generated. Adjust.
5078 (tree_loop_distribution): Report number of loops and library
5079 calls generated as opt-info.
5081 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5083 * tree-flow.h: Include new .h files. Move prototypes.
5084 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5085 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5086 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5087 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5088 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5089 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5090 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5091 * tree.h (get_ref_base_and_extent): Move prototype out.
5092 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5094 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5095 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5096 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5097 * tree-scalar-evolution.c: Include tree.h.
5098 * sese.c: Include tree.h.
5099 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5100 * dwarf2out.c: Include tree-dfa.h.
5101 * tree-chrec.c: Include tree.h.
5102 * tree-data-ref.c: Include tree.h.
5104 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5106 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5109 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5111 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5112 Use $(COMPILE) and $(POSTCOMPILE).
5113 (sol2-cxx.o): Likewise.
5114 (sol2-stubs.o): Likewise.
5116 * config/x-solaris (host-solaris.o): Likewise.
5118 * config/sparc/t-sparc (sparc.o): Remove.
5119 (sparc-c.o): Remove header dependencies.
5120 Use $(COMPILE) and $(POSTCOMPILE).
5121 * config/sparc/x-sparc: Likewise.
5123 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5125 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5126 * config/arc/arc700.md: Likewise.
5127 * config/arc/arc-modes.def: Likewise.
5128 * config/arc/arc-simd.h: Likewise.
5129 * config/arc/t-arc-uClibc: Likewise.
5130 * config/arc/t-arc-newlib: Likewise.
5132 2013-10-02 Renlin Li <renlin.li@arm.com>
5134 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5136 (aarch64_expand_epilogue): Likewise.
5138 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5139 Yufeng Zhang <yufeng.zhang@arm.com>
5141 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5143 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5144 'base_in' represent a conversion and legal_cast_p_1 holds; set
5145 'base_in' with the returned value from get_unwidened.
5147 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5149 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5150 plus_constant is not used.
5152 2013-10-01 Wei Mi <wmi@google.com>
5154 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5155 * config/i386/i386.md: Add define_peephole2 to
5156 break partial reg stall for cvtss2sd/cvtsd2ss.
5158 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5160 * config/arc/arc.c (pass_arc_ifcvt::clone):
5161 Update for ctxt_ -> m_ctxt change.
5163 2013-10-01 Jeff Law <law@redhat.com>
5165 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5166 outgoing_edge and intermediate_edge fields. Instead store the path.
5167 (redirection_data::hash): Hash on the last edge's destination index.
5168 (redirection_data::equal): Check the entire thread path.
5169 (lookup_redirectio_data): Corresponding changes.
5170 (create_edge_and_update_destination_phis): Likewise.
5171 (thread_single_edge): Likewise.
5173 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5174 Diego Novillo <dnovillo@google.com>
5176 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5177 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5178 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5179 versions of these patterns.
5181 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
5182 (__builtin_arc_aligned): Likewise.
5184 * config/arc/arc.md: Expand adc_0 comment stating the intended
5185 purpose and why it isn't ready.
5186 Replace commented out call_value_via_label_mixed with a
5187 plain comment about bl_s.
5189 * config/arc/arc.c (stdio.h): Don't include directly.
5190 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
5191 Remove if (1) condition.
5192 (arc_encode_section_info): Fix comment.
5194 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5196 * config/arc/arc.c (arc_conditional_register_usage):
5197 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
5198 Also set reg_alloc_order for DMA config regs.
5200 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5201 Jeremy Bennett <jeremy.bennett@embecosm.com>
5203 * doc/install.texi (--with-cpu): Mention ARC.
5204 (arc-*-elf32): New paragraph.
5205 (arc-linux-uclibc): Likewise.
5206 * doc/md.texi (Machine Constraints): Add ARC part.
5207 * doc/invoke.texi: (menu): Add ARC Options.
5208 (Machine Dependent Options) <ARC Options>: Add synopsis.
5209 (node ARC Options): Add.
5210 * doc/extend.texi (long_call / short_call attribute): Add ARC.
5211 (ARC Built-in Functions): New section defining
5212 generic ARC built-in functions.
5213 (ARC SIMD Built-in Functions): New section defining SIMD specific
5215 (Declaring Attributes of Functions): Extended
5216 description of short_call and long_call attributes for ARC and
5217 added index entries.
5219 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
5220 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
5221 Joern Rennecke <joern.rennecke@embecosm.com>
5222 Muhammad Khurram Riaz <khurram.riaz@arc.com>
5223 Brendan Kehoe <brendan@zen.org>
5224 Michael Eager <eager@eagercon.com>
5225 Simon Cook <simon.cook@embecosm.com>
5226 Jeremy Bennett <jeremy.bennett@embecosm.com>
5228 * config/arc, common/config/arc: New directories.
5230 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5231 Brendan Kehoe <brendan@zen.org>
5232 Simon Cook <simon.cook@embecosm.com>
5234 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
5236 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5238 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
5239 * tree-ssa-coalesce.h: New. Move prototype to here.
5240 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
5241 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
5242 (gimple_can_coalesce_p): Move to...
5243 * gimple.c (gimple_can_coalesce_p): Here.
5245 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5247 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
5248 (dump_decl_set): Move to gimple.c.
5249 * gimple.h: Don't include tree-ssa-operands.h.
5250 (dump_decl_set): Add prototype.
5251 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5252 Move to gimple-ssa.h.
5253 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
5254 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
5255 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
5256 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
5257 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
5259 (dump_decl_set): Relocate here.
5260 * gimple-ssa.h: New file.
5261 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5262 Relocate from gimple.h.
5263 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
5264 * tree-ssa-operands.c (swap_ssa_operands): Rename from
5265 swap_tree_operands and remove non-ssa path.
5266 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
5267 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
5269 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
5270 vect_is_simple_reduction_1): Use swap_ssa_operands.
5271 * tree-flow.h: Move various prototypes to tree-phinodes.h.
5272 (enum need_phi_state): Move to tree-into-ssa.c.
5273 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5274 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
5275 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
5276 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
5277 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5278 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5279 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5280 num_imm_uses): Move to ssa-iterators.h.
5281 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
5282 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
5284 (op_iter_done, op_iter_next_def, op_iter_next_tree,
5285 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
5286 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
5287 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
5288 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
5289 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
5290 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
5291 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
5292 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
5293 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
5294 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
5295 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
5296 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
5297 (set_phi_nodes): Move to tree-phinodes.h.
5298 * tree-ssa-operands.h (enum ssa_op_iter_type,
5299 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
5300 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
5301 (dump_decl_set): Remove prototype.
5302 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
5303 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
5304 (set_phi_nodes): Relocate from tree-flow-inline.h.
5305 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
5307 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
5308 include list. Temporarily add gimple.h to include list.
5309 * ssa-iterators.h: New file.
5310 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5311 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
5312 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
5313 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
5314 Relocate from tree-ssa-operands.h.
5315 (delink_imm_use, link_imm_use_to_list, link_imm_use,
5316 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5317 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5318 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5319 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
5320 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
5321 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
5322 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
5323 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
5324 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
5325 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
5326 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
5327 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
5328 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
5329 Relocate from tree-flow-inline.h.
5330 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
5332 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
5335 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
5336 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
5337 Insert '\t' to output template.
5338 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
5339 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
5340 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
5341 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
5343 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
5345 * doc/install.texi (Host/target specific installation notes for GCC):
5346 Put @anchor before @heading.
5347 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
5348 Use @email for email addresses.
5350 2013-10-01 Jeff Law <law@redhat.com>
5352 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
5353 a vec. Only delete the path if we create one without successfully
5354 registering a jump thread.
5355 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
5357 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
5358 (paths): New vector of jump threading paths.
5359 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
5360 (THREAD_PATH): New accessor macro for the entire thread path.
5361 (lookup_redirection_data): Get intermediate and final outgoing edge
5362 from the thread path.
5363 (create_edge_and_update_destination_phis): Copy the threading path.
5364 (ssa_fix_duplicate_block_edges): Get edges and block types from the
5365 jump threading path.
5366 (ssa_redirect_edges): Get edges and block types from the jump threading
5367 path. Free the path vector.
5368 (thread_block): Get edges from the jump threading path. Look at the
5369 entire path to see if we thread to a loop exit. If we cancel a jump
5370 thread request, then free the path vector.
5371 (thread_single_edge): Get edges and block types from the jump threading
5372 path. Free the path vector.
5373 (thread_through_loop_header): Get edges and block types from the jump
5374 threading path. Free the path vector.
5375 (mark_threaded_blocks): Iterate over the vector of paths and store
5376 the path on the appropriate edge. Get edges and block types from the
5377 jump threading path.
5378 (mark_threaded_blocks): Get edges and block types from the jump
5379 threading path. Free the path vector.
5380 (thread_through_all_blocks): Use the vector of paths rather than
5381 a vector of 3-edge sets.
5382 (register_jump_thread): Accept pointer to a path vector rather
5383 than the path vector itself. Store the path vector for later use.
5386 2013-10-01 Jakub Jelinek <jakub@redhat.com>
5387 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5390 * config/s390/s390.c (s390_split_branches): Modify check for table
5392 (s390_chunkify_start): Rearrange table jump insn check in order to
5393 deal with compare and branch insns correctly.
5395 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
5399 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
5400 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
5401 define_insn_and_split.
5402 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
5403 (shiftsi3_compare): New pattern.
5405 * config/arm/unspecs.md (UNSPEC_RRX): New.
5407 2013-10-01 Alan Modra <amodra@gmail.com>
5409 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
5410 casing inout operands.
5412 2013-10-01 Richard Biener <rguenther@suse.de>
5414 PR tree-optimization/58553
5415 * tree-loop-distribution.c (struct partition_s): Add niter member.
5416 (classify_partition): Populate niter member for the partition
5417 and properly identify whether the relevant store happens before
5418 or after the loop exit.
5419 (generate_memset_builtin): Use niter member from the partition.
5420 (generate_memcpy_builtin): Likewise.
5422 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
5424 * vec.h (vec_prefix, vec): Prefix member names with "m_".
5425 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
5427 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
5429 * basic-block.h (edge_list): Prefix member names with "m_".
5430 * context.h (context): Likewise.
5431 * domwalk.h (dom_walker): Likewise.
5432 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
5433 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
5434 * hash-table.h (hash_table): Likewise.
5435 * machmode.h (bit_field_mode_iterator): Likewise.
5436 * pass_manager.h (pass_list): Likewise.
5437 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
5438 * tree-pass.h (pass_data): Likewise.
5439 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
5440 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
5441 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
5442 * asan.c (pass_data_asan): Update accordingly.
5443 * cfganal.c (control_dependences::find_control_dependence): Likewise.
5444 (control_dependences::control_dependences): Likewise.
5445 (control_dependences::~control_dependences): Likewise.
5446 (control_dependences::~control_dependences): Likewise.
5447 (control_dependences::get_edges_dependent_on): Likewise.
5448 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
5449 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
5450 * context.c (gcc::context::context): Likewise.
5451 * cprop.c (pass_rtl_cprop::clone): Likewise.
5452 * domwalk.c (dom_walker::walk): Likewise.
5453 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
5454 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
5455 * mode-switching.c (pass_mode_switching::clone): Likewise.
5456 * passes.c (opt_pass::opt_pass): Likewise.
5457 (pass_manager::pass_manager): Likewise.
5458 * predict.c (pass_strip_predict_hints::clone): Likewise.
5459 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
5460 (pass_split_all_insns::clone): Likewise.
5461 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
5463 (bit_field_mode_iterator::next_mode): Likewise.
5464 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
5465 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
5466 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
5467 * tree-complex.c (pass_lower_complex::clone): Likewise.
5468 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
5469 * tree-object-size.c (pass_object_sizes::clone): Likewise.
5470 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
5471 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
5472 (pass_fold_builtins::clone): Likewise.
5473 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
5474 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
5475 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
5476 (pass_cd_dce::clone): Likewise.
5477 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
5478 (pass_phi_only_cprop::clone): Likewise.
5479 * tree-ssa-dse.c (pass_dse::clone): Likewise.
5480 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
5481 * tree-ssa-loop.c (pass_lim::clone): Likewise.
5482 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
5483 * tree-ssa-pre.c (pass_fre::clone): Likewise.
5484 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
5485 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
5486 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
5487 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
5488 * tree-vrp.c (pass_vrp::clone): Likewise.
5489 * tsan.c (pass_tsan::clone): Likewise.
5491 2013-09-30 Jakub Jelinek <jakub@redhat.com>
5494 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
5495 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
5498 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
5499 optimization, punt if sign_bit_p looked through any zero extension.
5501 2013-09-30 Teresa Johnson <tejohnson@google.com>
5503 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
5504 Update redirected out edge count in joiner case.
5505 (ssa_redirect_edges): Common the joiner and non-joiner cases
5506 so that joiner case gets profile updates.
5508 2013-09-30 Richard Biener <rguenther@suse.de>
5510 PR tree-optimization/58554
5511 * tree-loop-distribution.c (classify_partition): Require
5512 unconditionally executed stores for memcpy and memset recognition.
5513 (tree_loop_distribution): Calculate dominance info.
5515 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5517 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
5518 (NO_PROFILE_COUNTERS): Likewise.
5519 (PROFILE_HOOK): Likewise.
5520 (FUNCTION_PROFILER): Likewise.
5521 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
5523 2013-09-30 Iain Sandoe <iain@codesourcery.com>
5525 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
5526 calls and defines in TARGET_MACHO conditional.
5527 (load_macho_picbase_di): Likewise.
5528 (reload_macho_picbase): Likewise.
5529 (reload_macho_picbase_si): Likewise.
5530 (reload_macho_picbase_di): Likewise.
5531 (nonlocal_goto_receiver): Likewise.
5533 2013-09-30 Nick Clifton <nickc@redhat.com>
5535 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
5536 that use the MSP430X ISA.
5537 (msp430_option_override): Scan -mmcu command line option for any
5538 MCU name that supports the MSP430X ISA.
5539 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
5540 -mmcu options which enable the MSP430X ISA.
5542 2013-09-30 Richard Biener <rguenther@suse.de>
5545 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
5546 before looking for setjmp-like calls.
5548 2013-09-29 Iain Sandoe <iain@codesourcery.com>
5551 * config/darwin-protos.h (machopic_get_function_picbase): New.
5552 * config/darwin.c (machopic_get_function_picbase): New.
5553 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
5554 label for a new func. (load_macho_picbase_di): Likewise.
5555 (reload_macho_picbase): New expand.
5556 (reload_macho_picbase_si): New insn.
5557 (reload_macho_picbase_di): New insn.
5558 (nonlocal_goto_receiver): New define and split.
5559 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
5560 (unspecv enum): Add UNSPECV_NLGR.
5562 2013-09-29 Iain Sandoe <iain@codesourcery.com>
5564 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
5565 that altivec registers are correctly sized on Darwin.
5567 2013-09-29 Iain Sandoe <iain@codesourcery.com>
5569 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
5570 darwin-driver.o): Use COMPILE and POSTCOMPILE.
5571 * config/x-darwin (host-darwin.o): Likewise.
5572 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
5573 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
5574 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
5576 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
5578 * doc/invoke.texi: Fix usage of @tie{} command.
5580 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
5582 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
5584 2013-09-28 Jan Hubicka <jh@suse.cz>
5586 * config/i386/x86-tune.def: Add documentation for each of the options;
5589 2013-09-28 Jan Hubicka <jh@suse.cz>
5591 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
5593 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
5594 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
5595 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
5596 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
5597 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
5599 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
5601 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
5602 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
5603 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
5604 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
5605 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
5606 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
5607 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
5608 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
5609 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
5610 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
5611 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
5612 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
5613 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
5614 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
5615 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
5616 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
5617 gimple.h, godump.c, graphite-clast-to-gimple.c,
5618 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
5619 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
5620 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
5621 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5622 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
5623 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
5624 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
5625 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
5626 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
5627 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
5628 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
5629 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
5630 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
5631 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
5632 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
5633 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
5634 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
5635 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
5636 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
5637 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
5638 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
5639 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
5640 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
5641 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
5642 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5643 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5644 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
5645 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
5646 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
5647 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5648 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5649 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
5650 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
5651 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5652 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5653 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
5654 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
5655 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
5656 whitespace before "(".
5658 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
5660 * expr.h (extract_bit_field): Remove packedp parameter.
5661 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
5662 from forward declaration.
5663 (store_split_bit_field): Remove packedp arg from calls to
5664 extract_fixed_bit_field.
5665 (extract_bit_field_1): Remove packedp parameter and packedp
5666 argument from recursive calls and calls to extract_fixed_bit_field.
5667 (extract_bit_field): Remove packedp parameter and corresponding
5668 arg to extract_bit_field_1.
5669 (extract_fixed_bit_field): Remove packedp parameter. Remove code
5671 (extract_split_bit_field): Remove packedp arg from call to
5672 extract_fixed_bit_field.
5673 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
5674 (copy_blkmode_from_reg): Likewise.
5675 (copy_blkmode_to_reg): Likewise.
5676 (read_complex_part): Likewise.
5677 (store_field): Likewise.
5678 (expand_expr_real_1): Likewise.
5679 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
5680 to extract_bit_field.
5681 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
5682 call to extract_bit_field.
5683 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
5684 call to extract_bit_field.
5685 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
5686 and special packedp behavior from -fstrict-volatile-bitfields
5689 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5691 * lra-eliminations.c (init_elim_table): Guard value_p.
5693 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5695 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
5696 DFmode, DImode, and SFmode in the upper VSX registers based on the
5697 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
5698 if -mpower8-vector. Combine -mvsx-timode handling with the rest
5699 of the VSX register handling.
5701 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
5703 (zero_extendsidi2_lfiwzx): Add support for loading into the
5704 Altivec registers with -mpower8-vector. Use wu/wv constraints to
5705 only do VSX memory options on Altivec registers.
5706 (extendsidi2_lfiwax): Likewise.
5707 (extendsfdf2_fpr): Likewise.
5708 (mov<mode>_hardfloat, SF/SD modes): Likewise.
5709 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
5710 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
5711 (movdi_internal64): Likewise.
5713 2013-09-27 Xinliang David Li <davidxl@google.com>
5715 * opts.c (finish_options): Adjust parameters
5716 according to vect cost model.
5717 (common_handle_option): Set dynamic vect cost
5719 targhooks.c (default_add_stmt_cost): Compute stmt cost
5721 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
5722 Use helper function.
5723 * tree-vectorizer.h (unlimited_cost_model): New function.
5724 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
5725 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
5727 (vect_enhance_data_refs_alignment): Ditto.
5728 * flag-types.h: New enum.
5729 * common/config/i386/i386-common.c (ix86_option_init_struct):
5730 No need to initialize vect_cost_model flag.
5731 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
5734 2013-09-27 Diego Novillo <dnovillo@google.com>
5736 * gimple.h (enum ssa_mode): Remove.
5738 2013-09-27 Paulo Matos <pmatos@broadcom.com>
5740 * cfgloop.h (number_of_loops): Fix typo in check for null.
5742 2013-09-27 Jakub Jelinek <jakub@redhat.com>
5745 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
5746 are children of outermost saved_cfun's loop, and set it up to
5747 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
5748 if loop != loop0 and SESE region contains bbs that belong to loop0.
5750 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
5752 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
5753 (binary_scale_code_p, get_base_term, get_index_term): New functions.
5754 (set_address_segment, set_address_base, set_address_index)
5755 (set_address_disp): Accept the argument unconditionally.
5756 (baseness): Remove must_be_base_p and must_be_index_p checks.
5757 (decompose_normal_address): Classify as much as possible in the
5760 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
5762 * cse.c (count_reg_usage): Handle INT_LIST.
5763 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
5764 * reginfo.c (reg_scan_mark_refs): Likewise.
5765 * reload1.c (eliminate_regs_1): Likewise.
5767 2013-09-27 Iain Sandoe <iain@codesourcery.com>
5770 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
5773 2013-09-27 Richard Biener <rguenther@suse.de>
5775 PR tree-optimization/58459
5776 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
5777 restriction not propagating into loops.
5779 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
5781 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
5782 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
5783 * doc/tree-ssa.texi (Walking use-def chains): Delete.
5785 2013-09-26 Richard Biener <rguenther@suse.de>
5787 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
5789 2013-09-26 Richard Biener <rguenther@suse.de>
5791 * alias.h (component_uses_parent_alias_set): Rename to ...
5792 (component_uses_parent_alias_set_from): ... this.
5793 * alias.c (component_uses_parent_alias_set): Rename to ...
5794 (component_uses_parent_alias_set_from): ... this and return
5796 (reference_alias_ptr_type_1): Use the result from
5797 component_uses_parent_alias_set_from instead of stripping
5798 components one at a time.
5799 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
5801 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
5803 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
5804 Move prototypes to...
5805 * tree-ssa-ter.h: New File. Move prototypes here.
5806 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
5807 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
5808 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
5809 from is_replaceable_p.
5810 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
5811 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
5812 (process_replaceable): Use ter_is_replaceable_p.
5813 (find_replaceable_in_bb): Use ter_is_replaceable_p.
5814 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
5815 newly refactored ssa_is_replaceable_p.
5816 * cfgexpand.c: Include tree-outof-ssa.h.
5817 * ssaexpand.h: Delete.
5819 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
5821 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
5822 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
5824 (create_gimple_tmp): Delete.
5825 (get_expr_type, build_assign, build_type_cast): Move to...
5826 * gimple-builder.c: New File.
5827 (get_expr_type): Relocate from gimple.c.
5828 (build_assign, build_type_cast): Change to only create ssanames.
5829 * gimple.h: Move prototypes to...
5830 * gimple-builder.h: New File. Here.
5831 * tree-ssa.h: And here.
5832 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
5833 count_uses_and_derefs): Relocate from gimple.c.
5834 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
5835 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
5836 * tree-ssa-math-opts (execute_cse_reciprocals): Use
5837 gimple_replace_ssa_lhs.
5838 * asan.c: Include gimple-builder.h.
5839 * Makefile.in: Add gimple-builder.o.
5841 2013-09-26 Richard Biener <rguenther@suse.de>
5843 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
5844 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
5845 (loe_visit_block): Use gcc_checking_assert.
5846 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
5847 gimple_assign_ssa_name_copy_p.
5848 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
5851 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
5853 * config/rs6000/t-rs6000 (rs6000.o): Remove.
5854 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
5856 2013-09-26 Richard Biener <rguenther@suse.de>
5858 PR tree-optimization/58539
5859 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
5860 the fact that debug statements are not taking part in loop-closed
5863 2013-09-26 Nick Clifton <nickc@redhat.com>
5865 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
5866 time warning message.
5867 (msp430_print_operand_raw): Delete unused letter parameter.
5868 (TARGET_PRINT_OPERAND_ADDRESS): Define.
5869 (msp430_print_operand_address): New function.
5870 (msp430_print_operand): Move address printing code from here to
5872 * config/msp430/msp430.md (movsipsi2): Add comment in generated
5874 (zero_extendpsisi2): Likewise.
5875 (extendpsisi2): New pattern.
5876 (andneghi3): New pattern.
5878 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
5880 * config/aarch64/aarch64.opt (mlra): New option.
5881 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
5882 (TARGET_LRA_P): Define.
5884 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
5886 * expr.c (expand_assignment): Remove obsolete comment.
5888 2013-09-25 Jeff Law <law@redhat.com>
5890 * tree-flow.h (thread_through_all_blocks): Prototype moved into
5891 tree-ssa-threadupdate.h.
5892 (register_jump_thread): Similarly.
5893 * tree-ssa-threadupdate.h: New header file.
5894 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
5895 * tree-vrp.c: Likewise.
5896 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
5897 (thread_around_empty_blocks): Change type of path vector argument to
5898 an edge,type pair from just an edge. Initialize both elements when
5899 appending to a jump threading path. Tweak references to elements
5901 (thread_across_edge): Similarly. Release memory for the elements
5903 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
5904 (dump_jump_thread_path): New function broken out from
5905 register_jump_thread.
5906 (register_jump_thread): Use dump_jump_thread_path. Change type of
5907 path vector entries. Search the path for NULL edges and dump
5908 the path if one is found. Tweak the conversion of path to 3-edge
5909 form to use the block copy type information embedded in the path.
5911 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
5913 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
5915 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
5916 Vladimir Makarov <vmakarov@redhat.com>
5918 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
5919 from the least significant bit.
5920 (strip_address_mutations): Add bitfield operations handling.
5921 (must_be_index_p): Add shifting and rotate operations handling.
5922 (set_address_base): Use must_be_base_p predicate.
5923 (set_address_index): Use must_be_index_p predicate.
5925 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
5926 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5927 Sergey Lega <sergey.s.lega@intel.com>
5928 Anna Tikhonova <anna.tikhonova@intel.com>
5929 Ilya Tocar <ilya.tocar@intel.com>
5930 Andrey Turetskiy <andrey.turetskiy@intel.com>
5931 Ilya Verbin <ilya.verbin@intel.com>
5932 Kirill Yukhin <kirill.yukhin@intel.com>
5933 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5935 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
5937 (ix86_expand_vector_move_misalign): Support new unaligned load and
5938 stores and use new names.
5939 (CODE_FOR_sse2_storedqu): Rename to ...
5940 (CODE_FOR_sse2_storedquv16qi): ... this.
5941 (CODE_FOR_sse2_loaddqu): Rename to ...
5942 (CODE_FOR_sse2_loaddquv16qi): ... this.
5943 (CODE_FOR_avx_loaddqu256): Rename to ...
5944 (CODE_FOR_avx_loaddquv32qi): ... this.
5945 (CODE_FOR_avx_storedqu256): Rename to ...
5946 (CODE_FOR_avx_storedquv32qi): ... this.
5947 * config/i386/i386.md (fpint_logic): New.
5948 * config/i386/sse.md (VMOVE): Extend for AVX512.
5952 (VI_UNALIGNED_LOADSTORE): Ditto.
5953 (sse2_avx_avx512f): Ditto.
5954 (sse2_avx2): Extend for AVX512.
5955 (sse4_1_avx2): Ditto.
5956 (avx2_avx512f): New.
5957 (sse): Extend for AVX512.
5960 (avxsizesuffix): Ditto.
5961 (sseintvecmode): Ditto.
5963 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
5964 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
5965 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
5966 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
5967 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
5968 (<sse2_avx_avx512f>_storedqu<mode): ... this.
5969 (<sse>_movnt<mode>): Replace constraint "x" with "v".
5970 (STORENT_MODE): Extend for AVX512.
5971 (*absneg<mode>2): Replace constraint "x" with "v".
5972 (*mul<mode>3): Ditto.
5973 (*ieee_smin<mode>3): Ditto.
5974 (*ieee_smax<mode>3): Ditto.
5975 (avx_cmp<mode>3): Replace VF with VF_128_256.
5976 (*<sse>_maskcmp<mode>3_comm): Ditto.
5977 (<sse>_maskcmp<mode>3): Ditto.
5978 (<sse>_andnot<mode>3): Extend for AVX512.
5979 (<code><mode>3, anylogic): Replace VF with VF_128_256.
5980 (<code><mode>3, fpint_logic): New.
5981 (*<code><mode>3): Extend for AVX512.
5982 (avx512flogicsuff): New.
5983 (avx512f_<logic><mode>): Ditto.
5984 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
5986 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
5987 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
5988 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
5989 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
5990 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
5991 (xop_vpermil2<mode>3): Ditto.
5992 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
5993 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
5994 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
5995 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
5997 2013-09-25 Tom Tromey <tromey@redhat.com>
5999 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6000 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6001 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6002 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6003 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6004 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6005 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6006 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6007 (GRAPHITE_HTAB_H): Remove.
6009 2013-09-25 Tom Tromey <tromey@redhat.com>
6011 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6013 2013-09-25 Tom Tromey <tromey@redhat.com>
6015 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6017 2013-09-25 Tom Tromey <tromey@redhat.com>
6019 * config/i386/t-i386 (i386.o): Remove.
6020 (i386-c.o): Use COMPILE and POSTCOMPILE.
6022 2013-09-25 Tom Tromey <tromey@redhat.com>
6024 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6026 2013-09-25 Tom Tromey <tromey@redhat.com>
6028 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6029 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6030 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6031 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6032 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6033 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6034 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6035 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6036 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6037 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6038 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6039 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6040 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6041 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6042 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6043 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6044 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6045 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6046 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6047 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6048 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6049 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6050 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6051 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6052 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6053 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6054 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6055 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6056 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6057 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6058 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6059 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6060 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6061 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6062 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6063 (graphite-clast-to-gimple.o, graphite-dependences.o)
6064 (graphite-interchange.o, graphite-poly.o)
6065 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6066 (graphite-optimize-isl.o, tree-vect-loop.o)
6067 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6068 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6069 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6070 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6071 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6072 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6073 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6074 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6075 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6076 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6077 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6078 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6079 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6080 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6081 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6082 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6083 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6084 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6085 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6086 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6087 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6088 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6089 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6090 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6091 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6092 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6093 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6094 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6095 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6096 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6097 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6098 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6099 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6100 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6101 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6102 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6103 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6104 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6105 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6106 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6107 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6108 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6109 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6110 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6111 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6112 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6113 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6114 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6115 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6116 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6117 (gcov-dump.o): Remove.
6118 (default-c.o): Use COMPILE and POSTCOMPILE.
6119 (CFLAGS-gcc.o): New variable.
6120 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6122 2013-09-25 Tom Tromey <tromey@redhat.com>
6124 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6125 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6126 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6127 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6128 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6129 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6130 (c-family/array-notation-common.o, c-family/stub-objc.o)
6131 (c-family/c-ubsan.o): Remove.
6133 2013-09-25 Tom Tromey <tromey@redhat.com>
6135 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6137 2013-09-25 Tom Tromey <tromey@redhat.com>
6139 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6140 to add -DENABLE_SHARED_LIBGCC.
6141 (gcc.o): Don't use subshell.
6143 2013-09-25 Tom Tromey <tromey@redhat.com>
6145 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6146 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6147 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6148 * configure, config.in: Rebuild.
6150 2013-09-25 Tom Tromey <tromey@redhat.com>
6152 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6153 (COMPILE, POSTCOMPILE): New variables.
6154 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6155 (DEPFILES): New variable.
6156 Include ".Po" files.
6157 * configure.ac: Add checks for dependency checking.
6158 * configure, aclocal.m4: Regenerate.
6160 2013-09-25 Tom Tromey <tromey@redhat.com>
6162 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6163 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6165 2013-09-25 Tom Tromey <tromey@redhat.com>
6167 * Makefile.in (generated_files): Add options.h,
6168 target-hooks-def.h, insn-opinit.h,
6169 common/common-target-hooks-def.h, pass-instances.def,
6170 c-family/c-target-hooks-def.h.
6172 2013-09-25 Jeff Law <law@redhat.com>
6174 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6175 than foo[foo.length () - 1] to access last member in a vec.
6176 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6178 2013-09-25 Richard Biener <rguenther@suse.de>
6181 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
6183 2013-09-25 Jan Hubicka <jh@suse.cz>
6185 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
6188 2013-09-25 Marek Polacek <polacek@redhat.com>
6191 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
6192 when determining the type name.
6194 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
6196 * config/sh/sh.md: Fix formatting.
6198 2013-09-24 Xinliang David Li <davidxl@google.com>
6200 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
6201 max peel iterations parameter.
6202 * param.def: New parameter.
6203 * doc/invoke.texi: Document New parameter.
6205 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
6207 * gimple-pretty-print.c: Various whitespace tweaks.
6208 * tree-core.h: Likewise.
6209 * tree-pretty-print.c: Likewise.
6210 * tree-ssa-alias.c: Likewise.
6211 * tree-ssa-copy.c: Likewise.
6212 * tree-ssanames.c: Likewise.
6213 * tree-ssanames.h: Likewise.
6214 * tree-vrp.c: Likewise.
6216 2013-09-24 Alan Modra <amodra@gmail.com>
6220 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
6221 for output operands that disallow regs. Don't use EXPAND_WRITE on
6224 2013-09-24 Richard Biener <rguenther@suse.de>
6227 * tree.c (reference_alias_ptr_type): Move ...
6228 * alias.c (reference_alias_ptr_type): ... here and implement
6229 in terms of the new reference_alias_ptr_type_1.
6230 (ref_all_alias_ptr_type_p): New helper.
6231 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
6232 use ref_all_alias_ptr_type_p.
6233 (get_deref_alias_set): Add flag_strict_aliasing check here.
6234 (reference_alias_ptr_type_1): New function, split out from ...
6235 (get_alias_set): ... here.
6236 (alias_ptr_types_compatible_p): New function.
6237 * alias.h (reference_alias_ptr_type): Declare.
6238 (alias_ptr_types_compatible_p): Likewise.
6239 * tree.h (reference_alias_ptr_type): Remove.
6240 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
6241 to compare MEM_REF alias types.
6243 2013-09-24 Richard Biener <rguenther@suse.de>
6245 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
6247 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6249 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
6250 reload helper function arrays into a single array reg_addr.
6251 (reload_fpr_gpr): Likewise.
6252 (reload_gpr_vsx): Likewise.
6253 (reload_vsx_gpr): Likewise.
6254 (struct rs6000_reg_addr): Likewise.
6255 (reg_addr): Likewise.
6256 (rs6000_debug_reg_global): Change rs6000_vector_reload,
6257 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
6258 (rs6000_init_hard_regno_mode_ok): Likewise.
6259 (rs6000_secondary_reload_direct_move): Likewise.
6260 (rs6000_secondary_reload): Likewise.
6262 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
6263 constraints: wu, ww, and wy. Repurpose wv constraint added during
6264 power8 changes. Put wg constraint in alphabetical order.
6266 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
6267 for future work to add ISA 2.07 VSX single precision support.
6268 (-mvsx-scalar-double): Change default from -1 to 1, update
6269 documentation comment.
6270 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
6271 (-mupper-regs-df): New debug switch to control whether DF values
6272 can go in the traditional Altivec registers.
6273 (-mupper-regs-sf): New debug switch to control whether SF values
6274 can go in the traditional Altivec registers.
6276 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
6278 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
6279 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
6280 Add new constraints, wu/ww/wy. Repurpose wv constraint.
6281 (rs6000_debug_legitimate_address_p): Print if we are running
6282 before, during, or after reload.
6283 (rs6000_secondary_reload): Add a comment.
6284 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
6286 * config/rs6000/constraints.md (wa constraint): Sort w<x>
6287 constraints. Update documentation string.
6288 (wd constraint): Likewise.
6289 (wf constraint): Likewise.
6290 (wg constraint): Likewise.
6291 (wn constraint): Likewise.
6292 (ws constraint): Likewise.
6293 (wt constraint): Likewise.
6294 (wx constraint): Likewise.
6295 (wz constraint): Likewise.
6296 (wu constraint): New constraint for ISA 2.07 SFmode scalar
6298 (ww constraint): Likewise.
6299 (wy constraint): Likewise.
6300 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
6301 in the previous submissions.
6302 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
6304 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
6306 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
6307 (REG_BR_PROB): Say that the probability is stored in an int_list.
6308 * reg-notes.def: Update commentary to mention INT_LIST.
6309 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
6310 (INT_LIST): New rtx.
6311 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
6312 * rtlanal.c (int_reg_note_p): New function.
6313 (alloc_reg_note): Assert that the note does not have an int argument.
6314 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
6315 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
6316 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
6317 rather than an INSN_LIST. Handle INT_LIST.
6318 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
6319 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
6320 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
6321 Manipulate them as ints rather than rtxes.
6322 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
6323 * regmove.c (copy_src_to_dest): Likewise.
6324 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
6326 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
6327 into the cases that need it.
6328 * config/arm/arm.c (arm_unwind_emit): Likewise.
6330 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
6331 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
6332 * loop-doloop.c (add_test, doloop_modify): Likewise.
6333 * loop-unswitch.c (compare_and_jump_seq): Likewise.
6334 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
6335 * predict.c (combine_predictions_for_insn): Likewise.
6336 * print-rtl.c (print_rtx): Handle INT_LIST.
6337 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
6338 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
6339 * config/arm/arm.c (emit_unlikely_jump): Likewise.
6340 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
6341 (ix86_split_fp_branch, predict_jump): Likewise.
6342 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
6343 * config/sh/sh.c (expand_cbranchsi4): Likewise.
6344 * config/spu/spu.c (ea_load_store_inline): Likewise.
6346 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
6347 value of a REG_BR_PROB note.
6348 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
6349 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
6350 * emit-rtl.c (try_split): Likewise.
6351 * predict.c (br_prob_note_reliable_p): Likewise.
6352 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
6353 * reorg.c (mostly_true_jump): Likewise.
6354 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
6355 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
6356 * config/i386/i386.c (ix86_print_operand): Likewise.
6357 * config/ia64/ia64.c (ia64_print_operand): Likewise.
6358 * config/mmix/mmix.c (mmix_print_operand): Likewise.
6359 * config/rs6000/rs6000.c (output_cbranch): Likewise.
6360 * config/s390/s390.c (s390_expand_tbegin): Likewise.
6361 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
6362 * config/sparc/sparc.c (output_cbranch): Likewise.
6363 * config/spu/spu.c (get_branch_target): Likewise.
6364 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
6365 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
6367 2013-09-23 Jan Hubicka <jh@suse.cz>
6369 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
6371 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
6372 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
6374 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
6375 (gimple_fold_call): Dump inconsistent devirtualizations; add
6376 sanity check for type based devirtualizations.
6377 * ipa-prop.c: Include ipa-utils.h
6378 (ipa_intraprocedural_devirtualization): Add sanity check.
6379 (try_make_edge_direct_virtual_call): Likewise.
6381 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
6383 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
6384 assignment statements.
6386 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
6388 * gimple-pretty-print.c (dump_ssaname_info): New function.
6389 (dump_gimple_phi): Call it.
6390 (pp_gimple_stmt_1): Likewise.
6391 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
6392 (range_info_def): Declare.
6393 * tree-pretty-print.c (pp_double_int): New function.
6394 (dump_generic_node): Call it.
6395 * tree-pretty-print.h (pp_double_int): Declare.
6396 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
6397 * tree-ssanames.h (range_info_def): New structure.
6398 (value_range_type): Move definition here.
6399 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
6401 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
6403 (set_range_info): New function.
6404 (get_range_info): Likewise.
6405 (duplicate_ssa_name_range_info): Likewise.
6406 (duplicate_ssa_name_fn): Check pointer type and call
6407 duplicate_ssa_name_range_info.
6408 * tree-ssa-copy.c (fini_copy_prop): Likewise.
6409 * tree-vrp.c (value_range_type): Remove definition, now in
6411 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
6412 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
6413 (SSA_NAME_RANGE_INFO): New macro.
6415 2013-09-23 Richard Biener <rguenther@suse.de>
6417 PR tree-optimization/58464
6418 * tree-ssa-pre.c (phi_trans_lookup): Remove.
6419 (phi_trans_add): Change to add conditionally on being not present.
6420 (phi_translate_1): Remove recursion detection here.
6421 (phi_translate): Pre-seed the cache with NULL to catch
6422 recursion here in a more generic way.
6423 (bitmap_find_leader): Adjust comment.
6424 (get_representative_for): Dump value-numbers.
6425 (create_expression_by_pieces): Likewise.
6426 (insert_into_preds_of_block): Likewise.
6428 2013-09-23 Christian Bruel <christian.bruel@st.com>
6431 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
6432 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
6434 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
6437 2013-09-20 Renlin Li <renlin.li@arm.com>
6439 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6441 (aarch64_expand_epilogue): Likewise.
6442 (aarch64_legitimize_reload_address): Likewise.
6444 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
6446 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
6447 NULL_TREE before pushing them onto the vector. Likewise for labels.
6449 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
6451 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
6452 comparison operators when -fno-trapping-math is in effect.
6453 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
6454 comparison operators in TFmode and assert that unsupported operators
6456 (ia64_print_operand): Likewise.
6458 2013-09-21 Jan Hubicka <jh@suse.cz>
6460 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
6461 (sse_typeless_stores): Enable for core
6462 (sse_load0_by_pxor): Likewise.
6463 (four_jump_limit): Disable for core.
6464 (pad_returns): Likewise.
6465 (avoid_vector_decode): Likewise.
6466 (fuse_cmp_and_branch): Enable for cores.
6467 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
6469 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
6472 * config/pa/pa.c (pa_option_override): Disable auto increment and
6473 decrement instructions until reload is completed.
6475 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
6476 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
6477 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
6479 2013-09-20 DJ Delorie <dj@redhat.com>
6480 Nick Clifton <nickc@redhat.com>
6482 * config/rl78/rl78.c: Various whitespace and comment tweaks.
6483 (need_to_save): Save bank 0 on interrupts.
6484 (characterize_address): Strip far address wrappers.
6485 (rl78_as_legitimate_address): Likewise.
6486 (transcode_memory_rtx): Likewise.
6487 (rl78_peep_movhi_p): Disable this peephole after devirt.
6488 (rl78_propogate_register_origins): Forget all origins when a
6490 * config/rl78/rl78-virt.md: Various whitespace tweaks.
6491 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
6493 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
6494 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
6495 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
6497 * config/rl78/constraints.md: Various whitespace and paren tweaks.
6499 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
6501 * config/pa/pa.md: In "scc" insn patterns, change output template to
6502 handle const0_rtx in reg_or_0_operand operands.
6504 2013-09-20 Martin Husemann <martin@NetBSD.org>
6507 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
6508 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
6510 2013-09-20 Richard Biener <rguenther@suse.de>
6513 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
6514 remembering SSA name version and block index.
6515 (new_scev_info_str): Adjust.
6516 (hash_scev_info): Likewise. Also hash the block index.
6517 (eq_scev_info): Adjust.
6518 (find_var_scev_info): Likewise.
6519 (struct instantiate_cache_entry): Remove.
6520 (struct instantiate_cache_type): Use a htab to map name, block
6522 (instantiate_cache_type::~instantiate_cache_type): Adjust.
6523 (get_instantiated_value_entry): Likewise.
6524 (hash_idx_scev_info, eq_idx_scev_info): New functions.
6525 (instantiate_scev_name): Adjust.
6527 2013-09-20 Jeff Law <law@redhat.com>
6529 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
6531 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
6533 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
6534 Call aarch64_simd_expand_args to update op[argc].
6536 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
6538 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
6541 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
6543 * gengtype.c (file_rules): Added rule for *.cc files.
6544 (get_output_file_with_visibility): Give fatal message when no
6547 2013-09-20 Renlin Li <renlin.li@arm.com>
6549 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
6550 (aarch64_expand_epilogue): Likewise.
6551 (aarch64_legitimize_reload_address): Likewise.
6553 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
6556 * expr.c (expand_assignment): Remove misalignp code path.
6558 2013-09-20 Marek Polacek <polacek@redhat.com>
6561 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
6562 TYPE_PRECISION. Add asserts.
6564 2013-09-20 Richard Biener <rguenther@suse.de>
6566 PR tree-optimization/58453
6567 * tree-loop-distribution.c (distribute_loop): Apply the cost
6568 model for -ftree-loop-distribute-patterns, too.
6570 2013-09-20 Richard Biener <rguenther@suse.de>
6573 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
6574 make type comparison less strict.
6576 2013-09-20 Alan Modra <amodra@gmail.com>
6578 * configure: Regenerate.
6579 * aclocal.m4: Regenerate.
6581 2013-09-20 Marek Polacek <polacek@redhat.com>
6584 * doc/extend.texi: Document that attribute used is meant to be used
6585 on variables with static storage duration.
6587 2013-09-19 Jakub Jelinek <jakub@redhat.com>
6589 PR tree-optimization/58472
6590 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
6591 simd_lane_access set inv_p = false.
6592 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
6593 the simduid magic VAR_DECL.
6595 2013-09-19 Jan Hubicka <jh@suse.cz>
6597 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
6599 2013-09-17 Jeff Law <law@redhat.com>
6601 * tree-ssa-dom.c (record_temporary_equivalences): New function
6602 split out of dom_opt_dom_walker::after_dom_children.
6603 (dom_opt_dom_walker::thread_across_edge): Move common code
6604 in here from dom_opt_dom_walker::after_dom_children.
6605 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
6607 2013-09-19 Jan Hubicka <jh@suse.cz>
6609 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
6610 (TARGET_GENERIC): Use PROCESOR_GENERIC
6611 (enum processor_type): Unify generic32 and 64.
6612 * i386.md (cpu): Likewise.
6613 * x86-tune.def (use_leave): Enable for generic32.
6614 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
6615 * athlon.md: Change generic64 to generic in all occurences.
6616 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
6617 (ix86_target_macros_internal): Likewise.
6618 * driver-i386.c (host_detect_local_cpu): Likewise.
6619 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
6621 (generic_memcpy, generic_memset, generic_cost): This one.
6622 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
6623 (m_GENERIC32, m_GENERIC64): Remove.
6624 (m_GENERIC): Turn into one flag.
6625 (processor_target): Unify generic tunnings.
6626 (ix86_option_override_internal): Replace generic32/64 by generic.
6627 (ix86_issue_rate): Likewise.
6628 (ix86_adjust_cost): Likewise.
6630 2013-09-19 Jan Hubicka <jh@suse.cz>
6632 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
6633 of speculative flag.
6635 2013-09-19 Jakub Jelinek <jakub@redhat.com>
6637 * omp-low.c (expand_omp_sections): Always pass len - 1 to
6638 GOMP_sections_start, even if !exit_reachable.
6640 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
6642 * lra-constraints.c (need_for_all_save_p): Use macro
6643 HARD_REGNO_CALL_PART_CLOBBERED.
6644 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
6645 set up pseudo conflict hard regs.
6647 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6650 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
6653 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
6655 PR rtl-optimization/58438
6656 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
6657 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
6658 from previous subpasses.
6660 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
6662 * arm.c (arm_get_frame_offsets): Validate architecture supports
6663 LDRD/STRD before accepting the tuning preference.
6664 (arm_expand_prologue): Likewise.
6665 (arm_expand_epilogue): Likewise.
6667 2013-09-18 Richard Biener <rguenther@suse.de>
6669 PR tree-optimization/58417
6670 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
6671 have chrecs with symbols defined in the loop as operands.
6672 (chrec_fold_multiply): Likewise.
6673 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
6674 parameters before folding binary operations.
6675 (struct instantiate_cache_entry_hasher): Remove.
6676 (struct instantiate_cache_type): Use a pointer-map.
6677 (instantiate_cache_type::instantiate_cache_type): New function.
6678 (instantiate_cache_type::get): Likewise.
6679 (instantiate_cache_type::set): Likewise.
6680 (instantiate_cache_type::~instantiate_cache_type): Adjust.
6681 (get_instantiated_value_entry): Likewise.
6682 (global_cache): New global.
6683 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
6684 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
6685 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
6686 (instantiate_scev_name): Adjust.
6687 (instantiate_scev): Construct global instead of local cache.
6688 (resolve_mixers): Likewise.
6690 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
6691 Paolo Carlini <paolo.carlini@oracle.com>
6694 * doc/implement-cxx.texi: Fix references to the C++ standards.
6696 2013-09-18 Jakub Jelinek <jakub@redhat.com>
6698 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
6699 * tree-vect-data-refs.c (vect_analyze_data_refs): For
6700 simd_lane_access drs, update also DR_ALIGNED_TO.
6702 2013-09-18 Marek Polacek <polacek@redhat.com>
6705 * doc/extend.texi: Document no_sanitize_undefined attribute.
6706 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
6707 no_sanitize_undefined attribute.
6709 2013-09-18 Nick Clifton <nickc@redhat.com>
6711 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
6712 (ASM_DECLARE_FUNCTION_NAME): Define.
6714 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
6716 * compare-elim.c (find_comparison_dom_walker): New class
6717 (find_comparisons_in_bb): Rename to
6718 find_comparison_dom_walker::before_dom_children
6719 (find_comparisons): Adjust
6720 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
6722 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
6723 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
6724 (init_walk_dominator_tree): Remove declaration.
6725 (fini_walk_dominator_tree): Remove declaration.
6726 * fwprop.c (single_def_use_dom_walker): New class
6727 (single_def_use_enter_block): Convert to
6728 single_def_use_dom_walker::before_dom_children.
6729 (single_def_use_leave_block): Convert to
6730 single_def_use_dom_walker::after_dom_children.
6731 (build_single_def_use_links): Adjust.
6732 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
6734 (find_candidates_in_block): Convert to
6735 find_candidates_dom_walker::before_dom_children.
6736 (execute_strength_reduction): Adjust.
6737 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
6738 (sese_dom_walker): New class.
6739 (sese_dom_walker::sese_dom_walker): New constructor.
6740 (sese_dom_walker::~sese_dom_walker): New destructor.
6741 (build_sese_conditions_before): Convert to
6742 sese_dom_walker::before_dom_children.
6743 (build_sese_conditions_after): Convert to
6744 sese_dom_walker::after_dom_children.
6745 (build_poly_scop): Adjust
6746 * tree-into-ssa.c (rewrite_dom_walker): New class
6747 (rewrite_enter_block): Convert to
6748 rewrite_dom_walker::before_dom_children.
6749 (rewrite_leave_block): Convert to
6750 rewrite_dom_walker::after_dom_children.
6751 (rewrite_update_dom_walker): New class.
6752 (rewrite_update_enter_block): Convert to
6753 rewrite_update_dom_walker::before_dom_children.
6754 (rewrite_update_leave_block): Convert to
6755 rewrite_update_dom_walker::after_dom_children.
6756 (rewrite_blocks, rewrite_into_ssa): Adjust.
6757 (mark_def_dom_walker): New class.
6758 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
6759 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
6760 (mark_def_sites_blocks): Convert to
6761 mark_def_dom_walker::before_dom_children.
6762 (mark_def_site_blocks): Remove.
6763 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
6764 (tree_ssa_dominator_optimize): Adjust.
6765 (dom_thread_across_edge): Convert to method
6766 dom_opt_dom_walker::thread_across_edge.
6767 (dom_opt_enter_block): Convert to member function
6768 dom_opt_dom_walker::before_dom_children.
6769 (dom_opt_leave_block): Convert to member function
6770 dom_opt_dom_walker::after_dom_children.
6771 * tree-ssa-dse.c (dse_dom_walker): New class.
6772 (dse_enter_block): Convert to member function
6773 dse_dom_walker::before_dom_children.
6774 (tree_ssa_dse): Adjust.
6775 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
6776 (determine_invariantness_stmt): Convert to method
6777 invariantness_dom_walker::before_dom_children.
6778 (determine_invariantness): Remove
6779 (move_computations_dom_walker): New class.
6780 (move_computations_stmt): Convert to method
6781 move_computations_dom_walker::before_dom_children.
6782 (move_computations, tree_ssa_lim): Adjust.
6783 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
6784 (nt_init_block): Convert to method
6785 notrappping_dom_walker::before_dom_children.
6786 (nt_fini_block): Convert to method
6787 method nontrapping_dom_walker::after_dom_children.
6788 (get_non_trapping): Adjust.
6789 * tree-ssa-pre.c (eliminate_dom_walker): New class.
6790 (eliminate_bb): Convert to method
6791 eliminate_dom_walker::before_dom_children.
6792 (eliminate_leave_block): Convert to method
6793 eliminate_dom_walker::after_dom_children.
6794 (eliminate): Adjust.
6795 * tree-ssa-strlen.c (strlen_dom_walker): New class.
6796 (strlen_enter_block): Convert to method
6797 strlen_dom_walker::before_dom_children.
6798 (strlen_leave_block): Convert to method
6799 method strlen_dom_walker::after_dom_children.
6800 (tree_ssa_strlen): Adjust.
6801 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
6802 (tree_ssa_uncprop): Adjust.
6803 (uncprop_leave_block): Convert to method
6804 uncprop_dom_walker::after_dom_children.
6805 (uncprop_leave_block): Convert to method
6806 uncprop_dom_walker::before_dom_children.
6808 2013-09-18 Bin Cheng <bin.cheng@arm.com>
6810 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
6811 branch in same basic block. Check both src and dest of the move insn.
6813 2013-09-17 Nick Clifton <nickc@redhat.com>
6815 * config/rl78/rl78-real.md (bf): New pattern.
6817 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
6818 (rl78_print_operand): Do not put a # before a %B.
6819 * config/rl78/rl78.opt: Tweak doc strings.
6821 2013-09-17 DJ Delorie <dj@redhat.com>
6823 * config/rl78/constraints.md (Wcv): Allow up to $r31.
6824 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
6825 (rl78_option_override): Likewise, if -mallregs.
6826 (is_virtual_register): Likewise.
6827 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
6828 (REGNO_OK_FOR_BASE_P): Likewise.
6829 * config/rl78/rl78.opt (-mallregs): New.
6831 2013-09-17 Nick Clifton <nickc@redhat.com>
6833 * config/rl78/rl78.c (need_to_save): Change return type to bool.
6834 For interrupt functions: save all call clobbered registers if the
6835 interrupt handler is not a leaf function.
6836 (rl78_expand_prologue): Always recompute the frame information.
6837 For interrupt functions: only select bank 0 if one of the bank 0
6838 registers is going to be psuhed.
6840 2013-09-17 DJ Delorie <dj@redhat.com>
6842 * config/rl78/constraints.md: For each W* constraint, rename to C*
6843 and create a W* constraint that checks for an optional ES: prefix
6845 * config/rl78/rl78.md (UNS_ES_ADDR): New.
6846 (es_addr): New. Used to wrap far addresses.
6847 * config/rl78/rl78-protos.h (rl78_es_addr): New.
6848 (rl78_es_base): New.
6849 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
6850 wrapped far addresses.
6851 (rl78_print_operand_1): Unwrap far addresses before processing.
6852 (rl78_lo16): Wrap far addresses in unspecs.
6853 (rl78_es_addr): New.
6854 (rl78_es_base): New.
6855 (insn_ok_now): Check for not-yet-wrapped far addresses.
6856 (transcode_memory_rtx): Properly re-wrap far addresses.
6858 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
6860 * config/sparc/t-rtems: Add leon3 multilibs.
6862 2013-09-17 Cong Hou <congh@google.com>
6864 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
6865 when checking the dot production pattern. The type of rhs operand
6866 of multiply is now checked correctly.
6868 2013-09-17 Jeff Law <law@redhat.com>
6870 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6871 edge implied equivalences into successor phis.
6872 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
6873 here from tree-ssa-threadedge.c.
6874 (mark_threaded_blocks): When threading through a joiner, if both
6875 successors of the joiner's clone reach the same block, verify the
6876 PHI arguments are equal. If not, cancel the jump threading request.
6877 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
6878 tree-ssa-threadupdate.c
6879 (thread_across_edge): Don't check PHI argument equality when
6880 threading through joiner block here.
6882 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
6884 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
6885 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
6886 (warn_uninit, warn_uninitialized_vars,
6887 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
6888 Move to tree-ssa-uninit.c.
6889 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
6890 (has_undefined_value_p): New. Pass dependant parts of
6891 ssa_undefined_value_p.
6892 (uninit_undefined_value_p): Use has_undefined_value_p.
6893 (warn_uninit, warn_uninitialized_vars,
6894 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
6895 Move from tree-ssa.c.
6896 * tree-ssa.h: Adjust prototypes.
6898 2013-09-17 Jan Hubicka <jh@suse.cz>
6901 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
6902 * ipa-inline.c (can_inline_edge_p): Do not downgrade
6903 FUNCTION_NOT_OPTIMIZED.
6904 * ipa-inline-analysis.c (compute_inline_parameters): Function
6905 not optimized is not inlinable unless it is alwaysinline.
6906 (inline_analyze_function): Force calls in not optimized
6907 function not inlinable.
6909 2013-09-17 Jan Hubicka <jh@suse.cz>
6912 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
6914 * ipa.c (function_and_variable_visibility): Likewise.
6915 * ipa-profile.c (ipa_profile): Likewise.
6917 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
6920 * cgraph.c (cgraph_function_body_availability): Check for ifunc
6921 attribute, and don't inline the resolver in this case.
6923 2013-09-17 Teresa Johnson <tejohnson@google.com>
6925 * coverage.c (get_coverage_counts): Add missing newline.
6927 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6929 PR tree-optimization/58088
6930 * fold-const.c (mask_with_trailing_zeros): New function.
6931 (fold_binary_loc): Make sure we don't recurse infinitely
6932 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
6933 Use mask_with_trailing_zeros where appropriate.
6935 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
6937 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
6938 of 'prev' var to get better distance estimation.
6940 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
6942 * tree-inline.h (struct copy_body_data): Add transform_parameter.
6943 * tree-inline.c (is_parameter_of): New predicate.
6944 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
6945 a parameter has been remapped.
6946 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
6947 (optimize_inline_calls): Initialize transform_parameter.
6948 (copy_gimple_seq_and_replace_locals): Likewise.
6949 (tree_function_versioning): Likewise.
6950 (maybe_inline_call_in_expr): Likewise.
6952 2013-09-17 Nick Clifton <nickc@redhat.com>
6954 * config/msp430/msp430-protos.h: Add prototypes for new functions.
6955 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
6957 (is_attr_func): New function.
6958 (msp430_is_interrupt_func): New function.
6959 (is_naked_func): New function.
6960 (is_reentrant_func): New function.
6961 (is_critical_func): New function.
6962 (msp430_start_function): Add annotations for function attributes.
6963 (msp430_attr): New function.
6964 (msp430_attribute_table): New.
6965 (msp430_function_section): New function.
6966 (TARGET_ASM_FUNCTION_SECTION): Define.
6967 (msp430_builtin): New enum.
6968 (msp430_init_builtins): New function.
6969 (msp430_builtin_devl): New function.
6970 (msp430_expand_builtin): New function.
6971 (TARGET_INIT_BUILTINS): Define.
6972 (TARGET_EXPAND_BUILTINS): Define.
6973 (TARGET_BUILTIN_DECL): Define.
6974 (msp430_expand_prologue): Add support for naked, interrupt,
6975 critical and reentrant functions.
6976 (msp430_expand_epilogue): Likewise.
6977 (msp430_print_operand): Handle 'O' character.
6978 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
6980 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
6981 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
6982 (pushm): Use a 'n' rather than an 'i' constraint.
6983 (msp_return): Add generation of the interrupt return instruction.
6984 (disable_interrupts): New pattern.
6985 (enable_interrupts): New pattern.
6986 (push_intr_state): New pattern.
6987 (pop_intr_state): New pattern.
6988 (bic_SR): New pattern.
6989 (bis_SR): New pattern.
6990 * doc/extend.texi: Document MSP430 function attributes and builtin
6993 2013-09-17 Richard Biener <rguenther@suse.de>
6995 PR tree-optimization/58432
6996 * tree-loop-distribution.c (tree_loop_distribution): Also
6997 scan PHIs for outside loop uses and seed a partition from them.
6999 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7001 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7002 (restructure_reference): Call backtrace_base_for_ref.
7004 2013-09-17 Alan Modra <amodra@gmail.com>
7007 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7010 2013-09-16 DJ Delorie <dj@redhat.com>
7012 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7013 vregs location for RL78/G10.
7014 (rl78_expand_prologue): Avoid SEL on G10.
7015 (rl78_expand_epilogue): Likewise.
7016 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7017 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7018 __RL78_G10__ when appropriate.
7019 (ASM_SPEC): Pass -mg10 along to the assembler.
7020 * config/rl78/rl78.md (sel_rb): Disable for G10.
7021 * config/rl78/rl78.opt: Add -mg10 option.
7022 * config/rl78/t-rl78: Add -mg10 multilib.
7024 2013-09-16 Xinliang David Li <davidxl@google.com>
7026 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7027 * omp-low.c (omp_max_vf): Ditto.
7028 (expand_omp_simd): Ditto.
7029 * tree-vectorizer.c (vectorize_loops): Ditto.
7030 (gate_vect_slp): Ditto.
7031 (gate_increase_alignment): Ditto.
7032 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7033 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7034 (gate_tree_vectorize): Name change.
7035 (tree_vectorize): Ditto.
7036 (pass_vectorize::gate): Call new function.
7037 (pass_vectorize::execute): Ditto.
7038 * opts.c: O3 default setting change.
7039 (finish_options): Check new flag.
7040 * doc/invoke.texi: Document new flags.
7041 * common.opt: New flags.
7043 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7045 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7046 (Condition Code): Don't reference it.
7047 * doc/tm.texi: Regenerate.
7049 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7052 * lra-constraints.c (undo_optional_reloads): Consider all optional
7053 reload even if it did not get a hard reg.
7055 2013-09-16 Teresa Johnson <tejohnson@google.com>
7057 * dumpfile.c (dump_loc): Remove newline emission.
7058 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7059 emission to dump_printf_loc calls where missing.
7060 (vect_mark_for_runtime_alias_test): Ditto.
7061 (vect_analyze_data_ref_dependence): Ditto.
7062 (vect_analyze_data_ref_dependences): Ditto.
7063 (vect_slp_analyze_data_ref_dependence): Ditto.
7064 (vect_slp_analyze_data_ref_dependences): Ditto.
7065 (vect_compute_data_ref_alignment): Ditto.
7066 (vect_update_misalignment_for_peel): Ditto.
7067 (vect_verify_datarefs_alignment): Ditto.
7068 (vector_alignment_reachable_p): Ditto.
7069 (vect_get_data_access_cost): Ditto.
7070 (vect_enhance_data_refs_alignment): Ditto.
7071 (vect_find_same_alignment_drs): Ditto.
7072 (vect_analyze_data_refs_alignment): Ditto.
7073 (vect_analyze_group_access): Ditto.
7074 (vect_analyze_data_ref_access): Ditto.
7075 (vect_analyze_data_ref_accesses): Ditto.
7076 (vect_prune_runtime_alias_test_list): Ditto.
7077 (vect_analyze_data_refs): Ditto.
7078 (vect_create_addr_base_for_vector_ref): Ditto.
7079 (vect_create_data_ref_ptr): Ditto.
7080 (vect_grouped_store_supported): Ditto.
7081 (vect_grouped_load_supported): Ditto.
7082 * value-prof.c (check_counter): Ditto.
7083 (check_ic_target): Ditto.
7084 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7085 (vect_recog_widen_mult_pattern): Ditto.
7086 (vect_recog_widen_sum_pattern): Ditto.
7087 (vect_recog_over_widening_pattern): Ditto.
7088 (vect_recog_widen_shift_pattern): Ditto.
7089 (vect_recog_rotate_pattern): Ditto.
7090 (vect_recog_vector_vector_shift_pattern): Ditto.
7091 (vect_recog_divmod_pattern): Ditto.
7092 (vect_recog_mixed_size_cond_pattern): Ditto.
7093 (vect_recog_bool_pattern): Ditto.
7094 (vect_pattern_recog_1): Ditto.
7095 (vect_pattern_recog): Ditto.
7096 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7097 (vect_is_simple_iv_evolution): Ditto.
7098 (vect_analyze_scalar_cycles_1): Ditto.
7099 (vect_get_loop_niters): Ditto.
7100 (vect_analyze_loop_1): Ditto.
7101 (vect_analyze_loop_form): Ditto.
7102 (vect_analyze_loop_operations): Ditto.
7103 (vect_analyze_loop_2): Ditto.
7104 (vect_analyze_loop): Ditto.
7105 (report_vect_op): Ditto.
7106 (vect_is_slp_reduction): Ditto.
7107 (vect_is_simple_reduction_1): Ditto.
7108 (vect_get_known_peeling_cost): Ditto.
7109 (vect_estimate_min_profitable_iters): Ditto.
7110 (vect_model_reduction_cost): Ditto.
7111 (vect_model_induction_cost): Ditto.
7112 (get_initial_def_for_induction): Ditto.
7113 (vect_create_epilog_for_reduction): Ditto.
7114 (vectorizable_reduction): Ditto.
7115 (vectorizable_induction): Ditto.
7116 (vectorizable_live_operation): Ditto.
7117 (vect_loop_kill_debug_uses): Ditto.
7118 (vect_transform_loop): Ditto.
7119 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7120 (vect_stmt_relevant_p): Ditto.
7121 (process_use): Ditto.
7122 (vect_mark_stmts_to_be_vectorized): Ditto.
7123 (vect_model_simple_cost): Ditto.
7124 (vect_model_promotion_demotion_cost): Ditto.
7125 (vect_model_store_cost): Ditto.
7126 (vect_get_store_cost): Ditto.
7127 (vect_model_load_cost): Ditto.
7128 (vect_get_load_cost): Ditto.
7129 (vect_init_vector_1): Ditto.
7130 (vect_get_vec_def_for_operand): Ditto.
7131 (vect_finish_stmt_generation): Ditto.
7132 (vectorizable_call): Ditto.
7133 (vectorizable_conversion): Ditto.
7134 (vectorizable_assignment): Ditto.
7135 (vectorizable_shift): Ditto.
7136 (vectorizable_operation): Ditto.
7137 (vectorizable_store): Ditto.
7138 (vectorizable_load): Ditto.
7139 (vectorizable_condition): Ditto.
7140 (vect_analyze_stmt): Ditto.
7141 (vect_transform_stmt): Ditto.
7142 (vect_is_simple_use): Ditto.
7143 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7144 (vect_can_advance_ivs_p): Ditto.
7145 (vect_update_ivs_after_vectorizer): Ditto.
7146 (vect_do_peeling_for_loop_bound): Ditto.
7147 (vect_gen_niters_for_prolog_loop): Ditto.
7148 (vect_update_inits_of_drs): Ditto.
7149 (vect_create_cond_for_alias_checks): Ditto.
7150 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7151 (vect_build_slp_tree_1): Ditto.
7152 (vect_supported_load_permutation_p): Ditto.
7153 (vect_analyze_slp_instance): Ditto.
7154 (vect_analyze_slp): Ditto.
7155 (vect_make_slp_decision): Ditto.
7156 (vect_detect_hybrid_slp): Ditto.
7157 (vect_bb_vectorization_profitable_p): Ditto.
7158 (vect_slp_analyze_bb_1): Ditto.
7159 (vect_update_slp_costs_according_to_vf): Ditto.
7160 (vect_get_mask_element): Ditto.
7161 (vect_transform_slp_perm_load): Ditto.
7162 (vect_schedule_slp_instance): Ditto.
7163 (vect_schedule_slp): Ditto.
7164 (vect_slp_transform_bb): Ditto.
7165 * profile.c (read_profile_edge_counts): Ditto.
7166 (compute_branch_probabilities): Ditto.
7167 * coverage.c (get_coverage_counts): Ditto.
7169 2013-09-16 Diego Novillo <dnovillo@google.com>
7171 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7173 2013-09-16 Jan Hubicka <jh@suse.cz>
7175 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7176 abstract functions; for static functions check the presence of body.
7178 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7180 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7181 * config/aarch64/aarch64-simd.md
7182 (aarch64_mla_elt<mode>): New.
7183 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7184 (aarch64_mls_elt<mode>): Likewise.
7185 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7186 (aarch64_fma4_elt<mode>): Likewise.
7187 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7188 (aarch64_fma4_elt_to_128v2df): Likewise.
7189 (aarch64_fma4_elt_to_64df): Likewise.
7190 (fnma<mode>4): Likewise.
7191 (aarch64_fnma4_elt<mode>): Likewise.
7192 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7193 (aarch64_fnma4_elt_to_128v2df): Likewise.
7194 (aarch64_fnma4_elt_to_64df): Likewise.
7195 * config/aarch64/iterators.md (VDQSF): New.
7196 * config/aarch64/arm_neon.h
7197 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
7198 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
7200 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7202 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
7203 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
7204 (aarch64_mul3_elt_to_128df): Likewise.
7205 (aarch64_mul3_elt_to_64v2df): Likewise.
7206 * config/aarch64/iterators.md (VEL): Also handle DFmode.
7208 (VMUL_CHANGE_NLANES) Likewise.
7211 * config/aarch64/arm_neon.h
7212 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
7214 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7216 * config/aarch64/arm_neon.h
7217 (vcvtx_high_f32_f64): Fix parameters.
7219 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7220 Uros Bizjak <ubizjak@gmail.com>
7222 * config/alpha.c: Include tree-ssanames.h.
7224 2013-09-16 Richard Biener <rguenther@suse.de>
7226 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
7227 (dot_rdg_1): Handle control_dd.
7228 (create_edge_for_control_dependence): New function.
7229 (create_rdg_edges): Add control dependences if asked for.
7230 (build_rdg): Likewise.
7231 (generate_loops_for_partition): If there are not necessary
7232 control stmts remove all their dependencies.
7233 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
7234 (distribute_loop): Pass on control dependences.
7235 (tree_loop_distribution): Compute control dependences and remove
7236 restriction on number of loop nodes.
7238 2013-09-16 Jakub Jelinek <jakub@redhat.com>
7240 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
7243 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
7245 * cse.c (try_const_anchors): Punt on CC modes.
7247 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7249 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
7251 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
7254 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
7257 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7260 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
7262 (darwin_objc1_section): Likewise.
7263 (darwin_file_end): Emit Image Info section when required.
7265 2013-09-14 Jan Hubicka <jh@suse.cz>
7267 * tree-into-ssa.c (gate_into_ssa): New.
7268 (pass_data_build_ssa): Use it.
7269 * cgraph.h (expand_thunk): Update prototype.
7270 * cgraphunit.c (analyze_function): Expand thunks early.
7271 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
7272 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
7273 set lowered flag; do not add new function.
7274 (assemble_thunks_and_aliases): Update.
7275 * tree-ssa.c (gate_init_datastructures): New gate.
7276 (pass_data_init_datastructures): Use it.
7278 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7281 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
7282 xmm register set described in the psABI.
7284 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
7286 * dwarf2out.c (should_emit_struct_debug): Add check
7287 for type_decl variable is not NULL.
7289 2013-09-13 Jacek Caban <jacek@codeweavers.com>
7291 * config.gcc: Use new winnt-c.c target hooks
7292 * config/t-winnt: New file
7293 * config/winnt-c.c: New file
7294 * doc/tm.texi.in: Document new hook
7295 * doc/tm.texi: Regenerated
7297 2013-09-13 Jan Hubicka <jh@suse.cz>
7300 * ipa-inline.c (check_callers): New function.
7301 (check_caller_edge): Remove.
7302 (want_inline_function_to_all_callers_p): Also permit alises that are
7304 (inline_to_all_callers): Terminate the walk when devirtualization
7305 introduce new calls.
7307 2013-09-13 Jan Hubicka <jh@suse.cz>
7309 * ipa-inline-analysis.c (struct growth_data): Add node.
7310 (do_estimate_growth_1): Fix detection of recursion.
7312 2013-09-13 Jakub Jelinek <jakub@redhat.com>
7314 PR tree-optimization/58392
7315 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
7316 to avoid shadowing of outer loop variable. If
7317 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
7318 replace_by_duplicate_decl simduid of loops that have it set and
7319 set dest_cfun->has_simduid_loops and/or
7320 dest_cfun->has_force_vect_loops.
7321 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
7322 instead of maybe_lookup_decl.
7323 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
7324 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
7325 call. Copy over force_vect and copy and remap simduid. Set
7326 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
7327 (copy_cfg_body): Remove blocks_to_copy argument. Use
7328 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
7329 caller. Don't set cfun->has_simduid_loops and/or
7330 cfun->has_force_vect_loops here.
7331 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
7333 (expand_call_inline, tree_function_versioning): Adjust copy_body
7336 2013-09-13 Martin Jambor <mjambor@suse.cz>
7339 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
7340 the assert if the edge was a speculative one.
7342 2013-09-13 Richard Biener <rguenther@suse.de>
7344 * tree-data-ref.h (known_dependences_p): Move here ...
7345 * tree-loop-distribution.c (known_dependences_p): ... from here.
7346 (dump_rdg_component, debug_rdg_component): Remove.
7348 (generate_loops_for_partition): Use gimple_uid instead of
7349 relying on matching stmt visit order.
7350 (rdg_build_partitions): Take starting stmt vector.
7351 (ldist_gen): Merge into ...
7352 (distribute_loop): ... this function. Do not compute starting
7354 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
7356 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7358 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
7359 Set type attribute correctly. Set predicable_short_it attribute.
7360 (cmpsi_shiftsi): Remove %? from output template.
7362 2013-09-13 Richard Biener <rguenther@suse.de>
7364 * tree-loop-distribution.c (struct rdg_component,
7365 rdg_defs_used_in_other_loops_p, free_rdg_components,
7366 rdg_build_components): Remove.
7367 (stmts_from_loop): Do not record virtual PHIs.
7368 (generate_loops_for_partition): Skip virtual PHIs.
7369 (build_rdg_partition_for_component): Rename to ...
7370 (build_rdg_partition_for_vertex): ... this and adjust.
7371 (rdg_build_partitions): Take a vector of starting vertices
7372 instead of components. Remove unnecessary leftover handling.
7373 (ldist_gen): Do not build components or record other stores.
7374 (distribute_loop): Do not distribute loops containing stmts
7377 2013-09-13 Christian Bruel <christian.bruel@st.com>
7380 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
7382 2013-09-13 Kai Tietz <ktietz@redhat.com>
7384 * config.gcc: Separate cases for mingw and cygwin targets,
7385 and add 64-bit cygwin target case.
7387 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
7388 dll-export inline-functions.
7389 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
7392 2013-09-13 Jeff Law <law@redhat.com>
7396 2013-09-06 Jeff Law <law@redhat.com>
7398 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7399 edge implied equivalences into successor phis.
7401 2013-09-12 DJ Delorie <dj@redhat.com>
7403 * config/rl78/rl78-virt.md: Change from | to \; for asm line
7406 2013-09-12 Brooks Moses <bmoses@google.com>
7409 * Makefile.in: Do not install driver binaries in $(target)/bin.
7411 2013-09-12 DJ Delorie <dj@redhat.com>
7413 * config/rl78/rl78.opt (mrelax): New.
7414 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
7415 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
7417 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
7419 (rl78_expand_epilogue): Likewise.
7421 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
7424 * lra-eliminations.c (remove_reg_equal_offset_note): New.
7425 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
7426 pointer elimination with using remove_reg_equal_offset_note.
7428 2013-09-12 DJ Delorie <dj@redhat.com>
7430 * config/msp430/: New port.
7431 * config.gcc (msp430): Added.
7432 * doc/invoke.texi: Document MSP430 options.
7433 * doc/install.texi: Document msp430-elf
7434 * doc/md.texi: Document msp430-elf
7435 * doc/contrib.texi: Document msp430-elf
7437 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
7440 2013-09-12 Martin Jambor <mjambor@suse.cz>
7443 * ipa-prop.c (remove_described_reference): Give up if the edge in the
7444 reference descriptor is NULL.
7445 (ipa_edge_removal_hook): If owning a reference descriptor, set its
7448 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
7450 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
7451 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
7452 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
7453 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
7454 make_temp_ssa_name): move to tree-ssanames.h
7455 * tree-ssa-alias.h: Move prototype.
7456 * tree-ssa.h: Include tree-ssanames.h.
7457 * tree-ssanames.c (FREE_SSANAMES): Move to here.
7458 * tree-ssanames.h: New. Move items from tree-flow*.h
7459 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
7461 2013-09-12 Richard Biener <rguenther@suse.de>
7463 PR tree-optimization/58404
7464 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
7465 propagate non-invariant addresses into dereferences wrapped
7466 in component references.
7468 2013-09-12 Richard Biener <rguenther@suse.de>
7470 PR tree-optimization/58402
7471 * passes.def: Move pass_late_warn_uninitialized later.
7473 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
7475 * tree-ssa.h: New. Move content from tree-flow.h and
7477 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
7478 Move prototypes belonging to tree-ssa.c.
7479 * tree-flow-inline.h (redirect_edge_var_map_def,
7480 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
7482 * gimple.h: Adjust prototypes.
7483 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
7485 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7486 * tree.h: Move prototype to tree-ssa.h.
7487 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
7488 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
7489 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
7490 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
7491 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
7492 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
7493 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
7494 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
7495 graphite-blocking.c, graphite-clast-to-gimple.c,
7496 graphite-dependences.c, graphite-interchange.c,
7497 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
7498 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
7499 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
7500 ipa-reference.c, ipa-split.c, ipa-utils.c,
7501 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
7502 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
7503 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
7504 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
7505 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
7506 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
7507 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
7508 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
7509 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
7510 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
7511 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
7512 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
7513 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
7514 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
7515 value-prof.c, var-tracking.c,
7516 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
7518 2013-09-12 Richard Biener <rguenther@suse.de>
7520 PR tree-optimization/58396
7521 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
7522 (build_rdg): Take a loop-nest parameter, fix memleaks.
7523 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
7525 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
7527 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
7530 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
7532 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
7533 The FMA instruction names should have a 'v' prefix.
7535 2013-09-12 Richard Biener <rguenther@suse.de>
7537 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
7538 (dot_rdg): Use popen instead of system in optional code.
7539 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
7540 (already_processed_vertex_p): Adjust.
7541 (has_anti_or_output_dependence, predecessor_has_mem_write,
7542 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
7543 rdg_flag_uses): Remove.
7544 (rdg_flag_vertex): Simplify.
7545 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
7547 (build_rdg_partition_for_component): Process the first vertex
7548 of a component only.
7549 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
7551 2013-09-12 Alan Modra <amodra@gmail.com>
7553 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
7555 2013-09-11 DJ Delorie <dj@redhat.com>
7556 Nick Clifton <nickc@redhat.com>
7558 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
7559 (rl78_stack_based_mem): New.
7560 * config/rl78/constraints.md (Iv08): New.
7569 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
7571 (movsi): Change from expand to insn-and-split.
7572 (ashrsi3): Clobber AX.
7576 * config/rl78/rl78.md (CC_REG): Fix.
7577 (addsi3): Allow memory and immediate operands.
7578 (addsi3_internal): Split into...
7579 (addsi3_internal_virt): ...new, and ...
7580 (addsi3_internal_real): ...new.
7582 (subsi3_internal_virt): New.
7583 (subsi3_internal_real): New.
7584 (mulsi3): Add memory operand.
7585 (mulsi3_rl78): Likewise.
7586 (mulsi3_g13): Likewise.
7587 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
7588 (cbranchqi4_real): Add more constraint options.
7589 (cbranchhi4_real): Expand pattern.
7590 (cbranchhi4_real_signed): New.
7591 (cbranchhi4_real_inverted): New.
7592 (cbranchsi4_real_lt): New.
7593 (cbranchsi4_real_ge): New.
7594 (cbranchsi4_real_signed): New.
7595 (cbranchsi4_real): New.
7597 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
7599 (lshrsi3_virt): Likewise.
7600 (ashlsi3_virt): Likewise.
7601 (cbranchqi4_virt_signed): New.
7602 (cbranchhi4_virt_signed): New.
7603 (cbranchsi4_virt): New.
7604 * config/rl78/rl78.c: Whitespace fixes throughout.
7605 (move_elim_pass): New.
7606 (pass_data_rl78_move_elim): New.
7607 (pass_rl78_move_elim): New.
7608 (make_pass_rl78_move_elim): New.
7609 (rl78_devirt_info): Run devirt earlier.
7610 (rl78_move_elim_info): New.
7611 (rl78_asm_file_start): Register it.
7612 (rl78_split_movsi): New.
7613 (rl78_as_legitimate_address): Allow virtual base registers when
7615 (rl78_addr_space_convert): Remove spurious debug stuff.
7616 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
7617 (rl78_print_operand): More cases for not printing '#'.
7618 (rl78_expand_compare): Remove most of the logic.
7619 (content_memory): New.
7620 (clear_content_memory): New.
7621 (get_content_index): New.
7622 (get_content_name): New.
7623 (display_content_memory): New.
7624 (update_content): New.
7625 (record_content): New.
7626 (already_contains): New.
7627 (insn_ok_now): Re-recog insns with virtual registers.
7628 (add_postponed_content_update): New.
7629 (process_postponed_content_update): New.
7630 (gen_and_emit_move): New.
7631 (transcode_memory_rtx): Record new location content.
7632 Use gen_and_emit_move.
7633 (force_into_acc): New.
7634 (move_to_acc): Use gen_and_emit_move.
7635 (move_from_acc): Likewise.
7636 (move_acc_to_reg): Likewise.
7637 (move_to_x): Likewise.
7638 (move_to_hl): Likewise.
7639 (move_to_de): Likewise.
7640 (rl78_alloc_physical_registers_op1): Record location content.
7641 (has_constraint): New.
7642 (rl78_alloc_physical_registers_op2): Record location content.
7644 (rl78_alloc_physical_registers_ro1): Likewise.
7645 (rl78_alloc_physical_registers_cmp): Likewise.
7646 (rl78_alloc_physical_registers_umul): Likewise.
7647 (rl78_alloc_address_registers_macax): New.
7648 (rl78_alloc_physical_registers): Initialize and set location
7649 content memory as needed.
7650 (rl78_reorg): Make sure split2 is called.
7651 (rl78_rtx_costs): New.
7653 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
7655 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
7656 for (not (neg ...)) and (neg (not ...)) cases.
7658 2013-09-11 Richard Biener <rguenther@suse.de>
7661 * passes.def: Split critical edges before late uninit warning passes.
7662 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
7664 2013-09-11 Jakub Jelinek <jakub@redhat.com>
7666 PR tree-optimization/58385
7667 * fold-const.c (build_range_check): If both low and high are NULL,
7668 use omit_one_operand_loc to preserve exp side-effects.
7670 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7672 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
7674 2013-09-11 Richard Biener <rguenther@suse.de>
7676 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
7677 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
7678 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
7679 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
7680 stmts_from_loop, known_dependences_p, build_empty_rdg,
7681 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
7682 * tree-loop-distribution.c: ... here.
7683 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
7684 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
7685 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
7686 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
7687 * tree-loop-distribution.c: ... here.
7688 * tree-loop-distribution.c: Include gimple-pretty-print.h.
7689 (struct partition_s): Add loops member.
7690 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
7691 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
7692 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
7694 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7695 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7696 Sergey Lega <sergey.s.lega@intel.com>
7697 Anna Tikhonova <anna.tikhonova@intel.com>
7698 Ilya Tocar <ilya.tocar@intel.com>
7699 Andrey Turetskiy <andrey.turetskiy@intel.com>
7700 Ilya Verbin <ilya.verbin@intel.com>
7701 Kirill Yukhin <kirill.yukhin@intel.com>
7702 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7704 * config/i386/constraints.md (k): New.
7706 * config/i386/i386.c (const regclass_map): Add new mask registers.
7707 (dbx_register_map): Ditto.
7708 (dbx64_register_map): Ditto.
7709 (svr4_dbx_register_map): Ditto.
7710 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
7712 (ix86_preferred_reload_class): Disable constants for mask registers.
7713 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
7714 (ix86_hard_regno_mode_ok): Support new mask registers.
7715 (x86_order_regs_for_local_alloc): Ditto.
7716 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
7717 (FIXED_REGISTERS): Add new mask registers.
7718 (CALL_USED_REGISTERS): Ditto.
7719 (REG_ALLOC_ORDER): Ditto.
7720 (VALID_MASK_REG_MODE): New.
7721 (FIRST_MASK_REG): Ditto.
7722 (LAST_MASK_REG): Ditto.
7723 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
7724 (MAYBE_MASK_CLASS_P): New.
7725 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
7726 (REG_CLASS_CONTENTS): Ditto.
7727 (MASK_REGNO_P): New.
7728 (ANY_MASK_REG_P): Ditto.
7729 (HI_REGISTER_NAMES): Add new mask registers.
7730 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
7731 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
7733 (attribute "type"): Add mskmov, msklog.
7734 (attribute "length_immediate"): Support them.
7735 (attribute "memory"): Ditto.
7736 (attribute "prefix_0f"): Ditto.
7737 (*movhi_internal): Support new mask registers.
7738 (*movqi_internal): Ditto.
7739 (define_split): Split out clobber pattern is a logic
7740 insn on mask registers.
7741 (*k<logic><mode>): New.
7742 (*andhi_1): Extend to support mask regs.
7743 (*andqi_1): Extend to support mask regs.
7745 (define_split): Split and-not to and and not if operands
7747 (*<code><mode>_1): Separate HI mode to new pattern...
7748 (*<code>hi_1): This.
7749 (*<code>qi_1): Extend to support mask regs.
7751 (kortestzhi): Ditto.
7752 (kortestchi): Ditto.
7754 (*one_cmpl<mode>2_1): Remove HImode and handle it...
7755 (*one_cmplhi2_1): ...Here, now with mask registers support.
7756 (*one_cmplqi2_1): Support new mask registers.
7757 (HI/QImode arithmetics splitter): Don't split if mask registers
7759 (HI/QImode not splitter): Ditto.
7760 * config/i386/predicated.md (mask_reg_operand): New.
7761 (general_reg_operand): Ditto.
7763 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7765 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
7766 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
7768 2013-09-10 Jeff Law <law@redhat.com>
7770 PR tree-optimization/58380
7771 * tree-ssa-threadupdate.c (thread_block): Recognize another case
7772 of threading through a buried loop header.
7774 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
7775 return value for single successor case.
7777 2013-09-10 Jan Hubicka <jh@suse.cz>
7779 * ipa-devirt.c (ipa_devirt): Enable with LTO.
7781 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
7784 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
7785 support conditional execution.
7786 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
7788 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
7790 * lra.c (lra): Clear lra_optional_reload_pseudos before every
7792 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
7793 Check destination too to check move insn.
7794 (undo_optional_reloads): Add check that the original peudo did not
7795 changed its allocation and the optional reload was inherited on last
7796 inheritance pass. Break loop after deciding to keep optional reload.
7797 (lra_undo_inheritance): Add check that inherited pseudo still in
7800 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
7802 * config/aarch64/aarch64.md (generic_sched): New.
7803 * config/aarch64/aarch64-generic.md (load): Make conditional
7804 on generic_sched attribute.
7805 (nonload): Likewise.
7807 2013-09-10 Jan Hubicka <jh@suse.cz>
7809 * lto-cgraph.c: Include ipa-utils.h.
7810 (compute_ltrans_boundary): Also add possible targets into the boundary.
7812 2013-09-10 Jan Hubicka <jh@suse.cz>
7814 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
7815 VAR_DECL of vtable rather than full expression.
7817 2013-09-10 Jan Hubicka <jh@suse.cz>
7818 Paolo Carlini <paolo.carlini@oracle.com>
7820 * cgraphunit.c (analyze_functions): Save input_location, set it
7821 to UNKNOWN_LOCATION and restore it at the end.
7823 2013-09-10 Martin Jambor <mjambor@suse.cz>
7825 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
7826 represented by a thunk.
7828 2013-09-10 Jeff Law <law@redhat.com>
7830 PR tree-optimization/58343
7831 * tree-ssa-threadupdate.c (thread_block): Identify and disable
7832 jump threading requests through loop headers buried in the middle
7833 of a jump threading path.
7835 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
7836 in return value/type.
7838 2013-09-10 Jakub Jelinek <jakub@redhat.com>
7840 PR rtl-optimization/58365
7841 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
7842 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
7845 2013-09-10 Richard Biener <rguenther@suse.de>
7847 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
7848 * tree-data-ref.c (create_rdg_vertices): Collect all data
7849 references, signal failure to the caller, use data-ref API.
7850 (build_rdg): Compute data references only once. Maintain lifetime
7851 of data references and data dependences from within RDG.
7852 (free_rdg): Free dependence relations.
7853 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
7854 inventing extra dependences.
7855 (distribute_loop): Update for RDG API changes.
7857 2013-09-10 Kai Tietz <ktietz@redhat.com>
7859 * doc/invoke.texi (fms-extensions): Document changed
7860 behavior for ms-abi targets.
7861 * config/i386/i386.c (ix86_option_override_internal):
7862 Set default value of option -fms-extension for ms-abi targets.
7864 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7866 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
7868 2013-09-10 Alan Modra <amodra@gmail.com>
7871 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
7873 2013-09-10 Alan Modra <amodra@gmail.com>
7875 * config/rs6000/predicates.md (add_cint_operand): New.
7876 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
7877 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
7878 using add_cint_operand.
7879 (largetoc_high_plus_aix): Likewise.
7881 2013-09-09 Jakub Jelinek <jakub@redhat.com>
7883 PR tree-optimization/58364
7884 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
7885 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
7886 the current range can't be an unconditional true or false.
7888 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
7890 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
7892 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
7894 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
7896 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
7899 * doc/invoke.texi (-Wdelete-incomplete): Document it.
7901 2013-09-09 Ian Bolton <ian.bolton@arm.com>
7903 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
7904 NO_REGS for immediate that can't be moved directly into FP_REGS.
7906 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7908 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
7909 comparison with negated operand.
7910 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
7913 2013-09-09 Richard Biener <rguenther@suse.de>
7916 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
7917 of a subloop record all its block as affecting loop-closed SSA form.
7919 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7921 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
7922 of an rtx/bitpos pair.
7923 (store_fixed_bit_field): Update accordingly.
7925 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7927 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
7929 * builtins.c (expand_errno_check): Likewise.
7930 * dwarf2cfi.c (init_return_column_size): Likewise.
7931 * except.c (sjlj_mark_call_sites): Likewise.
7932 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
7933 * lra-constraints.c (emit_inc): Likewise.
7934 * ree.c (combine_set_extension): Likewise.
7935 * regmove.c (fixup_match_2): Likewise.
7936 * reload1.c (inc_for_reload): Likewise.
7938 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7940 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
7941 (force_to_mode, simplify_shift_const_1, simplify_comparison):
7942 Use gen_int_mode with the mode of the associated simplify_* call.
7943 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
7944 * expmed.c (expand_shift_1): Likewise.
7945 * function.c (instantiate_virtual_regs_in_insn): Likewise.
7946 * loop-iv.c (iv_number_of_iterations): Likewise.
7947 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
7948 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
7950 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7952 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
7953 of the associated expand_* call.
7954 (asan_emit_stack_protection): Likewise.
7955 * builtins.c (round_trampoline_addr): Likewise.
7956 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
7957 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
7958 (emit_store_flag): Likewise.
7959 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
7960 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
7962 * function.c (instantiate_virtual_regs_in_insn): Likewise.
7963 * ifcvt.c (noce_try_store_flag_constants): Likewise.
7964 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
7965 * modulo-sched.c (generate_prolog_epilog): Likewise.
7966 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
7967 (expand_ctz, expand_ffs, expand_unop): Likewise.
7969 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7971 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
7972 of the associated gen_rtx_* call.
7973 * caller-save.c (init_caller_save): Likewise.
7974 * combine.c (find_split_point, make_extraction): Likewise.
7975 (make_compound_operation): Likewise.
7976 * dwarf2out.c (mem_loc_descriptor): Likewise.
7977 * explow.c (plus_constant, probe_stack_range): Likewise.
7978 * expmed.c (expand_mult_const): Likewise.
7979 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
7980 * reload1.c (init_reload): Likewise.
7981 * valtrack.c (cleanup_auto_inc_dec): Likewise.
7982 * var-tracking.c (adjust_mems): Likewise.
7983 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
7984 rather than gen_rtx_fmt_ee.
7986 2013-09-09 Jan Hubicka <jh@suse.cz>
7989 * value-prof.c (gimple_ic): Copy also abnormal edges.
7991 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7993 * asan.c (asan_shadow_cst): Use gen_int_mode.
7995 2013-09-08 Jan Hubicka <jh@suse.cz>
7997 * ipa-profile.c: Add toplevel comment.
7998 (ipa_propagate_frequency_1): Be more conservative when profile is read.
7999 (contains_hot_call_p): New function.
8000 (ipa_propagate_frequency): Set frequencies based on counts when
8002 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8003 profile; do not tamper with profile after inlining if it is read.
8005 2013-09-08 Jan Hubicka <jh@suse.cz>
8007 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8010 2013-09-08 Jan Hubicka <jh@suse.cz>
8012 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8015 2013-09-08 Jeff Law <law@redhat.com>
8018 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8021 2013-09-08 Andi Kleen <ak@linux.intel.com>
8023 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8025 2013-09-08 Jan Hubicka <jh@suse.cz>
8027 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8029 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8031 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8032 for non-debug insns.
8033 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8034 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8035 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8038 2013-09-08 Jan Hubicka <jh@suse.cz>
8040 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8041 targets and devirtualize to BUILT_IN_UNREACHABLE.
8042 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8043 * ipa.c (walk_polymorphic_call_targets): New function.
8044 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8045 functions; use the new timevar.
8046 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8047 was removed from the program.
8048 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8049 not consider it in the walk when its vtable is dead.
8050 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8052 (devirt_variable_node_removal_hook): New function.
8053 (possible_polymorphic_call_targets): Also register
8054 devirt_variable_node_removal_hook.
8055 (ipa_devirt): Do not do non-speculative devirtualization.
8056 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8058 2013-09-08 Jan Hubicka <jh@suse.cz>
8060 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8061 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8062 varpool_remove_variable_insertion_hook): Declare.
8063 * varpool.c (varpool_node_hook_list): New structure.
8064 (first_varpool_node_removal_hook,
8065 first_varpool_variable_insertion_hook): New variables.
8066 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8067 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8068 varpool_remove_variable_insertion_hook,
8069 varpool_call_variable_insertion_hooks): New functions.
8070 (varpool_remove_node): Use it.
8072 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8075 * diagnostic.c (diagnostic_build_prefix): When s.file is
8076 "<built-in>" don't output line and column numbers.
8078 2013-09-06 Jan Hubicka <jh@suse.cz>
8080 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8081 * lto-streamer-out.c: Stream thunks, too.
8082 * lto-streamer-in.c (input_function): Pop cfun here
8083 (lto_read_body): Instead of here.
8085 2013-09-06 Caroline Tice <cmtice@google.com>
8087 * doc/install.texi: Add documentation for the --enable-vtable-verify
8088 and the --disable-libvtv configure options.
8090 2013-09-06 Jeff Law <law@redhat.com>
8092 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8093 edge implied equivalences into successor phis.
8095 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8097 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8099 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8101 * resource.c (mark_target_live_regs): Compute resources taking
8102 into account if a call is predicated or not.
8104 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8106 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8107 the compiler in the function names.
8109 2013-09-06 Jan Hubicka <jh@suse.cz>
8112 * ipa-inline.c (has_caller_p): New function.
8113 (want_inline_function_to_all_callers_p): Use it.
8114 (sum_callers, inline_to_all_callers): Break out from ...
8115 (ipa_inline): ... here.
8117 2013-09-06 Jan Hubicka <jh@suse.cz>
8119 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8120 only when AVX is enabled.
8122 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8124 * config/aarch64/aarch64.md
8125 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8126 is fpsimd_<load/store>2.
8127 (load_pair<mode>): Likewise.
8128 (store_pair<mode>): Likewise.
8130 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8132 * config/arm/types.md (type): Add "mrs" type.
8133 * config/aarch64/aarch64.md
8134 (aarch64_load_tp_hard): Make type "mrs".
8136 (load_tp_hard): Make type "mrs".
8137 * config/arm/cortex-a15.md: Update with new attributes.
8138 * config/arm/cortex-a5.md: Update with new attributes.
8139 * config/arm/cortex-a53.md: Update with new attributes.
8140 * config/arm/cortex-a7.md: Update with new attributes.
8141 * config/arm/cortex-a8.md: Update with new attributes.
8142 * config/arm/cortex-a9.md: Update with new attributes.
8143 * config/arm/cortex-m4.md: Update with new attributes.
8144 * config/arm/cortex-r4.md: Update with new attributes.
8145 * config/arm/fa526.md: Update with new attributes.
8146 * config/arm/fa606te.md: Update with new attributes.
8147 * config/arm/fa626te.md: Update with new attributes.
8148 * config/arm/fa726te.md: Update with new attributes.
8150 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8152 * config/aarch64/aarch64.md
8153 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8154 (*movtf_aarch64): Likewise.
8156 (thumb1_movdi_insn): Use "multiple" for type where more than one
8157 instruction is used for a move.
8158 (*arm32_movhf): Likewise.
8159 (*thumb_movdf_insn): Likewise.
8161 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8163 * config/arm/types.md (type): Rename fcpys to fmov.
8165 (*arm_movsi_vfp): Rename type fcpys as fmov.
8166 (*thumb2_movsi_vfp): Likewise
8167 (*movhf_vfp_neon): Likewise
8168 (*movhf_vfp): Likewise
8169 (*movsf_vfp): Likewise
8170 (*thumb2_movsf_vfp): Likewise
8171 (*movsfcc_vfp): Likewise
8172 (*thumb2_movsfcc_vfp): Likewise
8173 * config/aarch64/aarch64-simd.md
8174 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8175 * config/aarch64/aarch64.md
8176 (*movsi_aarch64): Replace type mov_reg with fmovs.
8177 (*movdi_aarch64): Likewise
8178 (*movsf_aarch64): Likewise
8179 (*movdf_aarch64): Likewise
8181 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
8182 * config/arm/iwmmxt.md
8183 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
8184 * config/arm/arm1020e.md: Update with new attributes.
8185 * config/arm/cortex-a15-neon.md: Update with new attributes.
8186 * config/arm/cortex-a5.md: Update with new attributes.
8187 * config/arm/cortex-a53.md: Update with new attributes.
8188 * config/arm/cortex-a7.md: Update with new attributes.
8189 * config/arm/cortex-a8-neon.md: Update with new attributes.
8190 * config/arm/cortex-a9.md: Update with new attributes.
8191 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8192 * config/arm/cortex-r4f.md: Update with new attributes.
8193 * config/arm/marvell-pj4.md: Update with new attributes.
8194 * config/arm/vfp11.md: Update with new attributes.
8196 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8198 * config/aarch64/aarch64.md
8199 (*madd<mode>): Fix type attribute.
8200 (*maddsi_uxtw): Likewise.
8201 (*msub<mode>): Likewise.
8202 (*msubsi_uxtw): Likewise.
8203 (<su_optab>maddsidi4): Likewise.
8204 (<su_optab>msubsidi4): Likewise.
8206 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8208 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
8209 * config/arm/arm.md (core_cycles): Remove fdiv.
8210 * config/arm/vfp.md:
8211 (*sqrtsf2_vfp): Update for attribute changes.
8212 (*sqrtdf2_vfp): Likewise.
8213 * config/aarch64/aarch64.md:
8214 (sqrt<mode>2): Update for attribute changes.
8215 * config/arm/arm1020e.md: Update with new attributes.
8216 * config/arm/cortex-a15-neon.md: Update with new attributes.
8217 * config/arm/cortex-a5.md: Update with new attributes.
8218 * config/arm/cortex-a53.md: Update with new attributes.
8219 * config/arm/cortex-a7.md: Update with new attributes.
8220 * config/arm/cortex-a8-neon.md: Update with new attributes.
8221 * config/arm/cortex-a9.md: Update with new attributes.
8222 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8223 * config/arm/cortex-r4f.md: Update with new attributes.
8224 * config/arm/marvell-pj4.md: Update with new attributes.
8225 * config/arm/vfp11.md: Update with new attributes.
8227 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8229 * config/arm/types.md
8230 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
8231 * config/aarch64/aarch64.md
8232 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
8234 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
8235 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
8236 (float<GPI:mode><GPF:mode>2): Likewise.
8238 (*truncsisf2_vfp): Update with new attributes.
8239 (*truncsidf2_vfp): Likewise.
8240 (fixuns_truncsfsi2): Likewise.
8241 (fixuns_truncdfsi2): Likewise.
8242 (*floatsisf2_vfp): Likewise.
8243 (*floatsidf2_vfp): Likewise.
8244 (floatunssisf2): Likewise.
8245 (floatunssidf2): Likewise.
8246 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
8247 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
8248 * config/arm/arm1020e.md: Update with new attributes.
8249 * config/arm/cortex-a15-neon.md: Update with new attributes.
8250 * config/arm/cortex-a5.md: Update with new attributes.
8251 * config/arm/cortex-a53.md: Update with new attributes.
8252 * config/arm/cortex-a7.md: Update with new attributes.
8253 * config/arm/cortex-a8-neon.md: Update with new attributes.
8254 * config/arm/cortex-a9.md: Update with new attributes.
8255 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8256 * config/arm/cortex-r4f.md: Update with new attributes.
8257 * config/arm/marvell-pj4.md: Update with new attributes.
8258 * config/arm/vfp11.md: Update with new attributes.
8260 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8262 * config/aarch64/arm_neon.h
8263 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
8264 (vqtbx<1,2,3,4><q>_s8): Likewise.
8266 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8268 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
8269 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
8270 (add<mode>3): Add type attribute.
8271 (add<mode>3): Likewise.
8272 (usadd<mode>3): Likewise.
8273 (ssadd<mode>3): Likewise.
8274 (sub<mode>3): Likewise.
8275 (sub<mode>3): Likewise.
8276 (ussub<mode>3): Likewise.
8277 (sssub<mode>3): Likewise.
8278 (ssmulsa3): Likewise.
8279 (usmulusa3): Likewise.
8280 (arm_usatsihi): Likewise.
8282 (*movdi_vfp): Add types for all instructions.
8283 (*movdi_vfp_cortexa8): Likewise.
8284 (*movhf_vfp_neon): Likewise.
8285 (*movhf_vfp): Likewise.
8286 (*movdf_vfp): Likewise.
8287 (*thumb2_movdf_vfp): Likewise.
8288 (*thumb2_movdfcc_vfp): Likewise.
8289 * config/arm/arm.md: Add type attribute to all insn patterns.
8290 (*thumb1_adddi3): Add type attribute.
8291 (*arm_adddi3): Likewise.
8292 (*adddi_sesidi_di): Likewise.
8293 (*adddi_zesidi_di): Likewise.
8294 (*thumb1_addsi3): Likewise.
8295 (addsi3_compare0): Likewise.
8296 (*addsi3_compare0_scratch): Likewise.
8297 (*compare_negsi_si): Likewise.
8298 (cmpsi2_addneg): Likewise.
8299 (*addsi3_carryin_<optab>): Likewise.
8300 (*addsi3_carryin_alt2_<optab>): Likewise.
8301 (*addsi3_carryin_clobercc_<optab>): Likewise.
8302 (*subsi3_carryin): Likewise.
8303 (*subsi3_carryin_const): Likewise.
8304 (*subsi3_carryin_compare): Likewise.
8305 (*subsi3_carryin_compare_const): Likewise.
8306 (*arm_subdi3): Likewise.
8307 (*thumb_subdi3): Likewise.
8308 (*subdi_di_zesidi): Likewise.
8309 (*subdi_di_sesidi): Likewise.
8310 (*subdi_zesidi_di): Likewise.
8311 (*subdi_sesidi_di): Likewise.
8312 (*subdi_zesidi_ze): Likewise.
8313 (thumb1_subsi3_insn): Likewise.
8314 (*arm_subsi3_insn): Likewise.
8315 (*anddi3_insn): Likewise.
8316 (*anddi_zesidi_di): Likewise.
8317 (*anddi_sesdi_di): Likewise.
8318 (*ne_zeroextracts): Likewise.
8319 (*ne_zeroextracts): Likewise.
8320 (*ite_ne_zeroextr): Likewise.
8321 (*ite_ne_zeroextr): Likewise.
8322 (*anddi_notdi_di): Likewise.
8323 (*anddi_notzesidi): Likewise.
8324 (*anddi_notsesidi): Likewise.
8325 (andsi_notsi_si): Likewise.
8326 (thumb1_bicsi3): Likewise.
8327 (*iordi3_insn): Likewise.
8328 (*iordi_zesidi_di): Likewise.
8329 (*iordi_sesidi_di): Likewise.
8330 (*thumb1_iorsi3_insn): Likewise.
8331 (*xordi3_insn): Likewise.
8332 (*xordi_zesidi_di): Likewise.
8333 (*xordi_sesidi_di): Likewise.
8334 (*arm_xorsi3): Likewise.
8335 (*andsi_iorsi3_no): Likewise.
8336 (*smax_0): Likewise.
8337 (*smax_m1): Likewise.
8338 (*arm_smax_insn): Likewise.
8339 (*smin_0): Likewise.
8340 (*arm_smin_insn): Likewise.
8341 (*arm_umaxsi3): Likewise.
8342 (*arm_uminsi3): Likewise.
8343 (*minmax_arithsi): Likewise.
8344 (*minmax_arithsi_): Likewise.
8345 (*satsi_<SAT:code>): Likewise.
8346 (arm_ashldi3_1bit): Likewise.
8347 (arm_ashrdi3_1bit): Likewise.
8348 (arm_lshrdi3_1bit): Likewise.
8349 (*arm_negdi2): Likewise.
8350 (*thumb1_negdi2): Likewise.
8351 (*arm_negsi2): Likewise.
8352 (*thumb1_negsi2): Likewise.
8353 (*negdi_extendsid): Likewise.
8354 (*negdi_zero_extend): Likewise.
8355 (*arm_abssi2): Likewise.
8356 (*thumb1_abssi2): Likewise.
8357 (*arm_neg_abssi2): Likewise.
8358 (*thumb1_neg_abss): Likewise.
8359 (one_cmpldi2): Likewise.
8360 (extend<mode>di2): Likewise.
8361 (*compareqi_eq0): Likewise.
8362 (*arm_extendhisi2addsi): Likewise.
8363 (*arm_movdi): Likewise.
8364 (*thumb1_movdi_insn): Likewise.
8365 (*arm_movt): Likewise.
8366 (*thumb1_movsi_insn): Likewise.
8367 (pic_add_dot_plus_four): Likewise.
8368 (pic_add_dot_plus_eight): Likewise.
8369 (tls_load_dot_plus_eight): Likewise.
8370 (*thumb1_movhi_insn): Likewise.
8371 (*thumb1_movsf_insn): Likewise.
8372 (*movdf_soft_insn): Likewise.
8373 (*thumb_movdf_insn): Likewise.
8374 (cbranchsi4_insn): Likewise.
8375 (cbranchsi4_scratch): Likewise.
8376 (*negated_cbranchsi4): Likewise.
8377 (*tbit_cbranch): Likewise.
8378 (*tlobits_cbranch): Likewise.
8379 (*tstsi3_cbranch): Likewise.
8380 (*cbranchne_decr1): Likewise.
8381 (*addsi3_cbranch): Likewise.
8382 (*addsi3_cbranch_scratch): Likewise.
8383 (*arm_cmpdi_insn): Likewise.
8384 (*arm_cmpdi_unsig): Likewise.
8385 (*arm_cmpdi_zero): Likewise.
8386 (*thumb_cmpdi_zero): Likewise.
8387 (*deleted_compare): Likewise.
8388 (*mov_scc): Likewise.
8389 (*mov_negscc): Likewise.
8390 (*mov_notscc): Likewise.
8391 (*cstoresi_eq0_thumb1_insn): Likewise.
8392 (cstoresi_nltu_thumb1): Likewise.
8393 (cstoresi_ltu_thu): Likewise.
8394 (thumb1_addsi3_addgeu): Likewise.
8395 (*arm_jump): Likewise.
8396 (*thumb_jump): Likewise.
8397 (*check_arch2): Likewise.
8398 (arm_casesi_internal): Likewise.
8399 (thumb1_casesi_dispatch): Likewise.
8400 (*arm_indirect_jump): Likewise.
8401 (*thumb1_indirect_jump): Likewise.
8403 (*and_scc): Likewise.
8404 (*ior_scc): Likewise.
8405 (*compare_scc): Likewise.
8406 (*cond_move): Likewise.
8407 (*cond_arith): Likewise.
8408 (*cond_sub): Likewise.
8409 (*cmp_ite0): Likewise.
8410 (*cmp_ite1): Likewise.
8411 (*cmp_and): Likewise.
8412 (*cmp_ior): Likewise.
8413 (*ior_scc_scc): Likewise.
8414 (*ior_scc_scc_cmp): Likewise.
8415 (*and_scc_scc): Likewise.
8416 (*and_scc_scc_cmp): Likewise.
8417 (*and_scc_scc_nod): Likewise.
8418 (*negscc): Likewise.
8419 (movcond_addsi): Likewise.
8420 (movcond): Likewise.
8421 (*ifcompare_plus_move): Likewise.
8422 (*if_plus_move): Likewise.
8423 (*ifcompare_move_plus): Likewise.
8424 (*if_move_plus): Likewise.
8425 (*ifcompare_arith_arith): Likewise.
8426 (*if_arith_arith): Likewise.
8427 (*ifcompare_arith_move): Likewise.
8428 (*if_arith_move): Likewise.
8429 (*ifcompare_move_arith): Likewise.
8430 (*if_move_arith): Likewise.
8431 (*ifcompare_move_not): Likewise.
8432 (*if_move_not): Likewise.
8433 (*ifcompare_not_move): Likewise.
8434 (*if_not_move): Likewise.
8435 (*ifcompare_shift_move): Likewise.
8436 (*if_shift_move): Likewise.
8437 (*ifcompare_move_shift): Likewise.
8438 (*if_move_shift): Likewise.
8439 (*ifcompare_shift_shift): Likewise.
8440 (*ifcompare_not_arith): Likewise.
8441 (*ifcompare_arith_not): Likewise.
8442 (*if_arith_not): Likewise.
8443 (*ifcompare_neg_move): Likewise.
8444 (*if_neg_move): Likewise.
8445 (*ifcompare_move_neg): Likewise.
8446 (*if_move_neg): Likewise.
8447 (prologue_thumb1_interwork): Likewise.
8448 (*cond_move_not): Likewise.
8449 (*sign_extract_onebit): Likewise.
8450 (*not_signextract_onebit): Likewise.
8451 (stack_tie): Likewise.
8452 (align_4): Likewise.
8453 (align_8): Likewise.
8454 (consttable_end): Likewise.
8455 (consttable_1): Likewise.
8456 (consttable_2): Likewise.
8457 (consttable_4): Likewise.
8458 (consttable_8): Likewise.
8459 (consttable_16): Likewise.
8460 (*thumb1_tablejump): Likewise.
8461 (prefetch): Likewise.
8462 (force_register_use): Likewise.
8463 (thumb_eh_return): Likewise.
8464 (load_tp_hard): Likewise.
8465 (load_tp_soft): Likewise.
8466 (tlscall): Likewise.
8467 (*arm_movtas_ze): Likewise.
8468 (*arm_rev): Likewise.
8469 (*arm_revsh): Likewise.
8470 (*arm_rev16): Likewise.
8471 * config/arm/thumb2.md
8472 (*thumb2_smaxsi3): Likewise.
8473 (*thumb2_sminsi3): Likewise.
8474 (*thumb32_umaxsi3): Likewise.
8475 (*thumb2_uminsi3): Likewise.
8476 (*thumb2_negdi2): Likewise.
8477 (*thumb2_abssi2): Likewise.
8478 (*thumb2_neg_abss): Likewise.
8479 (*thumb2_movsi_insn): Likewise.
8480 (tls_load_dot_plus_four): Likewise.
8481 (*thumb2_movhi_insn): Likewise.
8482 (*thumb2_mov_scc): Likewise.
8483 (*thumb2_mov_negs): Likewise.
8484 (*thumb2_mov_negs): Likewise.
8485 (*thumb2_mov_nots): Likewise.
8486 (*thumb2_mov_nots): Likewise.
8487 (*thumb2_movsicc_): Likewise.
8488 (*thumb2_movsfcc_soft_insn): Likewise.
8489 (*thumb2_indirect_jump): Likewise.
8490 (*thumb2_and_scc): Likewise.
8491 (*thumb2_ior_scc): Likewise.
8492 (*thumb2_ior_scc_strict_it): Likewise.
8493 (*thumb2_cond_move): Likewise.
8494 (*thumb2_cond_arith): Likewise.
8495 (*thumb2_cond_ari): Likewise.
8496 (*thumb2_cond_sub): Likewise.
8497 (*thumb2_negscc): Likewise.
8498 (*thumb2_movcond): Likewise.
8499 (thumb2_casesi_internal): Likewise.
8500 (thumb2_casesi_internal_pic): Likewise.
8501 (*thumb2_alusi3_short): Likewise.
8502 (*thumb2_mov<mode>_shortim): Likewise.
8503 (*thumb2_addsi_short): Likewise.
8504 (*thumb2_subsi_short): Likewise.
8505 (thumb2_addsi3_compare0): Likewise.
8506 (*thumb2_cbz): Likewise.
8507 (*thumb2_cbnz): Likewise.
8508 (*thumb2_one_cmplsi2_short): Likewise.
8509 (*thumb2_negsi2_short): Likewise.
8510 (*orsi_notsi_si): Likewise.
8511 * config/arm/arm1020e.md: Update with new attributes.
8512 * config/arm/arm1026ejs.md: Update with new attributes.
8513 * config/arm/arm1136jfs.md: Update with new attributes.
8514 * config/arm/arm926ejs.md: Update with new attributes.
8515 * config/arm/cortex-a15.md: Update with new attributes.
8516 * config/arm/cortex-a5.md: Update with new attributes.
8517 * config/arm/cortex-a53.md: Update with new attributes.
8518 * config/arm/cortex-a7.md: Update with new attributes.
8519 * config/arm/cortex-a8.md: Update with new attributes.
8520 * config/arm/cortex-a9.md: Update with new attributes.
8521 * config/arm/cortex-m4.md: Update with new attributes.
8522 * config/arm/cortex-r4.md: Update with new attributes.
8523 * config/arm/fa526.md: Update with new attributes.
8524 * config/arm/fa606te.md: Update with new attributes.
8525 * config/arm/fa626te.md: Update with new attributes.
8526 * config/arm/fa726te.md: Update with new attributes.
8528 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8530 * config/aarch64/aarch64-simd.md
8531 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
8532 <vwx> iterator to ensure correct register choice.
8533 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
8534 (aarch64_sqdmull_n<mode>): Likewise.
8535 (aarch64_sqdmull2_n<mode>_internal): Likewise.
8536 * config/aarch64/arm_neon.h
8537 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
8538 (vml<as><q>_n_<su>16): Likewise.
8539 (vml<as>l_high_lane<q>_<su>16): Likewise.
8540 (vml<as>l_high_n_<su>16): Likewise.
8541 (vml<as>l_lane<q>_<su>16): Likewise.
8542 (vml<as>l_n_<su>16): Likewise.
8543 (vmul<q>_lane<q>_<su>16): Likewise.
8544 (vmul<q>_n_<su>16): Likewise.
8545 (vmull_lane<q>_<su>16): Likewise.
8546 (vmull_n_<su>16): Likewise.
8547 (vmull_high_lane<q>_<su>16): Likewise.
8548 (vmull_high_n_<su>16): Likewise.
8549 (vqrdmulh<q>_n_s16): Likewise.
8551 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
8553 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
8554 have the correct lane parameter.
8556 2013-09-06 Richard Biener <rguenther@suse.de>
8558 * cfganal.c (control_dependences::~control_dependences):
8559 Properly free all of the vector.
8561 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
8564 * config/i386/i386.c (ix86_conditional_register_usage):
8565 Proper initialize extended SSE registers.
8567 2013-09-06 Jan Hubicka <jh@suse.cz>
8569 PR tree-optimization/58311
8570 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
8572 2013-09-06 Jan Hubicka <jh@suse.cz>
8574 * Makefile.in (tree-sra.o): Update dependencies.
8575 * tree-sra.c: Include ipa-utils.h
8576 (scan_function): Use recursive_call_p.
8577 (has_caller_p): New function.
8578 (cgraph_for_node_and_aliases): Count also callers of aliases.
8580 2013-09-06 Jan Hubicka <jh@suse.cz>
8583 * cgraph.h (symtab_semantically_equivalent_p): Declare.
8584 * tree-tailcall.c: Include ipa-utils.h.
8585 (find_tail_calls): Use it.
8586 * ipa-pure-const.c (check_call): Likewise.
8587 * ipa-utils.c (recursive_call_p): New function.
8588 * ipa-utils.h (recursive_call_p): Dclare.
8589 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
8590 (symtab_semantically_equivalent_p): New function.
8591 * Makefile.in (tree-tailcall.o): Update dependencies.
8593 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8595 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
8598 2013-09-06 Richard Biener <rguenther@suse.de>
8600 * lto-streamer.h (lto_global_var_decls): Remove.
8601 * Makefile.in (OBJS): Remove lto-symtab.o.
8602 (lto-symtab.o): Remove.
8603 (GTFILES): Remove lto-symtab.c
8604 * lto-symtab.c: Move to lto/
8606 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8608 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
8609 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
8610 (UNSPEC_FPINT_RINT): New constant definitions.
8611 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
8612 definition with 2 attributes.
8613 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
8614 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
8617 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8619 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
8620 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
8621 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
8622 attribute to "z196_alone".
8623 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
8626 2013-09-06 Richard Biener <rguenther@suse.de>
8628 * basic-block.h (class control_dependences): New.
8629 * tree-ssa-dce.c (control_dependence_map): Remove.
8631 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
8632 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
8633 find_pdom, find_control_dependence, find_all_control_dependences):
8635 (mark_control_dependent_edges_necessary,
8636 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
8637 tree_dce_done, perform_tree_ssa_dce): Adjust.
8638 * cfganal.c (set_control_dependence_map_bit,
8639 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
8640 find_all_control_dependences): Move from tree-ssa-dce.c and
8641 implement as methods of control_dependences class.
8642 (control_dependences::control_dependences): New.
8643 (control_dependences::~control_dependences): Likewise.
8644 (control_dependences::get_edges_dependent_on): Likewise.
8645 (control_dependences::get_edge): Likewise.
8647 2013-09-04 Jan Hubicka <jh@suse.cz>
8649 * tree.c (types_same_for_odr): Drop overactive check.
8650 * ipa-devirt.c (hash_type_name): Likewise.
8652 2013-09-04 Jan Hubicka <jh@suse.cz>
8654 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
8655 (analyze_functions): ... here.
8657 2013-09-04 Jan Hubicka <jh@suse.cz>
8660 * cgraphunit.c (analyze_functions): Clear AUX fields
8661 after processing; initialize assembler name has.
8663 2013-09-05 Jeff Law <law@redhat.com>
8665 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
8666 from thread_around_empty_block. Record threading path into PATH.
8667 Recurse if threading through the initial block is successful.
8668 (thread_across_edge): Corresponding changes to slightly simplify.
8670 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
8672 * config/aarch64/aarch64.md
8673 (type): Remove frecpe, frecps, frecpx.
8674 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
8675 fix to be a TARGET_SIMD instruction.
8676 (aarch64_frecps): Remove.
8677 * config/aarch64/aarch64-simd.md
8678 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
8679 (aarch64_frecps<mode>): Handle all float/vector of float modes.
8681 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
8682 Sofiane Naci <sofiane.naci@arm.com>
8684 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
8685 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
8686 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
8687 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
8688 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
8689 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
8690 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
8691 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
8692 into "clz, "rbit". Rename "shift" to "shift_imm".
8693 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
8694 changes. Update for attribute changes all occurrences of arlo_* and
8696 * config/arm/arm-fixed.md: Update for attribute changes
8697 all occurrences of arlo_* types.
8698 * config/arm/thumb2.md: Update for attribute changes all occurrences
8700 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
8701 (cortexa7_older_only): Likewise.
8702 (cortexa7_younger): Likewise.
8703 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8704 (1020alu_shift_op): Likewise.
8705 (1020alu_shift_reg_op): Likewise.
8706 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8707 (alu_shift_op): Likewise.
8708 (alu_shift_reg_op): Likewise.
8709 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8710 (11_alu_shift_op): Likewise.
8711 (11_alu_shift_reg_op): Likewise.
8712 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8713 (9_alu_shift_reg_op): Likewise.
8714 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
8716 (cortex_a15_alu_shift): Likewise.
8717 (cortex_a15_alu_shift_reg): Likewise.
8718 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
8720 (cortex_a5_alu_shift): Likewise.
8721 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
8723 (cortex_a53_alu_shift): Likewise.
8724 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
8726 (cortex_a7_alu_reg): Likewise.
8727 (cortex_a7_alu_shift): Likewise.
8728 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
8730 (cortex_a8_alu_shift): Likewise.
8731 (cortex_a8_alu_shift_reg): Likewise.
8732 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8733 (cortex_a9_dp_shift): Likewise.
8734 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
8736 * config/arm/cortex-r4.md
8737 (cortex_r4_alu): Update for attribute changes.
8738 (cortex_r4_mov): Likewise.
8739 (cortex_r4_alu_shift_reg): Likewise.
8740 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8741 (526_alu_shift_op): Likewise.
8742 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8743 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8744 (626te_alu_shift_op): Likewise.
8745 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
8746 (726te_alu_shift_op): Likewise.
8747 (726te_alu_shift_reg_op): Likewise.
8748 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8749 (mp626_alu_shift_op): Likewise.
8750 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
8751 (pj4_alu_conds): Likewise.
8752 (pj4_shift): Likewise.
8753 (pj4_shift_conds): Likewise.
8754 (pj4_alu_shift): Likewise.
8755 (pj4_alu_shift_conds): Likewise.
8756 * config/aarch64/aarch64.md: Update for attribute change
8757 all occurrences of arlo_* and shift* types.
8759 2013-09-05 Mike Stump <mikestump@comcast.net>
8761 * tree.h: Move documentation for tree_function_decl to tree-core.h
8762 with the declaration.
8764 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
8767 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
8768 looking for widest mode.
8770 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
8772 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
8774 2013-09-05 Richard Biener <rguenther@suse.de>
8776 PR tree-optimization/58137
8777 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
8778 Do not create vectors of pointers.
8779 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
8780 types for the components of the vector initializer.
8781 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
8782 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
8784 2013-09-05 Martin Jambor <mjambor@suse.cz>
8786 * ipa-prop.c (remove_described_reference): Accept missing references,
8787 return false if that hppens, otherwise return true.
8788 (cgraph_node_for_jfunc): New function.
8789 (try_decrement_rdesc_refcount): Likewise.
8790 (try_make_edge_direct_simple_call): Use them.
8791 (ipa_edge_removal_hook): Remove references from rdescs.
8792 (ipa_edge_duplication_hook): Clone rdescs and their references
8793 when the new edge has the same caller as the old one.
8794 * cgraph.c (cgraph_resolve_speculation): Remove speculative
8795 reference before removing any edges.
8797 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
8799 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
8801 * thumb2.md (thumb2_storewb_parisi): New pattern.
8803 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
8805 * config/aarch64/aarch64-option-extensions.def: Add
8806 AARCH64_OPT_EXTENSION of 'crc'.
8807 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
8808 (AARCH64_ISA_CRC): Ditto.
8809 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
8810 description of the CRC extension.
8812 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
8814 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
8815 * config/rs6000/linux.h: Ditto.
8816 * alpha/linux.h: Ditto.
8817 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
8818 no_c99_libc_has_function.
8819 * config/c6x/uclinux-elf.h: Ditto.
8820 * config/lm32/uclinux-elf.h: Ditto.
8821 * config/m68k/uclinux.h: Ditto.
8822 * config/moxie/uclinux.h: Ditto.
8823 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
8824 (crisv32-*-linux*, cris-*-linux*): Ditto.
8825 * config/bfin/bfin.c: Include "tm_p.h".
8827 2013-09-05 Richard Biener <rguenther@suse.de>
8829 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
8830 check for a definition without a basic-block.
8832 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
8833 Sofiane Naci <sofiane.naci@arm.com>
8835 * config/aarch64/aarch64.md
8836 (*movti_aarch64): Rename r_2_f and f_2_r.
8837 (*movsf_aarch64): Likewise.
8838 (*movdf_aarch64): Likewise.
8839 (*movtf_aarch64): Likewise.
8840 (aarch64_movdi_<mode>low): Likewise.
8841 (aarch64_movdi_<mode>high): Likewise.
8842 (aarch64_mov<mode>high_di): Likewise.
8843 (aarch64_mov<mode>low_di): Likewise.
8844 (aarch64_movtilow_tilow): Likewise.
8845 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
8846 values to config/arm/types.md
8847 (attribute "conds"): Update for attribute change.
8848 (anddi3_insn): Likewise.
8849 (iordi3_insn): Likewise.
8850 (xordi3_insn): Likewise.
8851 (one_cmpldi2): Likewise.
8852 * config/arm/types.md (type): Add Neon types.
8853 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
8854 use "type" attribute.
8855 (movmisalign<mode>_neon_store): Likewise.
8856 (movmisalign<mode>_neon_load): Likewise.
8857 (vec_set<mode>_internal): Likewise.
8858 (vec_setv2di_internal): Likewise.
8859 (vec_extract<mode>): Likewise.
8860 (vec_extractv2di): Likewise.
8861 (add<mode>3_neon): Likewise.
8862 (adddi3_neon): Likewise.
8863 (sub<mode>3_neon): Likewise.
8864 (subdi3_neon): Likewise.
8865 (mul<mode>3_neon): Likewise.
8866 (mul<mode>3add<mode>_neon): Likewise.
8867 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
8868 (fma<VCVTF:mode>4)): Likewise.
8869 (fma<VCVTF:mode>4_intrinsic): Likewise.
8870 (fmsub<VCVTF:mode>4)): Likewise.
8871 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
8872 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
8873 (ior<mode>3): Likewise.
8874 (and<mode>3): Likewise.
8875 (anddi3_neon): Likewise.
8876 (orn<mode>3_neon): Likewise.
8877 (orndi3_neon): Likewise.
8878 (bic<mode>3_neon): Likewise.
8879 (bicdi3_neon): Likewise.
8880 (xor<mode>3): Likewise.
8881 (one_cmpl<mode>2): Likewise.
8882 (abs<mode>2): Likewise.
8883 (neg<mode>2): Likewise.
8884 (umin<mode>3_neon): Likewise.
8885 (umax<mode>3_neon): Likewise.
8886 (smin<mode>3_neon): Likewise.
8887 (smax<mode>3_neon): Likewise.
8888 (vashl<mode>3): Likewise.
8889 (vashr<mode>3_imm): Likewise.
8890 (vlshr<mode>3_imm): Likewise.
8891 (ashl<mode>3_signed): Likewise.
8892 (ashl<mode>3_unsigned): Likewise.
8893 (neon_load_count): Likewise.
8894 (ashldi3_neon_noclobber): Likewise.
8895 (signed_shift_di3_neon): Likewise.
8896 (unsigned_shift_di3_neon): Likewise.
8897 (ashrdi3_neon_imm_noclobber): Likewise.
8898 (lshrdi3_neon_imm_noclobber): Likewise.
8899 (widen_ssum<mode>3): Likewise.
8900 (widen_usum<mode>3): Likewise.
8901 (quad_halves_<code>v4si): Likewise.
8902 (quad_halves_<code>v4sf): Likewise.
8903 (quad_halves_<code>v8hi): Likewise.
8904 (quad_halves_<code>v16qi): Likewise.
8905 (reduc_splus_v2di): Likewise.
8906 (neon_vpadd_internal<mode>): Likewise.
8907 (neon_vpsmin<mode>): Likewise.
8908 (neon_vpsmax<mode>): Likewise.
8909 (neon_vpumin<mode>): Likewise.
8910 (neon_vpumax<mode>): Likewise.
8911 (ss_add<mode>_neon): Likewise.
8912 (us_add<mode>_neon): Likewise.
8913 (ss_sub<mode>_neon): Likewise.
8914 (us_sub<mode>_neon): Likewise.
8915 (neon_vadd<mode>_unspec): Likewise.
8916 (neon_vaddl<mode>): Likewise.
8917 (neon_vaddw<mode>): Likewise.
8918 (neon_vhadd<mode>): Likewise.
8919 (neon_vqadd<mode>): Likewise.
8920 (neon_vaddhn<mode>): Likewise.
8921 (neon_vmul<mode>): Likewise.
8922 (neon_vmla<mode>): Likewise.
8923 (neon_vmlal<mode>): Likewise.
8924 (neon_vmls<mode>): Likewise.
8925 (neon_vmlsl<mode>): Likewise.
8926 (neon_vqdmulh<mode>): Likewise.
8927 (neon_vqdmlal<mode>): Likewise.
8928 (neon_vqdmlsl<mode>): Likewise.
8929 (neon_vmull<mode>): Likewise.
8930 (neon_vqdmull<mode>): Likewise.
8931 (neon_vsub<mode>_unspec): Likewise.
8932 (neon_vsubl<mode>): Likewise.
8933 (neon_vsubw<mode>): Likewise.
8934 (neon_vqsub<mode>): Likewise.
8935 (neon_vhsub<mode>): Likewise.
8936 (neon_vsubhn<mode>): Likewise.
8937 (neon_vceq<mode>): Likewise.
8938 (neon_vcge<mode>): Likewise.
8939 (neon_vcgeu<mode>): Likewise.
8940 (neon_vcgt<mode>): Likewise.
8941 (neon_vcgtu<mode>): Likewise.
8942 (neon_vcle<mode>): Likewise.
8943 (neon_vclt<mode>): Likewise.
8944 (neon_vcage<mode>): Likewise.
8945 (neon_vcagt<mode>): Likewise.
8946 (neon_vtst<mode>): Likewise.
8947 (neon_vabd<mode>): Likewise.
8948 (neon_vabdl<mode>): Likewise.
8949 (neon_vaba<mode>): Likewise.
8950 (neon_vabal<mode>): Likewise.
8951 (neon_vmax<mode>): Likewise.
8952 (neon_vmin<mode>): Likewise.
8953 (neon_vpaddl<mode>): Likewise.
8954 (neon_vpadal<mode>): Likewise.
8955 (neon_vpmax<mode>): Likewise.
8956 (neon_vpmin<mode>): Likewise.
8957 (neon_vrecps<mode>): Likewise.
8958 (neon_vrsqrts<mode>): Likewise.
8959 (neon_vqabs<mode>): Likewise.
8960 (neon_vqneg<mode>): Likewise.
8961 (neon_vcls<mode>): Likewise.
8962 (clz<mode>2): Likewise.
8963 (popcount<mode>2): Likewise.
8964 (neon_vrecpe): Likewise.
8965 (neon_vrsqrte): Likewise.
8966 (neon_vget_lane<mode>_sext_internal): Likewise.
8967 (neon_vget_lane<mode>_zext_internal): Likewise.
8968 (neon_vdup_n<mode>): Likewise.
8969 (neon_vdup_nv2di): Likewise.
8970 (neon_vdpu_lane<mode>_internal): Likewise.
8971 (neon_vswp<mode>): Likewise.
8972 (float<mode><V_cvtto>2): Likewise.
8973 (floatuns<mode><V_cvtto>2): Likewise.
8974 (fix_trunc<mode><V_cvtto>)2): Likewise
8975 (fixuns_trunc<mode><V_cvtto)2): Likewise.
8976 (neon_vcvt<mode>): Likewise.
8977 (neon_vcvtv4sfv4hf): Likewise.
8978 (neon_vcvtv4hfv4sf): Likewise.
8979 (neon_vcvt_n<mode>): Likewise.
8980 (neon_vmovn<mode>): Likewise.
8981 (neon_vqmovn<mode>): Likewise.
8982 (neon_vqmovun<mode>): Likewise.
8983 (neon_vmovl<mode>): Likewise.
8984 (neon_vmul_lane<mode>): Likewise.
8985 (neon_vmull_lane<mode>): Likewise.
8986 (neon_vqdmull_lane<mode>): Likewise.
8987 (neon_vqdmulh_lane<mode>): Likewise.
8988 (neon_vmla_lane<mode>): Likewise.
8989 (neon_vmlal_lane<mode>): Likewise.
8990 (neon_vqdmlal_lane<mode>): Likewise.
8991 (neon_vmls_lane<mode>): Likewise.
8992 (neon_vmlsl_lane<mode>): Likewise.
8993 (neon_vqdmlsl_lane<mode>): Likewise.
8994 (neon_vext<mode>): Likewise.
8995 (neon_vrev64<mode>): Likewise.
8996 (neon_vrev32<mode>): Likewise.
8997 (neon_vrev16<mode>): Likewise.
8998 (neon_vbsl<mode>_internal): Likewise.
8999 (neon_vshl<mode>): Likewise.
9000 (neon_vqshl<mode>): Likewise.
9001 (neon_vshr_n<mode>): Likewise.
9002 (neon_vshrn_n<mode>): Likewise.
9003 (neon_vqshrn_n<mode>): Likewise.
9004 (neon_vqshrun_n<mode>): Likewise.
9005 (neon_vshl_n<mode>): Likewise.
9006 (neon_vqshl_n<mode>): Likewise.
9007 (neon_vqshlu_n<mode>): Likewise.
9008 (neon_vshll_n<mode>): Likewise.
9009 (neon_vsra_n<mode>): Likewise.
9010 (neon_vsri_n<mode>): Likewise.
9011 (neon_vsli_n<mode>): Likewise.
9012 (neon_vtbl1v8qi): Likewise.
9013 (neon_vtbl2v8qi): Likewise.
9014 (neon_vtbl3v8qi): Likewise.
9015 (neon_vtbl4v8qi): Likewise.
9016 (neon_vtbx1v8qi): Likewise.
9017 (neon_vtbx2v8qi): Likewise.
9018 (neon_vtbx3v8qi): Likewise.
9019 (neon_vtbx4v8qi): Likewise.
9020 (neon_vtrn<mode>_internal): Likewise.
9021 (neon_vzip<mode>_internal): Likewise.
9022 (neon_vuzp<mode>_internal): Likewise.
9023 (neon_vld1<mode>): Likewise.
9024 (neon_vld1_lane<mode>): Likewise.
9025 (neon_vld1_dup<mode>): Likewise.
9026 (neon_vld1_dupv2di): Likewise.
9027 (neon_vst1<mode>): Likewise.
9028 (neon_vst1_lane<mode>): Likewise.
9029 (neon_vld2<mode>): Likewise.
9030 (neon_vld2_lane<mode>): Likewise.
9031 (neon_vld2_dup<mode>): Likewise.
9032 (neon_vst2<mode>): Likewise.
9033 (neon_vst2_lane<mode>): Likewise.
9034 (neon_vld3<mode>): Likewise.
9035 (neon_vld3qa<mode>): Likewise.
9036 (neon_vld3qb<mode>): Likewise.
9037 (neon_vld3_lane<mode>): Likewise.
9038 (neon_vld3_dup<mode>): Likewise.
9039 (neon_vst3<mode>): Likewise.
9040 (neon_vst3qa<mode>): Likewise.
9041 (neon_vst3qb<mode>): Likewise.
9042 (neon_vst3_lane<mode>): Likewise.
9043 (neon_vld4<mode>): Likewise.
9044 (neon_vld4qa<mode>): Likewise.
9045 (neon_vld4qb<mode>): Likewise.
9046 (neon_vld4_lane<mode>): Likewise.
9047 (neon_vld4_dup<mode>): Likewise.
9048 (neon_vst4<mode>): Likewise.
9049 (neon_vst4qa<mode>): Likewise.
9050 (neon_vst4qb<mode>): Likewise.
9051 (neon_vst4_lane<mode>): Likewise.
9052 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9053 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9054 (neon_vec_<US>mult_lo_<mode>): Likewise.
9055 (neon_vec_<US>mult_hi_<mode>): Likewise.
9056 (neon_vec_<US>shiftl_<mode>): Likewise.
9057 (neon_unpack<US>_<mode>): Likewise.
9058 (neon_vec_<US>mult_<mode>): Likewise.
9059 (vec_pack_trunc_<mode>): Likewise.
9060 (neon_vec_pack_trunk_<mode>): Likewise.
9061 (neon_vabd<mode>_2): Likewise.
9062 (neon_vabd<mode>_3): Likewise.
9063 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9064 (thumb2_movsi_vfp): Likewise.
9065 (movdi_vfp): Likewise.
9066 (movdi_vfp_cortexa8): Likewise.
9067 (movhf_vfp_neon): Likewise.
9068 (movhf_vfp): Likewiwse.
9069 (movsf_vfp): Likewiwse.
9070 (thumb2_movsf_vfp): Likewiwse.
9071 (movdf_vfp): Likewise.
9072 (thumb2_movdf_vfp): Likewise.
9073 (movsfcc_vfp): Likewise.
9074 (thumb2_movsfcc_vfp): Likewise.
9075 (movdfcc_vfp): Likewise.
9076 (thumb2_movdfcc_vfp): Likewise.
9077 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9078 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9079 (v10_v2c): Likewise.
9080 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9082 (cortex_a15_neon_int_2): Likewise.
9083 (cortex_a15_neon_int_3): Likewise.
9084 (cortex_a15_neon_int_4): Likewise.
9085 (cortex_a15_neon_int_5): Likewise.
9086 (cortex_a15_neon_vqneg_vqabs): Likewise.
9087 (cortex_a15_neon_vmov): Likewise.
9088 (cortex_a15_neon_vaba): Likewise.
9089 (cortex_a15_neon_vaba_qqq): Likewise.
9090 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9091 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9092 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9093 scalar_64_32_long_scalar): Likewise.
9094 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9095 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9096 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9097 lotype_qdd_64_32_long): Likewise.
9098 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9099 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9100 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9101 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9102 (cortex_a15_neon_shift_1): Likewise.
9103 (cortex_a15_neon_shift_2): Likewise.
9104 (cortex_a15_neon_shift_3): Likewise.
9105 (cortex_a15_neon_vshl_ddd): Likewise.
9106 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9107 (cortex_a15_neon_vsra_vrsra): Likewise.
9108 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9109 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9110 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9111 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9112 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9113 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9114 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9115 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9116 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9117 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9118 (cortex_a15_neon_bp_simple): Likewise.
9119 (cortex_a15_neon_bp_2cycle): Likewise.
9120 (cortex_a15_neon_bp_3cycle): Likewise.
9121 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9122 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9123 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9124 (cortex_a15_neon_vld2_4_regs): Likewise.
9125 (cortex_a15_neon_vld3_vld4): Likewise.
9126 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9127 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9128 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9129 (cortex_a15_neon_vst3_vst4): Likewise.
9130 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9131 (cortex_a15_neon_vld3_vld4_lane" 10
9132 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9133 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9134 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9135 (cortex_a15_neon_ldm_2): Likewise.0
9136 (cortex_a15_neon_stm_2): Likewise.
9137 (cortex_a15_neon_mcr): Likewise.
9138 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9139 (cortex_a15_neon_mrc): Likewise.
9140 (cortex_a15_neon_mrrc): Likewise.
9141 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9143 (cortex_a15_alu_shift): Likewise.
9144 (cortex_a15_alu_shift_reg): Likewise.
9145 (cortex_a15_mult32): Likewise.
9146 (cortex_a15_mult64): Likewise.
9147 (cortex_a15_block): Likewise.
9148 (cortex_a15_branch): Likewise.
9149 (cortex_a15_load1): Likewise.
9150 (cortex_a15_load3): Likewise.
9151 (cortex_a15_store1): Likewise.
9152 (cortex_a15_store3): Likewise.
9153 (cortex_a15_call): Likewise.
9154 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9155 (cortex_a5_f2r): Likewise.
9156 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9158 (cortex_a53_f2r): Likewise.
9159 * config/arm/cortex-a7.md
9160 (cortex_a7_branch): Update for attribute change.
9161 (cortex_a7_call): Likewise.
9162 (cortex_a7_alu_imm): Likewise.
9163 (cortex_a7_alu_reg): Likewise.
9164 (cortex_a7_alu_shift): Likewise.
9165 (cortex_a7_mul): Likewise.
9166 (cortex_a7_load1): Likewise.
9167 (cortex_a7_store1): Likewise.
9168 (cortex_a7_load2): Likewise.
9169 (cortex_a7_store2): Likewise.
9170 (cortex_a7_load3): Likewise.
9171 (cortex_a7_store3): Likewise.
9172 (cortex_a7_load4): Likewise.
9173 (cortex_a7_store4): Likewise.
9174 (cortex_a7_fpalu): Likewise.
9175 (cortex_a7_fconst): Likewise.
9176 (cortex_a7_fpmuls): Likewise.
9177 (cortex_a7_neon_mul): Likewise.
9178 (cortex_a7_fpmacs): Likewise.
9179 (cortex_a7_neon_mla: Likewise.
9180 (cortex_a7_fpmuld: Likewise.
9181 (cortex_a7_fpmacd: Likewise.
9182 (cortex_a7_fpfmad: Likewise.
9183 (cortex_a7_fdivs: Likewise.
9184 (cortex_a7_fdivd: Likewise.
9185 (cortex_a7_r2f: Likewise.
9186 (cortex_a7_f2r: Likewise.
9187 (cortex_a7_f_flags: Likewise.
9188 (cortex_a7_f_loads: Likewise.
9189 (cortex_a7_f_loadd: Likewise.
9190 (cortex_a7_f_stores: Likewise.
9191 (cortex_a7_f_stored: Likewise.
9192 (cortex_a7_neon): Likewise.
9193 * config/arm/cortex-a8-neon.md
9194 (cortex_a8_neon_mrc): Update for attribute change.
9195 (cortex_a8_neon_mrrc): Likewise.
9196 (cortex_a8_neon_int_1): Likewise.
9197 (cortex_a8_neon_int_2): Likewise.
9198 (cortex_a8_neon_int_3): Likewise.
9199 (cortex_a8_neon_int_4): Likewise.
9200 (cortex_a8_neon_int_5): Likewise.
9201 (cortex_a8_neon_vqneg_vqabs): Likewise.
9202 (cortex_a8_neon_vmov): Likewise.
9203 (cortex_a8_neon_vaba): Likewise.
9204 (cortex_a8_neon_vaba_qqq): Likewise.
9205 (cortex_a8_neon_vsma): Likewise.
9206 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9207 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9208 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9209 long_scalar): Likewise.
9210 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9211 (cortex_a8_neon_mla_qqq_8_16): Likewise.
9212 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9213 long_scalar_qdd_64_32_long): Likewise.
9214 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9215 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9216 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
9217 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9218 (cortex_a8_neon_shift_1): Likewise.
9219 (cortex_a8_neon_shift_2): Likewise.
9220 (cortex_a8_neon_shift_3): Likewise.
9221 (cortex_a8_neon_vshl_ddd): Likewise.
9222 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9223 (cortex_a8_neon_vsra_vrsra): Likewise.
9224 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
9225 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
9226 (cortex_a8_neon_fp_vsum): Likewise.
9227 (cortex_a8_neon_fp_vmul_ddd): Likewise.
9228 (cortex_a8_neon_fp_vmul_qqd): Likewise.
9229 (cortex_a8_neon_fp_vmla_ddd): Likewise.
9230 (cortex_a8_neon_fp_vmla_qqq): Likewise.
9231 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
9232 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
9233 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9234 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9235 (cortex_a8_neon_bp_simple): Likewise.
9236 (cortex_a8_neon_bp_2cycle): Likewise.
9237 (cortex_a8_neon_bp_3cycle): Likewise.
9238 (cortex_a8_neon_ldr): Likewise.
9239 (cortex_a8_neon_str): Likewise.
9240 (cortex_a8_neon_vld1_1_2_regs): Likewise.
9241 (cortex_a8_neon_vld1_3_4_regs): Likewise.
9242 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9243 (cortex_a8_neon_vld2_4_regs): Likewise.
9244 (cortex_a8_neon_vld3_vld4): Likewise.
9245 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9246 (cortex_a8_neon_vst1_3_4_regs): Likewise.
9247 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
9248 (cortex_a8_neon_vst3_vst4): Likewise.
9249 (cortex_a8_neon_vld1_vld2_lane): Likewise.
9250 (cortex_a8_neon_vld3_vld4_lane): Likewise.
9251 (cortex_a8_neon_vst1_vst2_lane): Likewise.
9252 (cortex_a8_neon_vst3_vst4_lane): Likewise.
9253 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
9254 (cortex_a8_neon_mcr): Likewise.
9255 (cortex_a8_neon_mcr_2_mcrr): Likewise.
9256 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9257 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
9259 (ca9_neon_mrrc): Likewise.
9260 (cortex_a9_neon_int_1): Likewise.
9261 (cortex_a9_neon_int_2): Likewise.
9262 (cortex_a9_neon_int_3): Likewise.
9263 (cortex_a9_neon_int_4): Likewise.
9264 (cortex_a9_neon_int_5): Likewise.
9265 (cortex_a9_neon_vqneg_vqabs): Likewise.
9266 (cortex_a9_neon_vmov): Likewise.
9267 (cortex_a9_neon_vaba): Likewise.
9268 (cortex_a9_neon_vaba_qqq): Likewise.
9269 (cortex_a9_neon_vsma): Likewise.
9270 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9271 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9272 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9273 long_scalar): Likewise.
9274 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9275 (cortex_a9_neon_mla_qqq_8_16): Likewise.
9276 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9277 long_scalar_qdd_64_32_long): Likewise.
9278 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9279 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9280 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
9281 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9282 (cortex_a9_neon_shift_1): Likewise.
9283 (cortex_a9_neon_shift_2): Likewise.
9284 (cortex_a9_neon_shift_3): Likewise.
9285 (cortex_a9_neon_vshl_ddd): Likewise.
9286 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9287 (cortex_a9_neon_vsra_vrsra): Likewise.
9288 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
9289 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
9290 (cortex_a9_neon_fp_vsum): Likewise.
9291 (cortex_a9_neon_fp_vmul_ddd): Likewise.
9292 (cortex_a9_neon_fp_vmul_qqd): Likewise.
9293 (cortex_a9_neon_fp_vmla_ddd): Likewise.
9294 (cortex_a9_neon_fp_vmla_qqq): Likewise.
9295 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
9296 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
9297 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9298 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9299 (cortex_a9_neon_bp_simple): Likewise.
9300 (cortex_a9_neon_bp_2cycle): Likewise.
9301 (cortex_a9_neon_bp_3cycle): Likewise.
9302 (cortex_a9_neon_ldr): Likewise.
9303 (cortex_a9_neon_str): Likewise.
9304 (cortex_a9_neon_vld1_1_2_regs): Likewise.
9305 (cortex_a9_neon_vld1_3_4_regs): Likewise.
9306 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9307 (cortex_a9_neon_vld2_4_regs): Likewise.
9308 (cortex_a9_neon_vld3_vld4): Likewise.
9309 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9310 (cortex_a9_neon_vst1_3_4_regs): Likewise.
9311 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
9312 (cortex_a9_neon_vst3_vst4): Likewise.
9313 (cortex_a9_neon_vld1_vld2_lane): Likewise.
9314 (cortex_a9_neon_vld3_vld4_lane): Likewise.
9315 (cortex_a9_neon_vst1_vst2_lane): Likewise.
9316 (cortex_a9_neon_vst3_vst4_lane): Likewise.
9317 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
9318 (cortex_a9_neon_mcr): Likewise.
9319 (cortex_a9_neon_mcr_2_mcrr): Likewise.
9320 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9321 (cortex_a9_fps): Likewise.
9322 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
9324 (cortex_m4_fmuls): Likewise.
9325 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
9327 (cortex_r4_mrc): Likewise.
9328 * config/arm/iterators.md: Update comment referring to neon_type.
9329 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
9330 (iwmmxt_movsi_insn): Likewise.
9331 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
9333 (pj4_core_to_vfp): Likewise.
9334 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
9336 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
9337 (vfp_fstore): Likewise.
9338 * doc/md.texi: Change references to neon_type to refer to type.
9340 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9342 * tree.h (DECL_BUILT_IN): Fix typo in comment.
9344 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
9346 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
9349 2013-09-04 Easwaran Raman <eraman@google.com>
9352 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
9353 (build_and_add_sum): Use it.
9354 (appears_later_in_bb): Simplify code.
9356 2013-09-04 Teresa Johnson <tejohnson@google.com>
9358 * dumpfile.c (dump_finish): Don't close stderr/stdout.
9360 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
9362 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
9364 2013-09-04 Jan Hubicka <jh@suse.cz>
9366 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
9367 * ipa-devirt.c: Include diganostic.h
9368 (odr_type_d): Add types and types_set.
9369 (hash_type_name): Work for types with vtables during LTO.
9370 (odr_hasher::remove): Fix comment; destroy types_set.
9371 (add_type_duplicate): New function,
9372 (get_odr_type): Use it.
9373 (dump_type_inheritance_graph): Dump type duplicates.
9374 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
9376 * tree.c (types_same_for_odr): Give exact answers on types with
9379 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9381 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
9382 explaining their differences.
9384 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
9386 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
9388 2013-09-03 Jeff Law <law@redhat.com>
9390 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
9391 when not threading through a joiner block. Pass joiner/no joiner
9392 state to register_jump_thread.
9393 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
9394 state from argument rather than implying on path length.
9395 Dump the entire jump thread path into debugging dump.
9396 * tree-flow.h (register_jump_thread): Update prototype.
9398 2013-08-29 Xinliang David Li <davidxl@google.com>
9400 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
9401 Remove a trivial gcc_assert.
9403 2013-08-29 Xinliang David Li <davidxl@google.com>
9405 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
9406 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
9407 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
9408 Delay base decl alignment adjustment.
9409 * tree-vectorizer.c (vect_destroy_datarefs): New function.
9410 * tree-vectorizer.h: New data structure.
9411 (set_dr_misalignment): New function.
9412 (dr_misalignment): Ditto.
9413 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
9414 (vectorizable_load): Ditto.
9415 (ensure_base_align): New function.
9416 (vectorize_loops): Add dbg_cnt support.
9417 (execute_vect_slp): Ditto.
9418 * dbgcnt.def: New debug counter.
9419 * Makefile: New dependency.
9421 2013-09-03 Meador Inge <meadori@codesourcery.com>
9425 2013-08-30 Meador Inge <meadori@codesourcery.com>
9427 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
9429 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
9431 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
9432 function descriptor.
9434 2013-09-03 Richard Biener <rguenther@suse.de>
9436 * tree-affine.c (add_elt_to_tree): Fix association issue,
9437 avoid useless converts and make sure to always return a
9438 properly typed result.
9440 2013-09-03 Richard Biener <rguenther@suse.de>
9443 * fold-const.c (negate_expr_p): Fix division case.
9444 (negate_expr): Likewise.
9446 2013-09-03 Richard Biener <rguenther@suse.de>
9449 * tree-streamer-out.c: Include tm.h.
9450 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
9452 2013-09-03 Jan Hubicka <jh@suse.cz>
9454 * tree-profile.c (tree_profiling): Cleanup CFG when done.
9456 2013-09-03 Alan Modra <amodra@gmail.com>
9458 * config.gcc (powerpc*-*-linux*): Add support for little-endian
9459 multilibs to big-endian target and vice versa.
9460 * config/rs6000/t-linux64: Use := assignment on all vars.
9461 (MULTILIB_EXTRA_OPTS): Remove fPIC.
9462 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
9463 * config/rs6000/t-linux64le: New file.
9464 * config/rs6000/t-linux64bele: New file.
9465 * config/rs6000/t-linux64lebe: New file.
9467 2013-09-02 Jan Hubicka <jh@suse.cz>
9469 * ipa-inline-transform.c (inline_transform): Do not
9470 optimize_inline_calls when not optimizing.
9472 2013-09-02 Jan Hubicka <jh@suse.cz>
9474 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
9475 duplicated nodes for assembler names.
9476 * symtab.c (symtab_unregister_node): Do not attempt to unlink
9477 hard registers from assembler name hash.
9479 2013-09-02 Jan Hubicka <jh@suse.cz>
9481 * ipa-split.c (execute_split_functions): Split externally visible
9482 functions called once.
9484 2013-09-02 Martin Jambor <mjambor@suse.cz>
9487 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
9488 linked list. When finding the correct duplicate, also consider also
9489 the caller in additon to its inlined_to node.
9491 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
9493 * config/aarch64/aarch64-simd-builtins.def
9494 (dup_lane_scalar): Remove.
9495 * config/aarch64/aarch64-simd.md
9496 (aarch64_simd_dup): Add 'w->w' alternative.
9497 (aarch64_dup_lane<mode>): Allow for VALL.
9498 (aarch64_dup_lane_scalar<mode>): Remove.
9499 (aarch64_dup_lane_<vswap_width_name><mode>): New.
9500 (aarch64_get_lane_signed<mode>): Add w->w altenative.
9501 (aarch64_get_lane_unsigned<mode>): Likewise.
9502 (aarch64_get_lane<mode>): Likewise.
9503 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
9504 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
9505 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
9506 (VCON): Change container of V2SF.
9507 (vswap_width_name): Likewise.
9508 * config/aarch64/arm_neon.h
9509 (__aarch64_vdup_lane_any): New.
9510 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
9511 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
9512 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
9514 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
9517 * expr.c (emit_move_complex): Do not move complex FP values as parts if
9518 the source or the destination is a single hard register.
9520 2013-09-02 Richard Biener <rguenther@suse.de>
9523 * tree-scalar-evolution.c (instantiate_scev_name): Allow
9526 2013-09-02 Richard Biener <rguenther@suse.de>
9528 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
9529 arithmetic to sizetype.
9531 2013-09-02 Bin Cheng <bin.cheng@arm.com>
9533 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
9534 Find auto-increment use both before and after candidate.
9536 2013-09-02 Marek Polacek <polacek@redhat.com>
9538 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
9540 2013-09-01 Jan Hubicka <jh@suse.cz>
9542 * Makefile.in: Add ipa-profile.o
9543 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
9544 * cgraph.c (struct cgraph_propagate_frequency_data,
9545 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
9546 ipa-profile.c; replace cgraph_ by ipa_ prefix.
9547 * cgraph.h (cgraph_propagate_frequency): Remove.
9548 * ipa-inline-analysis.c: Include ipa-utils.h;
9549 drop duplicated cfgloop.h.
9550 (inline_update_callee_summaries): Update.
9551 * ipa-profile.c: New file.
9552 * ipa-utils.h (ipa_propagate_frequency): Declare.
9553 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
9554 data-streamer.h, value-prof.h.
9555 (symtab_remove_unreachable_nodes): Update profile.
9556 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
9557 account_time_size, cmp_counts, dump_histogram,
9558 ipa_profile_generate_summary, ipa_profile_write_summary,
9559 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
9560 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
9561 Move to ipa-profile.c.
9563 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
9565 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
9567 2013-09-01 Jan Hubicka <jh@suse.cz>
9569 * common.opt (fdevirtualize-speculatively): New function.
9570 * invoke.texi (fdevirtualize-speculatively): Document.
9571 * ipa-devirt.c: Include ipa-inline.h
9572 (likely_target_p): New function.
9573 (ipa_devirt): New function.
9574 (gate_ipa_devirt): New function.
9575 (pass_data_ipa_devirt): New static var.
9576 (pass_ipa_devirt): Likewise.
9577 (make_pass_ipa_devirt): New function.
9578 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
9579 (common_handle_option): Disable devirtualization when
9580 value range profiling is available.
9581 * passes.def (pass_ipa_devirt): Add.
9582 * timever.def (TV_IPA_DEVIRT): New timevar.
9583 * tree-pass.h (make_pass_ipa_devirt):
9585 2013-09-01 Iain Sandoe <iain@codesourcery.com>
9587 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
9588 include sanitize(undefined).
9590 2013-08-31 Diego Novillo <dnovillo@google.com>
9592 * Makefile.in (TREE_CORE_H): Define.
9594 (GTFILES): Add tree-core.h.
9595 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
9597 * tree-core.h: New file.
9598 Move all data structures, enum, typedefs, global
9599 declarations and constants from ...
9602 2013-08-31 Jan Hubicka <jh@suse.cz>
9604 * bulitins.c (expand_builtin): Do not early exit for gcov
9605 instrumented functions.
9607 2013-08-31 Marek Polacek <polacek@redhat.com>
9609 * ubsan.c: Include tm_p.h.
9611 2013-08-31 Jan Hubicka <jh@suse.cz>
9613 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
9616 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
9617 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
9618 * tree-cfg.c (verify_gimple_label): ... here.
9619 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
9620 (ipa_merge_profiles): New function.
9621 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
9622 (lto_input_function_body): Likewise.
9623 * ipa-utils.h (ipa_merge_profiles): Declare.
9624 * lto-streamer.h (lto_input_function_body): Update prototype.
9625 (emit_label_in_global_context_p): Remove.
9626 * lto-symtab.c: Include ipa-utils.h
9627 (lto_cgraph_replace_node): Use ipa_merge_profiles.
9629 2013-08-31 Jan Hubicka <jh@suse.cz>
9631 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
9633 2013-08-31 Jan Hubicka <jh@suse.cz>
9635 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
9637 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
9639 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
9640 "cmp" RTX before signed_comparison_operator check to account
9643 2013-08-30 Jan Hubicka <jh@suse.cz>
9645 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
9646 (detect_type_change_1): Rename to ...
9647 (detect_type_change): ... this one; early return on non-polymorphic
9649 (detect_type_change_ssa): Add comp_type parameter; update
9650 use of detect_type_change.
9651 (compute_complex_assign_jump_func): Add param_type parameter;
9652 update use of detect_type_change_ssa.
9653 (compute_complex_ancestor_jump_func): Likewise.
9654 (ipa_get_callee_param_type): New function.
9655 (ipa_compute_jump_functions_for_edge): Compute parameter type;
9656 update calls to the jump function computation functions.
9658 2013-08-30 Teresa Johnson <tejohnson@google.com>
9659 Steven Bosscher <steven@gcc.gnu.org>
9661 * cfgrtl.c (fixup_new_cold_bb): New routine.
9662 (commit_edge_insertions): Invoke fixup_partitions.
9663 (find_partition_fixes): New routine.
9664 (fixup_partitions): Ditto.
9665 (verify_hot_cold_block_grouping): Update comments.
9666 (rtl_verify_edges): Invoke find_partition_fixes.
9667 (rtl_verify_bb_pointers): Update comments.
9668 (rtl_verify_bb_layout): Ditto.
9669 * basic-block.h (probably_never_executed_edge_p): Declare.
9670 (fixup_partitions): Ditto.
9671 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
9672 * bb-reorder.c (sanitize_hot_paths): New function.
9673 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
9675 * predict.c (probably_never_executed_edge_p): New routine.
9676 * cfg.c (check_bb_profile): Add partition insanity warnings.
9678 2013-08-30 Meador Inge <meadori@codesourcery.com>
9680 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
9682 2013-08-30 Marek Polacek <polacek@redhat.com>
9684 * Makefile.in (ubsan.o): Add.
9685 (c-family/c-ubsan.o): Add.
9686 (builtins.o): Add ubsan.h dependency.
9687 * ubsan.h: New file.
9688 * ubsan.c: New file.
9689 * common.opt: Add -fsanitize=undefined option.
9690 (flag_sanitize): Add variable.
9691 (fsanitize=): Add option. Add Driver.
9692 (fsanitize=thread): Remove option.
9693 (fsanitize=address): Likewise.
9694 (static-libubsan): New option.
9695 * doc/invoke.texi: Document the new flag and -static-libubsan.
9696 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
9697 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
9698 * builtin-attrs.def (ATTR_COLD): Define.
9699 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
9700 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
9701 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
9702 * flag-types.h (sanitize_code): New enum.
9703 * opts.c (common_handle_option): Parse command line arguments
9704 of -fsanitize=. Add -fsanitize=unreachable option.
9705 * varasm.c (get_variable_section): Adjust.
9706 (assemble_noswitch_variable): Likewise.
9707 (assemble_variable): Likewise.
9708 (output_constant_def_contents): Likewise.
9709 (categorize_decl_for_section): Likewise.
9710 (place_block_symbol): Likewise.
9711 (output_object_block): Likewise.
9712 * builtins.def: Likewise.
9713 * toplev.c (compile_file): Likewise.
9714 (process_options): Likewise.
9715 * cppbuiltin.c: Likewise.
9716 * tsan.c (tsan_pass): Likewise.
9717 (tsan_gate): Likewise.
9718 (tsan_gate_O0): Likewise.
9719 * cfgexpand.c (partition_stack_vars): Likewise.
9720 (expand_stack_vars): Likewise.
9721 (defer_stack_allocation): Likewise.
9722 (expand_used_vars): Likewise.
9723 * cfgcleanup.c (old_insns_match_p): Likewise.
9724 * asan.c (asan_finish_file): Likewise.
9725 (asan_instrument): Likewise.
9726 (gate_asan): Likewise.
9727 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
9728 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
9729 (asan_global_struct): Use pointer_sized_int_node instead
9730 calling build_nonstandard_integer_type.
9731 (initialize_sanitizer_builtins): Likewise.
9732 (asan_finish_file): Likewise.
9733 * gcc.c: Document %{%:function(args):X}.
9734 (static_spec_functions): Add sanitize.
9735 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
9736 store funcval != NULL there.
9737 (do_spec_1): Adjust handle_spec_function caller.
9738 (handle_braces): Allow %:function(args) as condition.
9739 (sanitize_spec_function): New function.
9740 (ADD_STATIC_LIBUBSAN_LIBS): Define.
9741 (LIBUBSAN_SPEC): Likewise.
9742 (LIBUBSAN_EARLY_SPEC): Likewise.
9743 (SANITIZER_SPEC): Handle libubsan.
9744 (SANITIZER_EARLY_SPEC): Likewise.
9745 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
9746 instead of fsanitize=address.
9747 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
9748 instead of fsanitize=address*.
9749 * builtins.c: Include ubsan.h.
9750 (fold_builtin_0): Instrument __builtin_unreachable.
9751 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
9752 instead of flag_asan.
9753 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
9754 (pointer_sized_int_node): Define.
9755 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
9757 2013-08-30 Mike Stump <mikestump@comcast.net>
9759 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
9762 2013-08-29 Jan Hubicka <jh@suse.cz>
9764 * cgraph.c (cgraph_function_body_availability): Handle weakref
9766 * passes.def: Remove pass_fixup_cfg.
9767 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
9768 track when we need to remove functions.
9769 (gate_ipa_inline): Execute inlining always; add comment why.
9770 (pass_data_ipa_inline): Remove TODO_remove_functions.
9771 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
9772 do not produce summaries.
9773 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
9774 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
9775 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
9776 constant pool and vtable.
9778 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
9780 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
9781 New arm_neon.h's internal macros to specify 64-bit constants.
9782 Avoid using stdint.h's macros.
9784 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
9786 * recog.c (verify_changes): Verify that changes[i].old is non-zero
9787 before applying REG_P.
9789 2013-08-30 Jakub Jelinek <jakub@redhat.com>
9791 PR tree-optimization/58277
9792 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
9793 after seeing too many stmts with vdef in between dombb and current
9794 bb, invalidate everything.
9796 2013-08-30 Richard Biener <rguenther@suse.de>
9798 * fold-const.c (fold_single_bit_test): Fix overflow test.
9800 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
9802 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
9803 and which can live in a register, always retrieve the value on entry.
9804 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
9805 passed by invisible reference specially.
9806 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
9807 (vt_add_function_parameter): Correctly deal with a parameter passed by
9808 invisible reference.
9810 2013-08-30 Jan Hubicka <jh@suse.cz>
9812 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
9814 2013-08-30 Richard Biener <rguenther@suse.de>
9816 PR tree-optimization/58228
9817 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
9818 allow invariant loads in nested loop vectorization.
9820 2013-08-30 Richard Biener <rguenther@suse.de>
9822 PR tree-optimization/58223
9823 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
9824 (has_anti_or_output_dependence): ... this and adjust to also
9825 look for output dependences.
9826 (mark_nodes_having_upstream_mem_writes): Adjust.
9827 (rdg_flag_uses): Likewise.
9829 2013-08-30 Richard Biener <rguenther@suse.de>
9831 PR tree-optimization/58010
9832 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
9833 assert that we have a loop-closed PHI.
9835 2013-08-29 Jan Hubicka <jh@suse.cz>
9837 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
9838 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
9839 * lto-section-in.c (lto_free_function_in_decl_state): New function.
9840 (lto_free_function_in_decl_state_for_node): New function.
9842 2013-08-29 Xinliang David Li <davidxl@google.com>
9844 * loop-unroll.c (report_unroll_peel): Minor message change.
9845 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
9846 Emit alignment peeling message with default -fopt-info.
9847 (vect_loop_versioning): Emit loop version info message.
9848 * tree-vectorizer.c (vectorize_loops): Minor message change.
9849 (execute_vect_slp): Ditto.
9851 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
9853 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
9854 of the clone from the DECL_NAME of the original function.
9856 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
9858 * passes.c (register_pass): Add overload.
9859 * tree-pass.h (register_pass): Forward declare it. Add comment.
9861 2013-08-29 Jan Hubicka <jh@suse.cz>
9863 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
9864 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
9865 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
9866 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
9867 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
9868 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
9869 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
9870 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
9872 2013-08-29 Teresa Johnson <tejohnson@google.com>
9874 * dumpfile.c (dump_loc): Output column number.
9875 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
9876 * doc/invoke.texi: Document optall -fopt-info flag.
9877 * profile.c (read_profile_edge_counts): Use new dump framework.
9878 (compute_branch_probabilities): Ditto.
9879 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
9880 when pass not in any opt group.
9881 * pass_manager.h (pass_manager::get_pass_profile): New method.
9882 * value-prof.c (check_counter): Use new dump framework.
9883 (check_ic_target): Ditto.
9884 * coverage.c (get_coverage_counts): Ditto.
9885 (coverage_init): Setup new dump framework.
9887 2013-08-29 Richard Biener <rguenther@suse.de>
9889 PR tree-optimization/58246
9890 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
9891 handle the dominance check inside a basic-block.
9893 2013-08-29 Richard Biener <rguenther@suse.de>
9896 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
9897 of default defs that appear in abnormal PHI nodes.
9899 2013-08-29 Richard Biener <rguenther@suse.de>
9901 PR tree-optimization/57685
9902 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
9903 single-use operands to avoid exponential complexity.
9905 2013-08-28 Dehao Chen <dehao@google.com>
9907 * ipa-inline.c (edge_badness): Fix integer underflow.
9909 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
9911 * gtm-builtins.def (_ITM_free): Declare leaf.
9913 2013-08-28 Jakub Jelinek <jakub@redhat.com>
9916 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
9917 (*tls_local_dynamic_base_64_largepic): Likewise.
9918 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
9919 Remove predicate from call operand.
9920 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
9921 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
9923 2013-08-28 Jeff Law <law@redhat.com>
9925 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
9926 checks for the number of predecessors and successors allowed.
9927 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
9928 which require copying a joiner block if there is a request which
9929 is a subpath that requires no joiner block copying.
9931 2013-08-28 Jan Hubicka <jh@suse.cz>
9933 * lto-streamer-out.c (DFS_write_tree_body): Drop
9934 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
9935 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
9936 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
9937 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
9938 Do not read DECL_ERROR_ISSUED.
9939 (unpack_ts_decl_with_vis_value_fields): Do not read
9941 (lto_input_ts_binfo_tree_pointers): Do not read
9942 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
9943 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
9944 write DECL_ERROR_ISSUED..
9945 (pack_ts_decl_with_vis_value_fields): Do not write
9947 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
9948 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
9949 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
9950 * tree.h (tree_decl_common): Update comment.
9951 (DECL_ERROR_ISSUED): Remove.
9953 2013-08-28 Jakub Jelinek <jakub@redhat.com>
9956 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
9958 2013-08-28 Jan Hubicka <jh@suse.cz>
9960 * builtins.def (free): Declare leaf.
9962 2013-08-27 David Malcolm <dmalcolm@redhat.com>
9965 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
9966 * configure: Regenerate.
9968 2013-08-27 Martin Jambor <mjambor@suse.cz>
9970 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
9971 (ipa_ancestor_jf_data): Likewise.
9972 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
9973 (ipa_get_jf_pass_through_type_preserved): New function.
9974 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
9975 (ipa_get_jf_ancestor_type_preserved): New function.
9976 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
9977 (ipa_get_jf_ancestor_result): Likewise.
9978 (propagate_vals_accross_pass_through): Use
9979 ipa_get_jf_pass_through_result to do all the value mappings.
9980 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
9981 type_preserved flag.
9982 (ipa_set_jf_cst_copy): New function.
9983 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
9984 (ipa_set_jf_arith_pass_through): Likewise.
9985 (ipa_set_ancestor_jf): Likewise.
9986 (compute_complex_assign_jump_func): Set type_preserved instead of
9988 (ipa_compute_jump_functions_for_edge): Likewise.
9989 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
9990 (update_jump_functions_after_inlining): Update type_preserved.
9991 Explicitely create jump functions when combining one with pass_through.
9992 (ipa_write_jump_function): Stream the type_preserved flags.
9993 (ipa_read_jump_function): Likewise.
9995 2013-08-27 Jakub Jelinek <jakub@redhat.com>
9996 Aldy Hernandez <aldyh@redhat.com>
9998 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
9999 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10000 * function.h (struct function): Add has_force_vect_loops and
10002 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10003 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10005 * gimple.def: Update CLAUSES comments.
10006 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10007 (gimple_build_omp_for): Add argument to prototype.
10008 (gimple_omp_for_kind): New.
10009 (gimple_omp_for_set_kind): New.
10010 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10011 GOVD_DATA_SHARE_CLASS.
10012 (enum omp_region_type): Add ORT_SIMD.
10013 (gimple_add_tmp_var): Handle ORT_SIMD.
10014 (gimplify_var_or_parm_decl): Same.
10015 (is_gimple_stmt): Same.
10016 (omp_firstprivatize_variable): Same.
10017 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10018 (omp_notice_variable): Handle ORT_SIMD.
10019 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10020 (omp_check_private): Handle ORT_SIMD.
10021 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10022 OMP_CLAUSE_SAFELEN.
10023 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10024 Handle OMP_CLAUSE_LASTPRIVATE.
10025 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10026 OMP_CLAUSE_SAFELEN.
10027 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10028 (gimplify_expr): Handle OMP_SIMD.
10029 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10030 (expand_GOMP_SIMD_VF): New.
10031 (expand_GOMP_SIMD_LAST_LANE): New.
10032 * internal-fn.def (GOMP_SIMD_LANE): New.
10033 (GOMP_SIMD_VF): New.
10034 (GOMP_SIMD_LAST_LANE): New.
10035 * omp-low.c: Include target.h.
10036 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10037 OMP_CLAUSE_SAFELEN.
10038 (check_omp_nesting_restrictions): Same.
10040 (lower_rec_simd_input_clauses): New.
10041 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10043 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10044 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10045 (expand_omp_build_assign): New.
10046 (expand_omp_for_init_counts): New.
10047 (expand_omp_for_init_vars): New.
10048 (extract_omp_for_update_vars): New.
10049 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10050 and rewrite accordingly.
10051 (expand_omp_simd): New.
10052 (expand_omp_for): Use expand_omp_simd.
10053 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10054 (lower_omp_for): Do not lower the body.
10055 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10056 in their own loops.
10057 * tree-flow.h (find_omp_clause): Remove prototype.
10058 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10059 forcing vectorization of the loop, or if flag_tree_vectorize.
10060 (gate_tree_if_conversion): Similarly.
10061 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10062 gimple_build_omp_for.
10063 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10064 * tree-parloops (create_parallel_loop): Pass kind argument to
10065 gimple_build_omp_for.
10066 * tree-pretty-print.c (dump_omp_clause): Add cases for
10067 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10068 OMP_CLAUSE__SIMDUID_.
10069 (dump_generic_node): Handle OMP_SIMD.
10070 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10071 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10072 unroll OMP_SIMD loops here.
10073 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10074 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10076 (vect_analyze_data_refs): Handle simd loops.
10077 * tree-vect-loop.c (vectorizable_live_operation): Handle
10079 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10080 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10081 (vectorizable_load): Same.
10082 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10083 (struct simduid_to_vf): New.
10084 (simduid_to_vf::hash): New.
10085 (simduid_to-vf::equal): New.
10086 (struct simd_array_to_simduid): New.
10087 (simd_array_to_simduid::hash): New.
10088 (simd_array_to_simduid::equal): New.
10089 (adjust_simduid_builtins): New.
10090 (struct note_simd_array_uses_struct): New.
10091 (note_simd_array_uses_cb): New.
10092 (note_simd_array_uses): New.
10093 (vectorize_loops): Handle simd hints and adjust simd builtins
10095 * tree-vectorizer.h (struct _stmt_vec_info): Add
10096 simd_lane_access_p field.
10097 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10098 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10099 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10100 (omp_clause_code_name): Same.
10101 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10102 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10103 * tree.def (OMP_SIMD): New entry.
10104 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10105 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10106 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10107 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10108 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10109 (OMP_CLAUSE_LINEAR_STEP): New.
10110 (OMP_CLAUSE_SAFELEN_EXPR): New.
10111 (OMP_CLAUSE__SIMDUID__DECL): New.
10112 (find_omp_clause): New prototype.
10114 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10116 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10117 Haswell processor detection.
10119 2013-08-27 Christian Widmer <shadow@umbrox.de>
10122 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10123 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10124 AVX2 capable processors.
10126 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10128 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10129 of vget_low_* with implementations in terms of other intrinsics.
10131 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10133 PR middle-end/57219
10134 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10135 values to -1, 0 and 1.
10137 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10139 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10140 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10141 (<optab><mode>3_insn): Remove.
10142 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10143 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10144 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10145 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10146 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10147 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10148 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10149 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10150 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10151 (ror<mode>3_insn): Likewise.
10152 * config/aarch64/predicates.md (aarch64_simd_register): New.
10154 2013-08-27 Richard Biener <rguenther@suse.de>
10156 PR tree-optimization/57521
10157 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10158 one edge is non-critical.
10159 (find_phi_replacement_condition): Make sure to use a non-critical
10160 edge. Cleanup and remove old bug workarounds.
10161 (bb_postdominates_preds): Remove.
10162 (if_convertible_loop_p_1): Do not compute post-dominators.
10163 (combine_blocks): Do not free post-dominators.
10164 (main_tree_if_conversion): Likewise.
10165 (pass_data_if_conversion): Add TODO_verify_ssa.
10167 2013-08-27 DJ Delorie <dj@redhat.com>
10169 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10171 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10173 * function.c (assign_parm_find_data_types): Set passed_mode and
10174 nominal_mode to the TYPE_MODE of nominal_type for the built
10175 pointer type in case of the struct-pass-by-reference.
10177 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10179 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10180 (UINT16_TYPE): Change default to "unsigned int".
10182 * config/avr/avr.opt (mfract-convert-truncate): New option.
10183 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
10184 is set, round negative fractional integers according to n1169
10185 when converting to integer types.
10187 2013-08-26 Jan Hubicka <jh@suse.cz>
10189 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
10190 methods can not be called indirectly when their address is not taken.
10192 2013-08-26 Jan Hubicka <jh@suse.cz>
10194 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
10197 2013-08-26 Jan Hubicka <jh@suse.cz>
10199 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
10202 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10204 * reload.c (find_valid_class): Allow classes that do not include
10205 FIRST_PSEUDO_REGISTER - 1.
10207 2013-08-26 Jan Hubicka <jh@suse.cz>
10209 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
10210 fix edge count/frequency when speculation failed; fix type check
10211 for the direct call.
10213 2013-08-26 Jan Hubicka <jh@suse.cz>
10215 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
10217 2013-08-26 Jan Hubicka <jh@suse.cz>
10219 * ipa-inline-transform.c (inline_transform): Be ready for basic block
10220 to be changed by edge redirection.
10222 2013-08-26 Jan Hubicka <jh@suse.cz>
10224 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
10225 formating; add sanity check.
10226 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
10227 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
10228 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
10229 (resolve_noninline_speculation): Update callee keys, too.
10231 2013-08-26 Jan Hubicka <jh@suse.cz>
10233 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
10234 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
10236 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10238 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
10241 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
10243 * config/i386/i386.c (ix86_debug_options): Remove prototype.
10244 (x86_64_elf_select_section): Ditto.
10245 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
10247 (ix86_pass_by_reference): Ditto.
10248 (output_set_got): Ditto.
10249 (ix86_unary_operator_ok): Ditto.
10250 (ix86_expand_builtin): Ditto.
10252 2013-08-23 Jan Hubicka <jh@suse.cz>
10254 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
10256 2013-08-23 Jan Hubicka <jh@suse.cz>
10258 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
10259 (tree_decl_with_vis): Add FINAL field.
10261 2013-08-23 Jeff Law <law@redhat.com>
10263 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
10264 the debugging dump when the expression is fully redundant.
10266 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10268 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
10269 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
10270 * pretty-print.c (pp_formatted_text_data): Likewise.
10271 (pp_write_text_to_stream): Likewise.
10272 (pp_write_text_as_dot_label_to_stream): Likewise.
10273 (pp_append_r): Likewise.
10274 (pp_format): Likewise.
10275 (pp_flush): Likewise.
10276 (pp_clear_output_area): Likewise.
10277 (pp_append_text): Likewise.
10278 (pp_formatted_text): Likewise.
10279 (pp_remaining_character_count_for_line): Likewise.
10280 (pp_newline): Likewise.
10281 (pp_character): Likewise.
10282 (output_buffer::~output_buffer): Define.
10283 (pretty_printer::~pretty_printer): Destruct output buffer.
10284 * pretty-print.h (output_buffer::~output_buffer): Declare.
10285 (pretty_printer::~pretty_printer): Declare virtual.
10287 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
10290 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
10291 HOST_WIDE_INT_M1U): New macros.
10292 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
10293 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
10294 unsigned -1 for lshift.
10295 * cse.c (cse_insn): Likewise.
10296 * double-int.c (rshift_double, lshift_double): Likewise.
10297 * builtins.c (fold_builtin_bitop): Likewise.
10298 * combine.c (force_to_mode): Likewise.
10299 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
10300 * simplify-rtx.c (simplify_const_unary_operation,
10301 simplify_const_binary_operation): Likewise.
10302 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
10303 check_va_list_escapes): Likewise.
10304 * rtlanal.c (nonzero_bits1): Likewise.
10305 * expmed.c (expand_smod_pow2): Likewise.
10306 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
10308 2013-08-23 Jan Hubicka <jh@suse.cz>
10310 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
10311 as having address taken.
10313 2013-08-23 Jan Hubicka <jh@suse.cz>
10315 * ipa-utils.h (method_class_type): Declare.
10316 * ipa-devirt.c (method_class_type): Export.
10318 * cgraphunit.c (analyze_functions): Do basic devirtualization;
10319 do not walk base classes of anonymous types.
10321 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
10323 PR rtl-optimization/58220
10324 PR regression/58221
10325 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
10326 handle SEQUENCE insns properly.
10328 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10330 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
10332 (pp_newline_and_indent): Likewise.
10333 (pp_separate_with): Likewise.
10334 * pretty-print.c (pp_newline_and_flush): Define.
10335 (pp_newline_and_indent): Likewise.
10336 (pp_separate_with): Likewise.
10338 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10341 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
10342 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
10344 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
10346 * config/i386/predicates.md (ext_sse_reg_operand): New.
10347 * config/i386/i386.md (*movti_internal): Use
10348 predicate to determine if EVEX is needed.
10349 (*movsi_internal): Ditto.
10350 (*movdf_internal): Ditto.
10351 (*movsf_internal): Ditto.
10352 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
10354 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10356 PR tree-optimization/58209
10357 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
10358 (find_tail_calls): Give up for pointer result types if m is non-NULL.
10359 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
10360 emit POINTER_PLUS_EXPR.
10361 (create_tailcall_accumulator): For pointer result type accumulate in
10364 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
10366 * configure.ac: Add backslashes missing from the last change.
10367 * configure: Regenerate.
10369 2013-08-22 Jan Hubicka <jh@susue.cz>
10371 * ipa.c (function_and_variable_visibility): First remember function
10372 was global and then make it local.
10374 2013-08-22 Julian Brown <julian@codesourcery.com>
10376 * configure.ac: Add aarch64 to list of arches which use "nop" in
10378 * configure: Regenerate.
10380 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10382 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
10383 gnu_libc_has_function.
10384 * config/s390/tpf.h: Likewise.
10386 2013-08-22 Jan Hubicka <jh@susue.cz>
10388 * timevar.c (validate_phases): Add cast.
10390 2013-08-22 Jan Hubicka <jh@susue.cz>
10392 * timevar.c (validate_phases): Use size_t for memory.
10393 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
10395 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
10397 * pretty-print.h (output_buffer::output_buffer): Declare.
10398 (pretty_printer::pretty_printer): Likewise.
10399 (pp_construct): Remove.
10400 * pretty-print.c (output_buffer::output_buffer): Define.
10401 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
10402 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
10403 (print_gimple_expr): Likewise.
10404 (print_gimple_seq): Likewise.
10405 (gimple_dump_bb): Likewise.
10406 * sched-vis.c (dump_value_slim): Likewise.
10407 (dump_insn_slim): Likewise.
10408 (dump_rtl_slim): Likewise.
10409 (str_pattern_slim): Likewise.
10410 * tree-mudflap.c (mf_varname_tree): Likewise.
10411 * graph.c (print_graph_cfg): Likewise.
10412 (start_graph_dump): Likewise.
10413 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
10415 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
10416 pretty printer initialization.
10417 * coretypes.h (diagnostic_context): Remove superflous type alias
10419 (pretty_printer): Likewise. Declare directly as a class.
10420 (pretty_print_info): Remove declaration as class.
10421 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
10422 and pp_clear_output_area.
10423 (asan_add_global): Likewise.
10425 2013-08-22 Jan Hubicka <jh@suse.cz>
10427 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
10428 * ipa-utils.h (update_type_inheritance_graph): Declare.
10429 (possible_polymorphic_call_target_p): Declare.
10430 (possible_polymorphic_call_target_p): New.
10431 * ipa-devirt.c: Update toplevel comments.
10432 (cached_polymorphic_call_targets): Move up.
10433 (odr_type_d): Move ID down.
10434 (polymorphic_type_binfo_p): Update comment.
10435 (odr_hasher::remove): Likewise;
10436 (get_odr_type): Set anonymous_namespace.
10437 (dump_odr_type): Dump it.
10438 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
10439 (maybe_record_node): Record node in cached_polymorphic_call_targets.
10440 (record_binfo): Add comment.
10441 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
10443 (devirt_node_removal_hook): Do not iCE when cache is freed.
10444 (possible_polymorphic_call_target_p): New predicate.
10445 (update_type_inheritance_graph): New function.
10447 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
10448 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
10449 Sergey Lega <sergey.s.lega@intel.com>
10450 Anna Tikhonova <anna.tikhonova@intel.com>
10451 Ilya Tocar <ilya.tocar@intel.com>
10452 Andrey Turetskiy <andrey.turetskiy@intel.com>
10453 Ilya Verbin <ilya.verbin@intel.com>
10454 Kirill Yukhin <kirill.yukhin@intel.com>
10455 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10457 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
10458 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
10459 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
10460 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
10461 (OPTION_MASK_ISA_AVX2_UNSET): Update.
10462 (OPTION_MASK_ISA_AVX512F_UNSET): New.
10463 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
10464 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
10465 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
10466 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
10467 OPT_mavx512pf, OPT_mavx512er cases.
10468 * config/i386/constraints.md (v): New constraint.
10469 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
10470 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
10471 (bit_AVX512CD): New.
10472 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
10473 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
10474 * config/i386/i386-c.c (ix86_target_macros_internal):
10475 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
10477 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
10478 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
10479 * config/i386/i386.c (regclass_map, dbx_register_map)
10480 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
10481 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
10482 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
10483 -mavx512pf options.
10484 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
10485 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
10486 -mavx512pf options. Fix formatting.
10487 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
10488 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
10489 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
10490 -mavx512cd, -mavx512pf options.
10491 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
10492 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
10493 (ix86_preferred_output_reload_class): Replace SSE_REGS with
10495 (ix86_hard_regno_mode_ok): Support 512-bit registers.
10496 (ix86_set_reg_reg_cost): Ditto.
10497 (x86_order_regs_for_local_alloc): Ditto.
10498 (MAX_VECT_LEN): Extend to 64-byte.
10499 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
10500 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
10501 (TARGET_AVX512ER, TARGET_AVX512CD): New.
10502 (BIGGEST_ALIGNMENT): Extend to 512-bits.
10503 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
10504 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
10505 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
10506 (SSE_REG_MODE_P): Support new modes.
10507 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
10508 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
10509 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
10510 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
10511 (REG_CLASS_CONTENTS): Add new registers.
10512 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
10513 (EXT_REX_SSE_REGNO_P): New.
10514 (HI_REGISTER_NAMES): Add new registers.
10515 * config/i386/i386.md: Define constants for new registers.
10516 (mode): Add new 512-bit modes.
10517 (prefix): Support evex prefix.
10518 (isa): Support avx512f, noavx512f, fma_avx512f.
10519 (ssemodesuffix): Add new 512-bit modes.
10521 (*movxi_internal_avx512f): Ditto.
10522 (*movdi_internal): Replace constraint "x" with the new constraint "v".
10524 (*movsi_internal): Likewise.
10525 (*movdf_internal): Likewise.
10526 (*movsf_internal): Likewise.
10527 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
10528 (<code><mode>3): Likewise.
10529 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
10531 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
10532 with the new constraint "v".
10533 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
10535 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
10536 with the new constraint "v".
10537 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
10538 (<sse2>_storedqu<avxsizesuffix>): Likewise.
10539 (*<plusminus_insn><mode>3): Likewise.
10540 (<sse>_vm<plusminus_insn><mode>3): Likewise.
10541 (*mul<mode>3): Likewise.
10542 (<sse>_vmmul<mode>3): Likewise.
10543 (<sse>_div<mode>3): Likewise.
10544 (<sse>_vmdiv<mode>3): Likewise.
10545 (<sse>_sqrt<mode>2): Likewise.
10546 (<sse>_vmsqrt<mode>2): Likewise.
10547 (*<code><mode>3_finite): Likewise.
10548 (*<code><mode>3) <smaxmin>: Likewise.
10549 (<sse>_vm<code><mode>3): Likewise.
10550 (*<code><mode>3) <any_logic>: Likewise.
10551 (*fma_fmadd_<mode>): Likewise.
10552 (*fma_fmsub_<mode>): Likewise.
10553 (*fma_fnmadd_<mode>): Likewise.
10554 (*fma_fnmsub_<mode>): Likewise.
10555 (*fma_fmaddsub_<mode>): Likewise.
10556 (*fma_fmsubadd_<mode>): Likewise.
10557 (*fmai_fmadd_<mode>): Likewise.
10558 (*fmai_fmsub_<mode>): Likewise.
10559 (*fmai_fnmadd_<mode>): Likewise.
10560 (*fmai_fnmsub_<mode>): Likewise.
10561 (sse_cvtsi2ss): Likewise.
10562 (sse_cvtsi2ssq): Likewise.
10563 (sse_cvtss2si): Likewise.
10564 (sse_cvtss2si_2): Likewise.
10565 (sse_cvtss2siq): Likewise.
10566 (sse_cvtss2siq_2): Likewise.
10567 (sse_cvttss2si): Likewise.
10568 (sse_cvtss2siq_2): Likewise.
10569 (float<sseintvecmodelower><mode>2): Likewise.
10570 (sse2_cvtsd2si_2): Likewise.
10571 (sse2_cvtsd2siq_2): Likewise.
10572 (*<plusminus_insn><mode>3): Likewise.
10573 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
10574 (*<sse4_1_avx2>_mul<mode>3): Likewise.
10575 (ashr<mode>3): Likewise.
10576 (<shift_insn><mode>3): Likewise.
10577 (avx2_<code><mode>3): Likewise.
10578 (*avx2_<code><mode>3): Likewise.
10579 (*andnot<mode>3): Likewise.
10580 (*<code><mode>3) <any_logic>: Likewise.
10581 (abs<mode>2): Likewise.
10582 (avx2_permvar<mode>): Likewise.
10583 (avx2_perm<mode>_1): Likewise.
10584 (*avx_vpermilp<mode>): Likewise.
10585 (avx_vpermilvar<mode>3): Likewise.
10586 (avx2_ashrv<mode>): Likewise.
10587 (avx2_<shift_insn>v<mode>): Likewise.
10588 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
10590 * doc/rtl.texi: Document XImode.
10592 2013-08-21 Jeff Law <law@redhat.com>
10594 * tree-flow.h (register_jump_thread): Pass vector of edges
10595 instead of each important edge.
10596 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
10597 thread path into a vector and pass that to register_jump_thread.
10598 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
10599 passed in edge vector to the current 3-edge form.
10602 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
10604 * dce.c (fini_dce): Call df_analyze again just in case
10605 delete_unmarked_insns removed anything.
10607 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
10609 * reload.h (struct reg_equivs): Rename to ..
10610 (struct reg_equivs_s): .. this.
10612 2013-08-20 Martin Liska <marxin.liska@gmail.com>
10614 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
10616 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10618 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
10620 2013-08-21 Jeff Law <law@redhat.com>
10622 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
10623 simplify assignments too. If the RHS collapses to a singleton
10624 range, then return the value for the range.
10626 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
10628 * config/i386/sse.md (V16): Rename to...
10630 (mov<mode>): Update iterator name.
10631 (*mov<mode>_internal): Ditto.
10632 (push<mode>1): Ditto.
10633 (movmisalign<mode>): Ditto.
10635 2013-08-20 Jan Hubicka <jh@suse.cz>
10638 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
10639 entry for direct edges.
10640 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
10642 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10644 Revert my last two changes, r201865 and r201864:
10647 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10649 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
10650 instances can own GC refs.
10652 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
10653 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
10654 (gcc::context::gt_pch_nx): Likewise.
10655 (gcc::context::gt_pch_nx): Likewise.
10656 * ggc.h (gt_ggc_mx <T>): New.
10657 (gt_pch_nx_with_op <T>): New.
10658 (gt_pch_nx <T>): New.
10659 * passes.c (opt_pass::gt_ggc_mx): New.
10660 (opt_pass::gt_pch_nx): New.
10661 (opt_pass::gt_pch_nx_with_op): New.
10662 (pass_manager::gt_ggc_mx): New.
10663 (pass_manager::gt_pch_nx): New.
10664 (pass_manager::gt_pch_nx_with_op): New.
10665 (pass_manager::operator new): Use
10666 ggc_internal_cleared_alloc_stat rather than xcalloc.
10667 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
10668 (pass_manager::gt_ggc_mx): New.
10669 (pass_manager::gt_pch_nx): New.
10670 (pass_manager::gt_pch_nx_with_op): New.
10671 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
10672 (opt_pass::operator new): New.
10673 (opt_pass::gt_ggc_mx): New.
10674 (opt_pass::gt_pch_nx): New.
10675 (opt_pass::gt_pch_nx_with_op): New.
10678 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10680 * Makefile.in (GTFILES): Add context.h.
10681 * context.c (gcc::context::operator new): New.
10682 (gcc::context::gt_ggc_mx): New.
10683 (gcc::context::gt_pch_nx): New.
10684 (gcc::context::gt_pch_nx): New.
10685 * context.h (gcc::context): Add GTY((user)) marking.
10686 (gcc::context::operator new): New.
10687 (gcc::context::gt_ggc_mx): New.
10688 (gcc::context::gt_pch_nx): New.
10689 (gcc::context::gt_pch_nx): New.
10690 (g): Add GTY marking.
10691 (gt_ggc_mx (gcc::context *)): New.
10692 (gt_pch_nx (gcc::context *)): New.
10693 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
10694 void *cookie)): New.
10695 * gengtype.c (open_base_files) <ifiles>: Add context.h.
10697 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
10699 * dce.c (fini_dce): Call df_analyze again just in case
10700 delete_unmarked_insns removed anything.
10702 2013-08-20 Teresa Johnson <tejohnson@google.com>
10704 PR rtl-optimizations/57451
10705 * final.c (reemit_insn_block_notes): Prevent lexical blocks
10706 from crossing split section boundaries.
10708 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10710 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
10711 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
10712 with MULTLIB_DEFAULTS.
10714 2013-08-20 Nick Clifton <nickc@redhat.com>
10716 * target.def (narrow_volatile_bitfield): Note that the default
10717 value is false, not !TARGET_STRICT_ALIGN.
10718 * doc/tm.texi: Regenerate.
10720 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
10722 Fix LIB_SPEC for systems without libpthread.
10724 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
10725 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
10727 * config/i386/linux-common.h: Likewise.
10728 * config/mips/linux-common.h: Likewise.
10730 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
10732 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
10735 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10737 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
10738 instances can own GC refs.
10740 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
10741 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
10742 (gcc::context::gt_pch_nx): Likewise.
10743 (gcc::context::gt_pch_nx): Likewise.
10744 * ggc.h (gt_ggc_mx <T>): New.
10745 (gt_pch_nx_with_op <T>): New.
10746 (gt_pch_nx <T>): New.
10747 * passes.c (opt_pass::gt_ggc_mx): New.
10748 (opt_pass::gt_pch_nx): New.
10749 (opt_pass::gt_pch_nx_with_op): New.
10750 (pass_manager::gt_ggc_mx): New.
10751 (pass_manager::gt_pch_nx): New.
10752 (pass_manager::gt_pch_nx_with_op): New.
10753 (pass_manager::operator new): Use
10754 ggc_internal_cleared_alloc_stat rather than xcalloc.
10755 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
10756 (pass_manager::gt_ggc_mx): New.
10757 (pass_manager::gt_pch_nx): New.
10758 (pass_manager::gt_pch_nx_with_op): New.
10759 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
10760 (opt_pass::operator new): New.
10761 (opt_pass::gt_ggc_mx): New.
10762 (opt_pass::gt_pch_nx): New.
10763 (opt_pass::gt_pch_nx_with_op): New.
10765 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10767 * Makefile.in (GTFILES): Add context.h.
10768 * context.c (gcc::context::operator new): New.
10769 (gcc::context::gt_ggc_mx): New.
10770 (gcc::context::gt_pch_nx): New.
10771 (gcc::context::gt_pch_nx): New.
10772 * context.h (gcc::context): Add GTY((user)) marking.
10773 (gcc::context::operator new): New.
10774 (gcc::context::gt_ggc_mx): New.
10775 (gcc::context::gt_pch_nx): New.
10776 (gcc::context::gt_pch_nx): New.
10777 (g): Add GTY marking.
10778 (gt_ggc_mx (gcc::context *)): New.
10779 (gt_pch_nx (gcc::context *)): New.
10780 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
10781 void *cookie)): New.
10782 * gengtype.c (open_base_files) <ifiles>: Add context.h.
10784 2013-08-20 Alan Modra <amodra@gmail.com>
10787 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
10788 (rs6000_emit_epilogue): Likewise.
10790 2013-08-19 Dehao Chen <dehao@google.com>
10792 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
10794 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
10795 Jakub Jelinek <jakub@redhat.com>
10797 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
10798 (BUILT_IN_FABSD64): Likewise.
10799 (BUILT_IN_FABSD128): Likewise.
10800 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
10801 (fold_builtin_1): Likewise.
10802 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
10803 destination and source operands.
10804 (*abstd2_fpr): Likewise.
10805 (*nabstd2_fpr): Likewise.
10807 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
10809 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
10812 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
10814 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
10817 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
10819 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
10820 * builtins.c (default_libc_has_function): New.
10821 (gnu_libc_has_function): Ditto.
10822 (no_c99_libc_has_function): Ditto.
10823 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
10824 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
10825 (fold_builtin_sincos): Likewise.
10826 (fold_builtin_cexp): Likewise.
10827 * builtins.def (DEF_C94_BUILTIN): Likewise.
10828 (DEF_C99_BUILTIN): Likewise.
10829 (DEF_C99_C90RES_BUILTIN): Likewise.
10830 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
10831 definitions to using this define.
10832 * config/darwin-protos.h (darwin_libc_has_function): New.
10833 * config/darwin.c (darwin_libc_has_function): Ditto.
10834 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
10835 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
10836 * config/darwin.h: Ditto.
10837 * config/elfos.h: Ditto.
10838 * config/freebsd.h: Ditto.
10839 * config/i386/cygming.h: Ditto.
10840 * config/i386/djgpp.h: Ditto.
10841 * config/i386/i386-interix.h: Ditto.
10842 * config/microblaze/microblaze.h: Ditto.
10843 * config/mmix/mmix.h: Ditto.
10844 * config/gnu-user.h: Ditto.
10845 * config/ia64/hpux.h: Ditto.
10846 * config/pa/pa-hpux.h: Ditto.
10847 * config/pdp11/pdp11.h: Ditto.
10848 * config/picochip/picochip.h: Ditto.
10849 * config/linux.h: Ditto.
10850 * config/netbsd.h: Ditto.
10851 * config/openbsd.h: Ditto.
10852 * config/rs6000/aix43.h: Ditto.
10853 * config/rs6000/aix51.h: Ditto.
10854 * config/rs6000/aix52.h: Ditto.
10855 * config/rs6000/aix53.h: Ditto.
10856 * config/rs6000/aix61.h: Ditto.
10857 * config/rs6000/darwin.h: Ditto.
10858 * config/rs6000/linux.h: Ditto.
10859 * config/rs6000/linux64.h: Ditto.
10860 * config/s390/tpf.h: Ditto.
10861 * config/sol2-10.h: Ditto.
10862 * config/sol2.h: Ditto.
10863 * config/vms/vms.h: Ditto.
10864 * config/vxworks.h: Ditto.
10865 * config/linux-android.c (linux_android_libc_has_function):
10866 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
10867 * config/linux-protos.h (linux_android_libc_has_function):
10869 * config/i386/i386.c (ix86_libc_has_function): New.
10870 * config/i386/i386-protos.h
10871 (ix86_libc_has_function): New declaration.
10872 * config/i386/i386.md
10873 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
10874 ("isinf<mode>2): Likewise.
10875 * convert.c (convert_to_integer): Using new target hook
10876 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
10877 TARGET_C99_FUNCTIONS.
10878 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
10879 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
10880 * coretypes.h (function_class): New enum for different
10881 classes of functions.
10882 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
10883 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
10884 (TARGET_HAS_SINCOS): Likewise.
10885 (TARGET_LIBC_HAS_FUNCTION): New.
10886 * doc/tm.texi: Regenerated.
10887 * targhooks.h (default_libc_has_function): New declaration.
10888 (no_c99_libc_has_function): Ditto.
10889 (gnu_libc_has_function): Ditto.
10890 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
10891 and TARGET_HAS_SINCOS.
10893 2013-08-18 Jan Hubicka <jh@suse.cz>
10895 * Makeifle-in (ipa-devirt.o): New.
10896 (GTFILES): Add ipa-utils.h and ipa-devirt.c
10897 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
10898 (analyze_functions): Look into possible targets of polymorphic call.
10899 * dumpfile.c (dump_files): Add type-inheritance dump.
10900 * dumpfile.h (TDI_inheritance): New.
10901 * ipa-devirt.c: New file.
10902 * ipa-utils.h (odr_type_d): Forward declare.
10903 (odr_type): New type.
10904 (build_type_inheritance_graph): Declare.
10905 (possible_polymorphic_call_targets): Declare and introduce inline
10906 variant when only edge is pased.
10907 (dump_possible_polymorphic_call_targets): Likewise.
10908 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
10909 * tree.c (type_in_anonymous_namespace_p): Break out from ...
10910 (types_same_for_odr): ... here.
10911 * tree.h (type_in_anonymous_namespace_p): Declare.
10913 2013-08-18 Jakub Jelinek <jakub@redhat.com>
10915 PR tree-optimization/58006
10916 * tree-parloops.c (take_address_of): Don't ICE if get_name
10918 (eliminate_local_variables_stmt): Remove clobber stmts.
10920 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
10922 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
10923 error message is issued for an alias to undefined symbol.
10925 2013-08-18 Jan Hubicka <jh@suse.cz>
10927 * cgraph.c (cgraph_create_indirect_edge): Discover
10928 polymorphic calls and record basic info into indirect_info.
10929 * gimple-fold.c (gimple_fold_call): When doing BINFO based
10930 devirtualization, ignore objc function calls.
10931 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
10932 call with no parm index info.
10933 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
10934 * tree.c (virtual_method_call_p): New function.
10935 * tree.h (virtual_method_call_p): Declare.
10937 2013-08-16 Jan Hubicka <jh@suse.cz>
10939 PR middle-end/58179
10940 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
10942 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
10944 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
10947 2013-08-16 David Malcolm <dmalcolm@redhat.com>
10949 * gengtype.c (type_for_name): Add special-case support for
10950 locating types within the "gcc::" namespace.
10951 (open_base_files): Emit a "using namespace gcc" directive.
10953 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10956 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
10957 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
10959 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
10960 array instead of each individual operand as a separate argument.
10961 (emit_fusion_gpr_load): Likewise.
10962 (expand_fusion_gpr_load): Add new function declaration.
10964 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
10965 signature to have the operands passed as an array, instead of as
10966 separate arguments. Allow ZERO_EXTEND to be in the memory
10967 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
10968 depend on the register live/dead flags when peepholes are run.
10969 (expand_fusion_gpr_load): New function to be called from the
10970 peephole2 pass, to change the register that addis sets to be the
10972 (emit_fusion_gpr_load): Change the calling signature to have the
10973 operands passed as an array, instead of as separate arguments.
10974 Allow ZERO_EXTEND to be in the memory address, and also
10975 SIGN_EXTEND if -mpower8-fusion-sign.
10977 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
10978 unspec enumeration.
10979 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
10980 adjust the register addis loads up in the peephole2 pass. Do not
10981 depend on the register live/dead state when the peephole pass is done.
10983 2013-08-16 David Malcolm <dmalcolm@redhat.com>
10985 * gengtype.c (create_user_defined_type): Ensure that the kind
10986 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
10987 declaration is seen before the GTY((user)) marking.
10989 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
10992 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
10994 2013-08-16 Jan Hubicka <jh@suse.cz>
10996 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
10997 arugment expected_type.
10998 (gimple_fold_call): Use it.
10999 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11000 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11001 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11002 (try_make_edge_direct_virtual_call): Likewise.
11003 * tree.c (obj_type_ref_class): New.
11004 * tree.h (obj_type_ref_class): Use it.
11006 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11008 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11009 (rtl_slim_pp): Likewise.
11010 (init_rtl_slim_pretty_print): Likewise.
11011 (dump_value_slim): Don't call it. Use local pretty printer.
11012 (dump_insn_slim): Likewise.
11013 (dump_rtl_slim): Likewise.
11014 (str_pattern_slim): Likewise.
11015 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11018 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11020 PR tree-optimization/58164
11021 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11022 walk gimple_goto_dest of GIMPLE_GOTO.
11024 PR tree-optimization/58165
11025 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11026 bi_call must be the last stmt in a bb, don't split_block, instead
11027 use fallthru edge from it and give up if there is none.
11028 Release conds vector when returning early.
11030 2013-08-14 Xinliang David Li <davidxl@google.com>
11032 * config/i386/i386.c (ix86_option_override_internal):
11033 Remove unused variable and field.
11035 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11038 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11039 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11040 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11041 and Linux, correct BLKmode alignment when 128-bit alignment is
11042 required and compatibility flag is not set.
11043 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11044 for zero-size arguments when compatibility flag is not set.
11046 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11048 PR tree-optimization/58145
11049 * tree-sra.c (build_ref_for_offset): If prev_base has
11050 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11052 2013-08-14 Xinliang David Li <davidxl@google.com>
11054 * config/i386/i386.c (ix86_option_override_internal):
11055 Fix uninitialized variable error.
11057 2013-08-14 Xinliang David Li <davidxl@google.com>
11059 * config/i386/i386.opt: Define two new options.
11060 * config/i386/x86-tune.def: Add arch selector field in macros.
11061 * config/i386/i386.h: Adjust macro definition.
11062 * config/i386/i386.c (ix86_option_override_internal):
11064 (parse_mtune_ctrl_str): New function.
11065 (set_ix86_tune_features): New function.
11066 (ix86_function_specific_restore): Call the new helper function.
11068 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11070 PR rtl-optimization/57662
11071 * sel-sched.c (code_motion_process_successors): When the current insn
11072 is removed after the recursive traversal, break from the loop.
11073 Add comments and debug printouts.
11075 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11076 Alexandre Oliva <aoliva@redhat.com>
11079 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11080 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11081 there also UNSPEC_PLTOFF.
11083 2013-08-14 Marek Polacek <polacek@redhat.com>
11085 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11086 NUM_CONDITIONS bit positions.
11088 2013-08-13 Cary Coutant <ccoutant@google.com>
11090 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11091 (attr_checksum): Hash vector contents instead of pointer.
11092 (attr_checksum_ordered): Likewise.
11094 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11096 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11097 when Pmode != word_mode. Add length_address attribute.
11098 (sse3_monitor_<mode>): Merge from sse3_monitor and
11099 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11100 Pmode != word_mode. Update insn length attribute.
11101 * config/i386/i386.c (ix86_option_override_internal): Update
11102 ix86_gen_monitor selection for merged sse3_monitor insn.
11104 2013-08-13 Julian Brown <julian@codesourcery.com>
11106 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11107 perform invalid legitimization on greater-than-word-size modes for
11108 TARGET_E500_DOUBLE.
11110 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11112 * ira.c (setup_class_translate_array): Use aclass instead of cl
11113 for classes not fully covered by allocno classes.
11115 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11117 PR tree-optimization/57661
11118 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11119 * tree-inline.c (tree_function_versioning): Initialize it.
11120 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11121 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11122 that is not being copied.
11125 * asan.c (instrument_strlen_call): Fix typo in comment.
11126 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11128 2013-08-13 Steve Ellcey <sellcey@mips.com>
11130 * config/mips/mips.md (prefetch): Use lw instead of ld on
11131 loongson in 32bit mode.
11133 2013-08-13 Nick Clifton <nickc@redhat.com>
11135 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11137 2013-08-13 Jan Hubicka <jh@suse.cz>
11139 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11140 introduced edge; fix typo in sanity check.
11141 (cgraph_resolve_speculation): Export; improve diagnostic.
11142 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11143 speculation at type mismatch.
11144 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11145 (cgraph_resolve_speculation): Declare.
11146 (symtab_can_be_discarded): New function.
11147 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11148 * ipa-inline-transform.c (speculation_removed): New global var.
11149 (clone_inlined_nodes): See if speculation can be removed.
11150 (inline_call): If speculations was removed, we growths may not match.
11151 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11152 (speculation_useful_p): New function.
11153 (resolve_noninline_speculation): New function.
11154 (inline_small_functions): Resolve useless speculations.
11155 * ipa-inline.h (speculation_useful_p): Declare
11156 * ipa.c (can_replace_by_local_alias): Simplify.
11157 (ipa_profile): Produce speculative calls in non-lto, too;
11158 add simple cost model; produce local aliases.
11160 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11162 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11165 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11167 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11169 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11171 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11172 iterator instead of X87MODEF.
11174 2013-08-12 Perez Read <netfirewall@gmail.com>
11177 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11178 operand 0 for intel asm alternative.
11179 (*movabs<mode>_2): Ditto for operand 1.
11181 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
11183 * config/aarch64/arm_none.h
11184 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
11186 2013-08-12 Nick Clifton <nickc@redhat.com>
11188 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
11190 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
11192 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
11193 expand for QI/HImode operand to produce more effictive code for
11194 unsigned char(short) --> float(double) conversion.
11196 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
11198 * doc/invoke.texi: Mention that -ftls-model does not force the final
11201 2013-08-12 Marek Polacek <polacek@redhat.com>
11202 Marc Glisse <marc.glisse@inria.fr>
11204 PR tree-optimization/57980
11205 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
11206 when creating -1 constant.
11208 2013-08-10 Jan Hubicka <jh@suse.cz>
11210 Workaround binutils PR14342.
11211 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
11212 (gimple_init_edge_profiler): Likewise.
11213 (gimple_gen_ic_func_profiler): Likewise.
11215 2013-08-09 Jan Hubicka <jh@suse.cz>
11217 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
11219 2013-08-09 Xinliang David Li <davidxl@google.com>
11221 * config/i386/stringop.def: New file.
11222 * config/i386/stringop.opt: New file.
11223 * config/i386/i386-opts.h: Include stringopt.def.
11224 * config/i386/i386.opt: Include stringopt.opt.
11225 * config/i386/i386.c (ix86_option_override_internal):
11226 Override default size based stringop inline strategies with options.
11227 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
11230 2013-08-09 Jan Hubicka <jh@suse.cz>
11232 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
11234 2013-08-09 Jan Hubicka <jh@suse.cz>
11236 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
11238 (dump_cgraph_node): Dump profile-id.
11239 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
11240 and common_target_probability.
11241 * lto-cgraph.c (lto_output_edge): Stream common targets.
11242 (lto_output_node): Stream profile ids.
11243 (input_node): Stream profile ids.
11244 (input_edge): Stream common targets.
11245 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
11246 * ipa.c: Include value-prof.h
11247 (ipa_profile_generate_summary): Turn indirect call statement histograms
11248 into common targets.
11249 (ipa_profile): Turn common targets into speculative edges.
11251 2013-08-09 Jan Hubicka <jh@suse.cz>
11253 * cgraph.h (cgraph_node): Add profile_id.
11254 * value-prof.c (cgraph_node_map): Turn into pointer_map.
11255 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
11256 (del_node_map): Update.
11257 (find_func_by_funcdef_no): Replace by ...
11258 (find_func_by_profile_id): ... this one.
11259 (gimple_ic_transform): Do not remove useful histograms when
11260 speculation is not done; dump info when indirect call removal
11262 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
11263 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
11264 (__gcov_indirect_call_profiler_v2): .. this one.
11265 * profile.h (init_node_map): Update.
11266 * coverage.c (coverage_compute_profile_id): New function.
11267 * coverage.h (coverage_compute_profile_id): Declare.
11268 * tree-profile.c (init_ic_make_global_vars): Make
11269 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
11270 (gimple_init_edge_profiler): Update prototype of
11271 __gcov_indirect_call_profiler.
11272 (gimple_gen_ic_func_profiler): Simplify.
11273 (tree_profiling): Use init_node_map
11275 2013-08-09 Jan Hubicka <jh@suse.cz>
11277 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
11278 non-speculative refs.
11279 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
11280 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
11281 (cgraph_set_call_stmt): Likewise.
11282 (cgraph_create_edge_1): Fix release checking compilatoin;
11283 clear lto_stmt_uid.
11284 (cgraph_free_edge): Free indirect info.
11285 (cgraph_turn_edge_to_speculative): New function.
11286 (cgraph_speculative_call_info): New function.
11287 (cgraph_make_edge_direct): Return direct edge; handle speculation.
11288 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
11289 (dump_cgraph_node): Dump speculation.
11290 (verify_edge_count_and_frequency): Accept speculative edges.
11291 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
11292 (verify_cgraph_node): Handle speculation.
11293 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
11294 (cgraph_set_call_stmt): Update prototype.
11295 (cgraph_make_edge_direct): Update prototype.
11296 (cgraph_speculative_call_info): Declare.
11297 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
11298 to change; update call of ipa_find_references.
11299 * ipa-ref.c (ipa_record_reference): Fix return value; clear
11300 lto_stmt_uid and speculative flags.
11301 (ipa_dump_references): Dump speculation.
11302 (ipa_clone_references): Clone speculative flag.
11303 (ipa_clone_referring): Likewise.
11304 (ipa_clone_ref): New function.
11305 (ipa_find_reference): Look into lto_stmt_uids
11306 (ipa_clear_stmts_in_references): Do not clear speculative calls.
11307 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
11308 (ipa_find_reference): Update declaration.
11309 (ipa_clone_ref): Declare.
11310 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
11311 stream speculative flag.
11312 (lto_output_ref): Stream statements uids and speculation.
11313 (input_ref): Likewise.
11314 (input_edge): Stream speuclation.
11315 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
11316 (cgraph_set_call_stmt_including_clones): Handle speculation.
11317 * ipa-inline.c (heap_edge_removal_hook): New function.
11318 (inline_small_functions): Register it.
11319 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
11320 also initialize refs.
11321 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
11323 (try_make_edge_direct_simple_call): Likewise.
11324 (try_make_edge_direct_simple_call): Likewise.
11325 (update_indirect_edges_after_inlining): Likewise.
11326 (remove_described_reference): Look proper lto_stmt_uid.
11327 (propagate_controlled_uses): Likewise.
11328 (propagate_controlled_uses): Liekwise.
11329 * tree-inline.c (copy_bb): Copy speculative edges.
11330 (redirect_all_calls): New function.
11331 (copy_cfg_body): Do redirection after loop info is updated.
11332 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
11334 2013-08-09 Jan Hubicka <jh@suse.cz>
11336 * lto-streamer-out.c (output_function): Renumber PHIs.
11337 * lto-streamer-in.c (input_function): Likewise.
11339 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
11341 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
11342 (get_lane_unsigned): Likewise.
11343 (dup_lane_scalar): Likewise.
11344 (get_lane): enable for VALL.
11345 * config/aarch64/aarch64-simd.md
11346 (aarch64_dup_lane_scalar<mode>): Remove.
11347 (aarch64_get_lane_signed<mode>): Likewise.
11348 (aarch64_get_lane_unsigned<mode>): Likewise.
11349 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
11350 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
11351 (aarch64_get_lane<mode>): Enable for all vector modes.
11352 (aarch64_get_lanedi): Remove misleading constraints.
11353 * config/aarch64/arm_neon.h
11354 (__aarch64_vget_lane_any): Define.
11355 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
11356 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
11357 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
11358 * config/aarch64/iterators.md (VDQQH): New.
11359 (VDQQHS): Likewise.
11360 (vwcore): Likewise.
11362 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
11364 * configure.ac: Add GAS check for LEON instructions on SPARC.
11365 * configure: Regenerate.
11366 * config.in: Likewise.
11367 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
11369 * config/sparc/sparc.opt (LEON, LEON3): New masks.
11370 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
11372 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
11373 (AS_LEON_FLAG): New macro.
11374 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
11375 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
11376 Deal with LEON and LEON3 for the memory model.
11377 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
11378 (atomic_compare_and_swap<mode>_1): Likewise.
11379 (*atomic_compare_and_swap<mode>_1): Likewise.
11381 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11383 * config/arm/neon.md (vcond): Fix floating-point vector
11384 comparisons against 0.
11386 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
11388 * lra-constraints.c (emit_spill_move): Remove assert.
11389 (process_alt_operands): Add more debugging
11390 output. Increase reject for spilling into memory. Decrease
11391 reject for reloading scratch.
11392 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
11394 2013-08-08 Steve Ellcey <sellcey@mips.com>
11396 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
11397 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
11398 micromips incompatible. Add nan2008.
11399 (MULTILIB_DIRNAMES): Add nan2008.
11400 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
11401 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
11402 and micromips incompatible. Add nan2008.
11403 (MULTILIB_DIRNAMES): Add nan2008.
11404 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
11406 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
11408 PR rtl-optimization/58079
11409 * combine.c (combine_simplify_rtx): Avoid using SUBST if
11410 simplify_comparison has widened a comparison with an integer.
11412 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11414 * config/arm/neon.md (movmisalign<mode>): Disable when we
11415 don't allow unaligned accesses.
11416 (*movmisalign<mode>_neon_store): Likewise.
11417 (*movmisalign<mode>_neon_load): Likewise.
11418 (*movmisalign<mode>_neon_store): Likewise.
11419 (*movmisalign<mode>_neon_load): Likewise.
11421 2013-08-08 Jan Hubicka <jh@suse.cz>
11423 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
11424 (make_pass_rebuild_cgraph_edges): Also clear references.
11425 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
11426 * ipa-inline-transform.c (inline_transform): Remove all references
11428 * cgraphunit.c (expand_function): Remove all references after
11430 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
11431 (ipa_find_reference): Rewrite to iterator.
11432 (remove_stmt_references): Likewise.
11433 (ipa_clear_stmts_in_references): New function.
11434 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
11435 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
11437 * ipa-split.c (split_function): Remove references in split function.
11439 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
11442 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
11443 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
11445 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
11448 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
11449 suggested mode for the assignment isn't compatible with the
11450 registers required.
11452 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11455 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
11457 2013-08-07 Xinliang David Li <davidxl@google.com>
11459 * config/i386/i386.opt: New option -mtune-ctrl=.
11460 * config/i386/x86-tune.def: New file.
11461 * config/i386/i386.h: include x86-tune.def.
11462 * config/i386/i386.c (ix86_option_override_internal):
11463 Parsing -mtune-ctrl= option and set tune features.
11465 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
11468 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
11469 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
11472 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
11474 * rtl.h (update_alignments): Declare.
11475 * final.c (grow_label_align): New function extracted from...
11476 (shorten_branches): ...here. Call it.
11477 (update_alignments): New function.
11478 * reorg.c (sibling_labels): New variable.
11479 (get_label_before): Add SIBLING parameter. If it is non-zero, push
11480 the new label along with it onto the sibling_labels vector.
11481 (fill_simple_delay_slots): Adjust call to get_label_before.
11482 (fill_slots_from_thread): Likewise.
11483 (relax_delay_slots): Likewise.
11484 (make_return_insns): Likewise.
11485 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
11487 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
11489 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
11490 document its semantics.
11491 (diagnostic_report_diagnostic): Adjust accordingly.
11493 2013-08-07 David Malcolm <dmalcolm@redhat.com>
11495 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
11496 (sparc_option_override): ...and port to new C++ pass API.
11497 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
11499 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
11501 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
11503 2013-08-06 Caroline Tice <cmtice@google.com>
11505 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
11506 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
11507 * tree-pass.h: Add pass_vtable_verify.
11508 * varasm.c (assemble_variable): Add code to properly set the comdat
11509 section and name for the .vtable_map_vars section.
11510 (assemble_vtyv_preinit_initializer): New function.
11511 (default_sectin_type_flags): Make sure .vtable_map_vars section has
11513 * output.h: Add function decl for assemble_vtv_preinit_initializer.
11514 * vtable-verify.c: New file.
11515 * vtable-verify.h: New file.
11516 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
11517 initialiation levels.
11518 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
11519 * passes.def: Insert pass_vtable_verify.
11520 * aclocal.m4: Reorder includes.
11521 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
11522 -fvtv-counts options.
11523 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
11524 as appropriate, if -fvtable-verify=... is used.
11525 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
11526 -fvtable-verify=... is used.
11527 * Makefile.in (OBJS): Add vtable-verify.o to list.
11528 (vtable-verify.o): Add new build rule.
11529 (GTFILES): Add vtable-verify.c to list.
11530 * common.opt (fvtable-verify=): New flag.
11531 (vtv_priority): Values for fvtable-verify= flag.
11532 (fvtv-counts): New flag.
11533 (fvtv-debug): New flag.
11534 * tree.h (save_vtable_map_decl): New extern function decl.
11536 2013-08-07 David Malcolm <dmalcolm@redhat.com>
11538 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
11539 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
11540 (pass_data_rl78_devirt): ...new pass_data instance and...
11541 (make_pass_rl78_devirt): ...new function.
11542 (rl78_asm_file_start): Port pass registration to new C++ API.
11544 2013-08-07 David Malcolm <dmalcolm@redhat.com>
11546 * coretypes.h (rtl_opt_pass): Add.
11547 (gcc::context): Add.
11548 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
11549 (epiphany_init): Port to new C++ pass API.
11550 (epiphany_optimize_mode_switching): Likewise.
11551 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
11552 (pass_manager::get_pass_mode_switching): New.
11553 (pass_manager::get_pass_peephole2): New.
11554 * mode-switching.c (pass_mode_switching): Add clone method.
11555 * recog.c (pass_peephole2): Add clone method.
11556 (pass_split_all_insns): Add clone method.
11558 2013-08-06 David Malcolm <dmalcolm@redhat.com>
11560 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
11561 (mips_option_override): ...here, porting to new C++ API for passes.
11563 2013-08-06 Jan Hubicka <jh@suse.cz>
11565 * cgraph.c (cgraph_get_body): New function based on lto.c
11567 * cgraph.h (cgraph_get_body): Declare.
11568 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
11570 * cgraphunit.c (expand_function): Get body prior expanding.
11571 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
11572 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
11574 * passes.c (do_per_function_toporder): Get body.
11575 * tree-inline.c (expand_call_inline): Get body prior inlining it.
11576 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
11578 2013-08-06 Martin Jambor <mjambor@suse.cz>
11581 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
11582 is not re-finalized. Rename second parameter to no_collect.
11584 2013-08-06 Martin Jambor <mjambor@suse.cz>
11586 PR middle-end/58041
11587 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
11588 MEM_REF has proper alignment information.
11590 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
11593 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
11595 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
11596 rtx (*) (rtx, ...) with insn_gen_fn.
11597 * genoutput.c (output_insn_data): Cast gen_? function pointers to
11598 insn_gen_fn::stored_funcptr. Add initializer braces.
11600 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11602 Rewrite how instances of passes are cloned to remove assumptions
11603 about their sizes (thus allowing pass subclasses to have
11604 additional data fields, albeit non-GC-managed ones at this point).
11606 * passes.c (make_pass_instance): Now that passes have clone
11607 methods, rewrite this function to eliminate XNEW and memcpy
11608 calls that used hardcoded sizes. Since this function no longer
11609 creates pass instances, rename it to...
11610 (add_pass_instance): ...this. Document the old way that passes were
11611 numbered and flagged, and rework this function to continue using it.
11612 (next_pass_1): Add an initial_pass argument for use by
11614 (position_pass): When adding multiple instances of a pass, use
11615 the pass's clone method, rather than relying on the XNEW/memcpy
11616 within the former make_pass_instance (now add_pass_instance).
11617 (pass_manager::pass_manager): When invoking next_pass_1, also supply
11618 the initial instance of the current pass within the pass manager.
11620 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11622 This is the automated part of the conversion of passes from C
11623 structs to C++ classes.
11625 Patch autogenerated by refactor_passes.py from
11626 https://github.com/davidmalcolm/gcc-refactoring-scripts
11627 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
11629 * asan.c (pass_asan): Convert from a global struct to a subclass of
11630 gimple_opt_pass along with...
11631 (pass_data_asan): ...new pass_data instance and...
11632 (make_pass_asan): ...new function.
11633 (pass_asan_O0): Convert from a global struct to a subclass of
11634 gimple_opt_pass along with...
11635 (pass_data_asan_O0): ...new pass_data instance and...
11636 (make_pass_asan_O0): ...new function.
11637 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
11638 subclass of rtl_opt_pass along with...
11639 (pass_data_inc_dec): ...new pass_data instance and...
11640 (make_pass_inc_dec): ...new function.
11641 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
11642 a subclass of rtl_opt_pass along with...
11643 (pass_data_reorder_blocks): ...new pass_data instance and...
11644 (make_pass_reorder_blocks): ...new function.
11645 (pass_duplicate_computed_gotos): Convert from a global struct to a
11646 subclass of rtl_opt_pass along with...
11647 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
11648 (make_pass_duplicate_computed_gotos): ...new function.
11649 (pass_partition_blocks): Convert from a global struct to a subclass of
11650 rtl_opt_pass along with...
11651 (pass_data_partition_blocks): ...new pass_data instance and...
11652 (make_pass_partition_blocks): ...new function.
11653 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
11654 struct to a subclass of rtl_opt_pass along with...
11655 (pass_data_branch_target_load_optimize1): ...new pass_data instance
11657 (make_pass_branch_target_load_optimize1): ...new function.
11658 (pass_branch_target_load_optimize2): Convert from a global struct to a
11659 subclass of rtl_opt_pass along with...
11660 (pass_data_branch_target_load_optimize2): ...new pass_data instance
11662 (make_pass_branch_target_load_optimize2): ...new function.
11663 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
11664 of rtl_opt_pass along with...
11665 (pass_data_jump): ...new pass_data instance and...
11666 (make_pass_jump): ...new function.
11667 (pass_jump2): Convert from a global struct to a subclass of
11668 rtl_opt_pass along with...
11669 (pass_data_jump2): ...new pass_data instance and...
11670 (make_pass_jump2): ...new function.
11671 * cfgexpand.c (pass_expand): Convert from a global struct to a
11672 subclass of rtl_opt_pass along with...
11673 (pass_data_expand): ...new pass_data instance and...
11674 (make_pass_expand): ...new function.
11675 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
11676 of rtl_opt_pass along with...
11677 (pass_data_free_cfg): ...new pass_data instance and...
11678 (make_pass_free_cfg): ...new function.
11679 (pass_into_cfg_layout_mode): Convert from a global struct to a
11680 subclass of rtl_opt_pass along with...
11681 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
11682 (make_pass_into_cfg_layout_mode): ...new function.
11683 (pass_outof_cfg_layout_mode): Convert from a global struct to a
11684 subclass of rtl_opt_pass along with...
11685 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
11686 (make_pass_outof_cfg_layout_mode): ...new function.
11687 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
11688 struct to a subclass of gimple_opt_pass along with...
11689 (pass_data_build_cgraph_edges): ...new pass_data instance and...
11690 (make_pass_build_cgraph_edges): ...new function.
11691 (pass_rebuild_cgraph_edges): Convert from a global struct to a
11692 subclass of gimple_opt_pass along with...
11693 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
11694 (make_pass_rebuild_cgraph_edges): ...new function.
11695 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
11696 subclass of gimple_opt_pass along with...
11697 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
11699 (make_pass_remove_cgraph_callee_edges): ...new function.
11700 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
11701 struct to a subclass of rtl_opt_pass along with...
11702 (pass_data_stack_adjustments): ...new pass_data instance and...
11703 (make_pass_stack_adjustments): ...new function.
11704 * combine.c (pass_combine): Convert from a global struct to a subclass
11705 of rtl_opt_pass along with...
11706 (pass_data_combine): ...new pass_data instance and...
11707 (make_pass_combine): ...new function.
11708 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
11709 global struct to a subclass of rtl_opt_pass along with...
11710 (pass_data_compare_elim_after_reload): ...new pass_data instance
11712 (make_pass_compare_elim_after_reload): ...new function.
11713 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
11714 of rtl_opt_pass along with...
11715 (pass_data_rtl_cprop): ...new pass_data instance and...
11716 (make_pass_rtl_cprop): ...new function.
11717 * cse.c (pass_cse): Convert from a global struct to a subclass of
11718 rtl_opt_pass along with...
11719 (pass_data_cse): ...new pass_data instance and...
11720 (make_pass_cse): ...new function.
11721 (pass_cse2): Convert from a global struct to a subclass of
11722 rtl_opt_pass along with...
11723 (pass_data_cse2): ...new pass_data instance and...
11724 (make_pass_cse2): ...new function.
11725 (pass_cse_after_global_opts): Convert from a global struct to a
11726 subclass of rtl_opt_pass along with...
11727 (pass_data_cse_after_global_opts): ...new pass_data instance and...
11728 (make_pass_cse_after_global_opts): ...new function.
11729 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
11730 of rtl_opt_pass along with...
11731 (pass_data_ud_rtl_dce): ...new pass_data instance and...
11732 (make_pass_ud_rtl_dce): ...new function.
11733 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
11734 rtl_opt_pass along with...
11735 (pass_data_fast_rtl_dce): ...new pass_data instance and...
11736 (make_pass_fast_rtl_dce): ...new function.
11737 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
11738 a subclass of rtl_opt_pass along with...
11739 (pass_data_df_initialize_opt): ...new pass_data instance and...
11740 (make_pass_df_initialize_opt): ...new function.
11741 (pass_df_initialize_no_opt): Convert from a global struct to a
11742 subclass of rtl_opt_pass along with...
11743 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
11744 (make_pass_df_initialize_no_opt): ...new function.
11745 (pass_df_finish): Convert from a global struct to a subclass of
11746 rtl_opt_pass along with...
11747 (pass_data_df_finish): ...new pass_data instance and...
11748 (make_pass_df_finish): ...new function.
11749 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
11750 rtl_opt_pass along with...
11751 (pass_data_rtl_dse1): ...new pass_data instance and...
11752 (make_pass_rtl_dse1): ...new function.
11753 (pass_rtl_dse2): Convert from a global struct to a subclass of
11754 rtl_opt_pass along with...
11755 (pass_data_rtl_dse2): ...new pass_data instance and...
11756 (make_pass_rtl_dse2): ...new function.
11757 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
11758 subclass of rtl_opt_pass along with...
11759 (pass_data_dwarf2_frame): ...new pass_data instance and...
11760 (make_pass_dwarf2_frame): ...new function.
11761 * except.c (pass_set_nothrow_function_flags): Convert from a global
11762 struct to a subclass of rtl_opt_pass along with...
11763 (pass_data_set_nothrow_function_flags): ...new pass_data instance
11765 (make_pass_set_nothrow_function_flags): ...new function.
11766 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
11767 subclass of rtl_opt_pass along with...
11768 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
11770 (make_pass_convert_to_eh_region_ranges): ...new function.
11771 * final.c (pass_compute_alignments): Convert from a global struct to a
11772 subclass of rtl_opt_pass along with...
11773 (pass_data_compute_alignments): ...new pass_data instance and...
11774 (make_pass_compute_alignments): ...new function.
11775 (pass_final): Convert from a global struct to a subclass of
11776 rtl_opt_pass along with...
11777 (pass_data_final): ...new pass_data instance and...
11778 (make_pass_final): ...new function.
11779 (pass_shorten_branches): Convert from a global struct to a subclass of
11780 rtl_opt_pass along with...
11781 (pass_data_shorten_branches): ...new pass_data instance and...
11782 (make_pass_shorten_branches): ...new function.
11783 (pass_clean_state): Convert from a global struct to a subclass of
11784 rtl_opt_pass along with...
11785 (pass_data_clean_state): ...new pass_data instance and...
11786 (make_pass_clean_state): ...new function.
11787 * function.c (pass_instantiate_virtual_regs): Convert from a global
11788 struct to a subclass of rtl_opt_pass along with...
11789 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
11790 (make_pass_instantiate_virtual_regs): ...new function.
11791 (pass_leaf_regs): Convert from a global struct to a subclass of
11792 rtl_opt_pass along with...
11793 (pass_data_leaf_regs): ...new pass_data instance and...
11794 (make_pass_leaf_regs): ...new function.
11795 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
11796 subclass of rtl_opt_pass along with...
11797 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
11799 (make_pass_thread_prologue_and_epilogue): ...new function.
11800 (pass_match_asm_constraints): Convert from a global struct to a
11801 subclass of rtl_opt_pass along with...
11802 (pass_data_match_asm_constraints): ...new pass_data instance and...
11803 (make_pass_match_asm_constraints): ...new function.
11804 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
11805 subclass of rtl_opt_pass along with...
11806 (pass_data_rtl_fwprop): ...new pass_data instance and...
11807 (make_pass_rtl_fwprop): ...new function.
11808 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
11809 rtl_opt_pass along with...
11810 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
11811 (make_pass_rtl_fwprop_addr): ...new function.
11812 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
11813 rtl_opt_pass along with...
11814 (pass_data_rtl_pre): ...new pass_data instance and...
11815 (make_pass_rtl_pre): ...new function.
11816 (pass_rtl_hoist): Convert from a global struct to a subclass of
11817 rtl_opt_pass along with...
11818 (pass_data_rtl_hoist): ...new pass_data instance and...
11819 (make_pass_rtl_hoist): ...new function.
11820 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
11821 subclass of gimple_opt_pass along with...
11822 (pass_data_lower_cf): ...new pass_data instance and...
11823 (make_pass_lower_cf): ...new function.
11824 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
11825 from a global struct to a subclass of gimple_opt_pass along with...
11826 (pass_data_strength_reduction): ...new pass_data instance and...
11827 (make_pass_strength_reduction): ...new function.
11828 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
11829 of rtl_opt_pass along with...
11830 (pass_data_rtl_ifcvt): ...new pass_data instance and...
11831 (make_pass_rtl_ifcvt): ...new function.
11832 (pass_if_after_combine): Convert from a global struct to a subclass of
11833 rtl_opt_pass along with...
11834 (pass_data_if_after_combine): ...new pass_data instance and...
11835 (make_pass_if_after_combine): ...new function.
11836 (pass_if_after_reload): Convert from a global struct to a subclass of
11837 rtl_opt_pass along with...
11838 (pass_data_if_after_reload): ...new pass_data instance and...
11839 (make_pass_if_after_reload): ...new function.
11840 * init-regs.c (pass_initialize_regs): Convert from a global struct to
11841 a subclass of rtl_opt_pass along with...
11842 (pass_data_initialize_regs): ...new pass_data instance and...
11843 (make_pass_initialize_regs): ...new function.
11844 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
11845 of ipa_opt_pass_d along with...
11846 (pass_data_ipa_cp): ...new pass_data instance and...
11847 (make_pass_ipa_cp): ...new function.
11848 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
11849 global struct to a subclass of gimple_opt_pass along with...
11850 (pass_data_inline_parameters): ...new pass_data instance and...
11851 (make_pass_inline_parameters): ...new function.
11852 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
11853 subclass of gimple_opt_pass along with...
11854 (pass_data_early_inline): ...new pass_data instance and...
11855 (make_pass_early_inline): ...new function.
11856 (pass_ipa_inline): Convert from a global struct to a subclass of
11857 ipa_opt_pass_d along with...
11858 (pass_data_ipa_inline): ...new pass_data instance and...
11859 (make_pass_ipa_inline): ...new function.
11860 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
11861 struct to a subclass of gimple_opt_pass along with...
11862 (pass_data_local_pure_const): ...new pass_data instance and...
11863 (make_pass_local_pure_const): ...new function.
11864 (pass_ipa_pure_const): Convert from a global struct to a subclass of
11865 ipa_opt_pass_d along with...
11866 (pass_data_ipa_pure_const): ...new pass_data instance and...
11867 (make_pass_ipa_pure_const): ...new function.
11868 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
11869 to a subclass of ipa_opt_pass_d along with...
11870 (pass_data_ipa_reference): ...new pass_data instance and...
11871 (make_pass_ipa_reference): ...new function.
11872 * ipa-split.c (pass_split_functions): Convert from a global struct to
11873 a subclass of gimple_opt_pass along with...
11874 (pass_data_split_functions): ...new pass_data instance and...
11875 (make_pass_split_functions): ...new function.
11876 (pass_feedback_split_functions): Convert from a global struct to a
11877 subclass of gimple_opt_pass along with...
11878 (pass_data_feedback_split_functions): ...new pass_data instance and...
11879 (make_pass_feedback_split_functions): ...new function.
11880 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
11881 global struct to a subclass of simple_ipa_opt_pass along with...
11882 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
11884 (make_pass_ipa_function_and_variable_visibility): ...new function.
11885 (pass_ipa_free_inline_summary): Convert from a global struct to a
11886 subclass of simple_ipa_opt_pass along with...
11887 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
11888 (make_pass_ipa_free_inline_summary): ...new function.
11889 (pass_ipa_whole_program_visibility): Convert from a global struct to a
11890 subclass of ipa_opt_pass_d along with...
11891 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
11893 (make_pass_ipa_whole_program_visibility): ...new function.
11894 (pass_ipa_profile): Convert from a global struct to a subclass of
11895 ipa_opt_pass_d along with...
11896 (pass_data_ipa_profile): ...new pass_data instance and...
11897 (make_pass_ipa_profile): ...new function.
11898 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
11899 ipa_opt_pass_d along with...
11900 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
11901 (make_pass_ipa_cdtor_merge): ...new function.
11902 * ira.c (pass_ira): Convert from a global struct to a subclass of
11903 rtl_opt_pass along with...
11904 (pass_data_ira): ...new pass_data instance and...
11905 (make_pass_ira): ...new function.
11906 (pass_reload): Convert from a global struct to a subclass of
11907 rtl_opt_pass along with...
11908 (pass_data_reload): ...new pass_data instance and...
11909 (make_pass_reload): ...new function.
11910 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
11911 subclass of rtl_opt_pass along with...
11912 (pass_data_cleanup_barriers): ...new pass_data instance and...
11913 (make_pass_cleanup_barriers): ...new function.
11914 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
11915 of rtl_opt_pass along with...
11916 (pass_data_loop2): ...new pass_data instance and...
11917 (make_pass_loop2): ...new function.
11918 (pass_rtl_loop_init): Convert from a global struct to a subclass of
11919 rtl_opt_pass along with...
11920 (pass_data_rtl_loop_init): ...new pass_data instance and...
11921 (make_pass_rtl_loop_init): ...new function.
11922 (pass_rtl_loop_done): Convert from a global struct to a subclass of
11923 rtl_opt_pass along with...
11924 (pass_data_rtl_loop_done): ...new pass_data instance and...
11925 (make_pass_rtl_loop_done): ...new function.
11926 (pass_rtl_move_loop_invariants): Convert from a global struct to a
11927 subclass of rtl_opt_pass along with...
11928 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
11929 (make_pass_rtl_move_loop_invariants): ...new function.
11930 (pass_rtl_unswitch): Convert from a global struct to a subclass of
11931 rtl_opt_pass along with...
11932 (pass_data_rtl_unswitch): ...new pass_data instance and...
11933 (make_pass_rtl_unswitch): ...new function.
11934 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
11935 subclass of rtl_opt_pass along with...
11936 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
11938 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11939 (pass_rtl_doloop): Convert from a global struct to a subclass of
11940 rtl_opt_pass along with...
11941 (pass_data_rtl_doloop): ...new pass_data instance and...
11942 (make_pass_rtl_doloop): ...new function.
11943 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
11944 a subclass of rtl_opt_pass along with...
11945 (pass_data_lower_subreg): ...new pass_data instance and...
11946 (make_pass_lower_subreg): ...new function.
11947 (pass_lower_subreg2): Convert from a global struct to a subclass of
11948 rtl_opt_pass along with...
11949 (pass_data_lower_subreg2): ...new pass_data instance and...
11950 (make_pass_lower_subreg2): ...new function.
11951 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
11952 struct to a subclass of ipa_opt_pass_d along with...
11953 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
11954 (make_pass_ipa_lto_gimple_out): ...new function.
11955 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
11956 of ipa_opt_pass_d along with...
11957 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
11958 (make_pass_ipa_lto_finish_out): ...new function.
11959 * mode-switching.c (pass_mode_switching): Convert from a global struct
11960 to a subclass of rtl_opt_pass along with...
11961 (pass_data_mode_switching): ...new pass_data instance and...
11962 (make_pass_mode_switching): ...new function.
11963 * modulo-sched.c (pass_sms): Convert from a global struct to a
11964 subclass of rtl_opt_pass along with...
11965 (pass_data_sms): ...new pass_data instance and...
11966 (make_pass_sms): ...new function.
11967 * omp-low.c (pass_expand_omp): Convert from a global struct to a
11968 subclass of gimple_opt_pass along with...
11969 (pass_data_expand_omp): ...new pass_data instance and...
11970 (make_pass_expand_omp): ...new function.
11971 (pass_lower_omp): Convert from a global struct to a subclass of
11972 gimple_opt_pass along with...
11973 (pass_data_lower_omp): ...new pass_data instance and...
11974 (make_pass_lower_omp): ...new function.
11975 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
11976 of gimple_opt_pass along with...
11977 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
11978 (make_pass_diagnose_omp_blocks): ...new function.
11979 * passes.c (pass_early_local_passes): Convert from a global struct to
11980 a subclass of simple_ipa_opt_pass along with...
11981 (pass_data_early_local_passes): ...new pass_data instance and...
11982 (make_pass_early_local_passes): ...new function.
11983 (pass_all_early_optimizations): Convert from a global struct to a
11984 subclass of gimple_opt_pass along with...
11985 (pass_data_all_early_optimizations): ...new pass_data instance and...
11986 (make_pass_all_early_optimizations): ...new function.
11987 (pass_all_optimizations): Convert from a global struct to a subclass
11988 of gimple_opt_pass along with...
11989 (pass_data_all_optimizations): ...new pass_data instance and...
11990 (make_pass_all_optimizations): ...new function.
11991 (pass_all_optimizations_g): Convert from a global struct to a subclass
11992 of gimple_opt_pass along with...
11993 (pass_data_all_optimizations_g): ...new pass_data instance and...
11994 (make_pass_all_optimizations_g): ...new function.
11995 (pass_rest_of_compilation): Convert from a global struct to a subclass
11996 of rtl_opt_pass along with...
11997 (pass_data_rest_of_compilation): ...new pass_data instance and...
11998 (make_pass_rest_of_compilation): ...new function.
11999 (pass_postreload): Convert from a global struct to a subclass of
12000 rtl_opt_pass along with...
12001 (pass_data_postreload): ...new pass_data instance and...
12002 (make_pass_postreload): ...new function.
12003 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12004 subclass of rtl_opt_pass along with...
12005 (pass_data_gcse2): ...new pass_data instance and...
12006 (make_pass_gcse2): ...new function.
12007 * postreload.c (pass_postreload_cse): Convert from a global struct to
12008 a subclass of rtl_opt_pass along with...
12009 (pass_data_postreload_cse): ...new pass_data instance and...
12010 (make_pass_postreload_cse): ...new function.
12011 * predict.c (pass_profile): Convert from a global struct to a subclass
12012 of gimple_opt_pass along with...
12013 (pass_data_profile): ...new pass_data instance and...
12014 (make_pass_profile): ...new function.
12015 (pass_strip_predict_hints): Convert from a global struct to a subclass
12016 of gimple_opt_pass along with...
12017 (pass_data_strip_predict_hints): ...new pass_data instance and...
12018 (make_pass_strip_predict_hints): ...new function.
12019 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12020 of rtl_opt_pass along with...
12021 (pass_data_peephole2): ...new pass_data instance and...
12022 (make_pass_peephole2): ...new function.
12023 (pass_split_all_insns): Convert from a global struct to a subclass of
12024 rtl_opt_pass along with...
12025 (pass_data_split_all_insns): ...new pass_data instance and...
12026 (make_pass_split_all_insns): ...new function.
12027 (pass_split_after_reload): Convert from a global struct to a subclass
12028 of rtl_opt_pass along with...
12029 (pass_data_split_after_reload): ...new pass_data instance and...
12030 (make_pass_split_after_reload): ...new function.
12031 (pass_split_before_regstack): Convert from a global struct to a
12032 subclass of rtl_opt_pass along with...
12033 (pass_data_split_before_regstack): ...new pass_data instance and...
12034 (make_pass_split_before_regstack): ...new function.
12035 (pass_split_before_sched2): Convert from a global struct to a subclass
12036 of rtl_opt_pass along with...
12037 (pass_data_split_before_sched2): ...new pass_data instance and...
12038 (make_pass_split_before_sched2): ...new function.
12039 (pass_split_for_shorten_branches): Convert from a global struct to a
12040 subclass of rtl_opt_pass along with...
12041 (pass_data_split_for_shorten_branches): ...new pass_data instance
12043 (make_pass_split_for_shorten_branches): ...new function.
12044 * ree.c (pass_ree): Convert from a global struct to a subclass of
12045 rtl_opt_pass along with...
12046 (pass_data_ree): ...new pass_data instance and...
12047 (make_pass_ree): ...new function.
12048 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12049 subclass of rtl_opt_pass along with...
12050 (pass_data_stack_regs): ...new pass_data instance and...
12051 (make_pass_stack_regs): ...new function.
12052 (pass_stack_regs_run): Convert from a global struct to a subclass of
12053 rtl_opt_pass along with...
12054 (pass_data_stack_regs_run): ...new pass_data instance and...
12055 (make_pass_stack_regs_run): ...new function.
12056 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12057 subclass of rtl_opt_pass along with...
12058 (pass_data_cprop_hardreg): ...new pass_data instance and...
12059 (make_pass_cprop_hardreg): ...new function.
12060 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12061 subclass of rtl_opt_pass along with...
12062 (pass_data_reginfo_init): ...new pass_data instance and...
12063 (make_pass_reginfo_init): ...new function.
12064 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12065 of rtl_opt_pass along with...
12066 (pass_data_regmove): ...new pass_data instance and...
12067 (make_pass_regmove): ...new function.
12068 * regrename.c (pass_regrename): Convert from a global struct to a
12069 subclass of rtl_opt_pass along with...
12070 (pass_data_regrename): ...new pass_data instance and...
12071 (make_pass_regrename): ...new function.
12072 * reorg.c (pass_delay_slots): Convert from a global struct to a
12073 subclass of rtl_opt_pass along with...
12074 (pass_data_delay_slots): ...new pass_data instance and...
12075 (make_pass_delay_slots): ...new function.
12076 (pass_machine_reorg): Convert from a global struct to a subclass of
12077 rtl_opt_pass along with...
12078 (pass_data_machine_reorg): ...new pass_data instance and...
12079 (make_pass_machine_reorg): ...new function.
12080 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12081 of rtl_opt_pass along with...
12082 (pass_data_sched): ...new pass_data instance and...
12083 (make_pass_sched): ...new function.
12084 (pass_sched2): Convert from a global struct to a subclass of
12085 rtl_opt_pass along with...
12086 (pass_data_sched2): ...new pass_data instance and...
12087 (make_pass_sched2): ...new function.
12088 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12089 to a subclass of rtl_opt_pass along with...
12090 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12091 (make_pass_stack_ptr_mod): ...new function.
12092 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12093 to a subclass of rtl_opt_pass along with...
12094 (pass_data_rtl_store_motion): ...new pass_data instance and...
12095 (make_pass_rtl_store_motion): ...new function.
12096 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12097 of gimple_opt_pass along with...
12098 (pass_data_tracer): ...new pass_data instance and...
12099 (make_pass_tracer): ...new function.
12100 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12101 to a subclass of gimple_opt_pass along with...
12102 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12103 (make_pass_diagnose_tm_blocks): ...new function.
12104 (pass_lower_tm): Convert from a global struct to a subclass of
12105 gimple_opt_pass along with...
12106 (pass_data_lower_tm): ...new pass_data instance and...
12107 (make_pass_lower_tm): ...new function.
12108 (pass_tm_init): Convert from a global struct to a subclass of
12109 gimple_opt_pass along with...
12110 (pass_data_tm_init): ...new pass_data instance and...
12111 (make_pass_tm_init): ...new function.
12112 (pass_tm_mark): Convert from a global struct to a subclass of
12113 gimple_opt_pass along with...
12114 (pass_data_tm_mark): ...new pass_data instance and...
12115 (make_pass_tm_mark): ...new function.
12116 (pass_tm_edges): Convert from a global struct to a subclass of
12117 gimple_opt_pass along with...
12118 (pass_data_tm_edges): ...new pass_data instance and...
12119 (make_pass_tm_edges): ...new function.
12120 (pass_tm_memopt): Convert from a global struct to a subclass of
12121 gimple_opt_pass along with...
12122 (pass_data_tm_memopt): ...new pass_data instance and...
12123 (make_pass_tm_memopt): ...new function.
12124 (pass_ipa_tm): Convert from a global struct to a subclass of
12125 simple_ipa_opt_pass along with...
12126 (pass_data_ipa_tm): ...new pass_data instance and...
12127 (make_pass_ipa_tm): ...new function.
12128 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12129 subclass of gimple_opt_pass along with...
12130 (pass_data_call_cdce): ...new pass_data instance and...
12131 (make_pass_call_cdce): ...new function.
12132 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12133 subclass of gimple_opt_pass along with...
12134 (pass_data_build_cfg): ...new pass_data instance and...
12135 (make_pass_build_cfg): ...new function.
12136 (pass_split_crit_edges): Convert from a global struct to a subclass of
12137 gimple_opt_pass along with...
12138 (pass_data_split_crit_edges): ...new pass_data instance and...
12139 (make_pass_split_crit_edges): ...new function.
12140 (pass_warn_function_return): Convert from a global struct to a
12141 subclass of gimple_opt_pass along with...
12142 (pass_data_warn_function_return): ...new pass_data instance and...
12143 (make_pass_warn_function_return): ...new function.
12144 (pass_warn_function_noreturn): Convert from a global struct to a
12145 subclass of gimple_opt_pass along with...
12146 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12147 (make_pass_warn_function_noreturn): ...new function.
12148 (pass_warn_unused_result): Convert from a global struct to a subclass
12149 of gimple_opt_pass along with...
12150 (pass_data_warn_unused_result): ...new pass_data instance and...
12151 (make_pass_warn_unused_result): ...new function.
12152 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12153 a subclass of gimple_opt_pass along with...
12154 (pass_data_merge_phi): ...new pass_data instance and...
12155 (make_pass_merge_phi): ...new function.
12156 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12157 a subclass of gimple_opt_pass along with...
12158 (pass_data_lower_complex): ...new pass_data instance and...
12159 (make_pass_lower_complex): ...new function.
12160 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12161 gimple_opt_pass along with...
12162 (pass_data_lower_complex_O0): ...new pass_data instance and...
12163 (make_pass_lower_complex_O0): ...new function.
12164 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12165 subclass of gimple_opt_pass along with...
12166 (pass_data_lower_eh): ...new pass_data instance and...
12167 (make_pass_lower_eh): ...new function.
12168 (pass_refactor_eh): Convert from a global struct to a subclass of
12169 gimple_opt_pass along with...
12170 (pass_data_refactor_eh): ...new pass_data instance and...
12171 (make_pass_refactor_eh): ...new function.
12172 (pass_lower_resx): Convert from a global struct to a subclass of
12173 gimple_opt_pass along with...
12174 (pass_data_lower_resx): ...new pass_data instance and...
12175 (make_pass_lower_resx): ...new function.
12176 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12177 of gimple_opt_pass along with...
12178 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12179 (make_pass_lower_eh_dispatch): ...new function.
12180 (pass_cleanup_eh): Convert from a global struct to a subclass of
12181 gimple_opt_pass along with...
12182 (pass_data_cleanup_eh): ...new pass_data instance and...
12183 (make_pass_cleanup_eh): ...new function.
12184 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
12185 to a subclass of simple_ipa_opt_pass along with...
12186 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
12187 (make_pass_ipa_lower_emutls): ...new function.
12188 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
12189 a subclass of gimple_opt_pass along with...
12190 (pass_data_if_conversion): ...new pass_data instance and...
12191 (make_pass_if_conversion): ...new function.
12192 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
12193 subclass of gimple_opt_pass along with...
12194 (pass_data_build_ssa): ...new pass_data instance and...
12195 (make_pass_build_ssa): ...new function.
12196 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
12197 global struct to a subclass of gimple_opt_pass along with...
12198 (pass_data_loop_distribution): ...new pass_data instance and...
12199 (make_pass_loop_distribution): ...new function.
12200 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
12201 subclass of gimple_opt_pass along with...
12202 (pass_data_mudflap_1): ...new pass_data instance and...
12203 (make_pass_mudflap_1): ...new function.
12204 (pass_mudflap_2): Convert from a global struct to a subclass of
12205 gimple_opt_pass along with...
12206 (pass_data_mudflap_2): ...new pass_data instance and...
12207 (make_pass_mudflap_2): ...new function.
12208 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
12209 subclass of gimple_opt_pass along with...
12210 (pass_data_mudflap_1): ...new pass_data instance and...
12211 (make_pass_mudflap_1): ...new function.
12212 (pass_mudflap_2): Convert from a global struct to a subclass of
12213 gimple_opt_pass along with...
12214 (pass_data_mudflap_2): ...new pass_data instance and...
12215 (make_pass_mudflap_2): ...new function.
12216 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
12217 gimple_opt_pass along with...
12218 (pass_data_nrv): ...new pass_data instance and...
12219 (make_pass_nrv): ...new function.
12220 (pass_return_slot): Convert from a global struct to a subclass of
12221 gimple_opt_pass along with...
12222 (pass_data_return_slot): ...new pass_data instance and...
12223 (make_pass_return_slot): ...new function.
12224 * tree-object-size.c (pass_object_sizes): Convert from a global struct
12225 to a subclass of gimple_opt_pass along with...
12226 (pass_data_object_sizes): ...new pass_data instance and...
12227 (make_pass_object_sizes): ...new function.
12228 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
12229 global struct to a subclass of gimple_opt_pass along with...
12230 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
12232 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12233 (pass_fixup_cfg): Convert from a global struct to a subclass of
12234 gimple_opt_pass along with...
12235 (pass_data_fixup_cfg): ...new pass_data instance and...
12236 (make_pass_fixup_cfg): ...new function.
12237 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
12238 (make_pass_mudflap_1): ...new function.
12239 (pass_mudflap_2): Replace declaration with that of...
12240 (make_pass_mudflap_2): ...new function.
12241 (pass_asan): Replace declaration with that of...
12242 (make_pass_asan): ...new function.
12243 (pass_asan_O0): Replace declaration with that of...
12244 (make_pass_asan_O0): ...new function.
12245 (pass_tsan): Replace declaration with that of...
12246 (make_pass_tsan): ...new function.
12247 (pass_tsan_O0): Replace declaration with that of...
12248 (make_pass_tsan_O0): ...new function.
12249 (pass_lower_cf): Replace declaration with that of...
12250 (make_pass_lower_cf): ...new function.
12251 (pass_refactor_eh): Replace declaration with that of...
12252 (make_pass_refactor_eh): ...new function.
12253 (pass_lower_eh): Replace declaration with that of...
12254 (make_pass_lower_eh): ...new function.
12255 (pass_lower_eh_dispatch): Replace declaration with that of...
12256 (make_pass_lower_eh_dispatch): ...new function.
12257 (pass_lower_resx): Replace declaration with that of...
12258 (make_pass_lower_resx): ...new function.
12259 (pass_build_cfg): Replace declaration with that of...
12260 (make_pass_build_cfg): ...new function.
12261 (pass_early_tree_profile): Replace declaration with that of...
12262 (make_pass_early_tree_profile): ...new function.
12263 (pass_cleanup_eh): Replace declaration with that of...
12264 (make_pass_cleanup_eh): ...new function.
12265 (pass_sra): Replace declaration with that of...
12266 (make_pass_sra): ...new function.
12267 (pass_sra_early): Replace declaration with that of...
12268 (make_pass_sra_early): ...new function.
12269 (pass_early_ipa_sra): Replace declaration with that of...
12270 (make_pass_early_ipa_sra): ...new function.
12271 (pass_tail_recursion): Replace declaration with that of...
12272 (make_pass_tail_recursion): ...new function.
12273 (pass_tail_calls): Replace declaration with that of...
12274 (make_pass_tail_calls): ...new function.
12275 (pass_tree_loop): Replace declaration with that of...
12276 (make_pass_tree_loop): ...new function.
12277 (pass_tree_loop_init): Replace declaration with that of...
12278 (make_pass_tree_loop_init): ...new function.
12279 (pass_lim): Replace declaration with that of...
12280 (make_pass_lim): ...new function.
12281 (pass_tree_unswitch): Replace declaration with that of...
12282 (make_pass_tree_unswitch): ...new function.
12283 (pass_predcom): Replace declaration with that of...
12284 (make_pass_predcom): ...new function.
12285 (pass_iv_canon): Replace declaration with that of...
12286 (make_pass_iv_canon): ...new function.
12287 (pass_scev_cprop): Replace declaration with that of...
12288 (make_pass_scev_cprop): ...new function.
12289 (pass_empty_loop): Replace declaration with that of...
12290 (make_pass_empty_loop): ...new function.
12291 (pass_record_bounds): Replace declaration with that of...
12292 (make_pass_record_bounds): ...new function.
12293 (pass_graphite): Replace declaration with that of...
12294 (make_pass_graphite): ...new function.
12295 (pass_graphite_transforms): Replace declaration with that of...
12296 (make_pass_graphite_transforms): ...new function.
12297 (pass_if_conversion): Replace declaration with that of...
12298 (make_pass_if_conversion): ...new function.
12299 (pass_loop_distribution): Replace declaration with that of...
12300 (make_pass_loop_distribution): ...new function.
12301 (pass_vectorize): Replace declaration with that of...
12302 (make_pass_vectorize): ...new function.
12303 (pass_slp_vectorize): Replace declaration with that of...
12304 (make_pass_slp_vectorize): ...new function.
12305 (pass_complete_unroll): Replace declaration with that of...
12306 (make_pass_complete_unroll): ...new function.
12307 (pass_complete_unrolli): Replace declaration with that of...
12308 (make_pass_complete_unrolli): ...new function.
12309 (pass_parallelize_loops): Replace declaration with that of...
12310 (make_pass_parallelize_loops): ...new function.
12311 (pass_loop_prefetch): Replace declaration with that of...
12312 (make_pass_loop_prefetch): ...new function.
12313 (pass_iv_optimize): Replace declaration with that of...
12314 (make_pass_iv_optimize): ...new function.
12315 (pass_tree_loop_done): Replace declaration with that of...
12316 (make_pass_tree_loop_done): ...new function.
12317 (pass_ch): Replace declaration with that of...
12318 (make_pass_ch): ...new function.
12319 (pass_ccp): Replace declaration with that of...
12320 (make_pass_ccp): ...new function.
12321 (pass_phi_only_cprop): Replace declaration with that of...
12322 (make_pass_phi_only_cprop): ...new function.
12323 (pass_build_ssa): Replace declaration with that of...
12324 (make_pass_build_ssa): ...new function.
12325 (pass_build_alias): Replace declaration with that of...
12326 (make_pass_build_alias): ...new function.
12327 (pass_build_ealias): Replace declaration with that of...
12328 (make_pass_build_ealias): ...new function.
12329 (pass_dominator): Replace declaration with that of...
12330 (make_pass_dominator): ...new function.
12331 (pass_dce): Replace declaration with that of...
12332 (make_pass_dce): ...new function.
12333 (pass_dce_loop): Replace declaration with that of...
12334 (make_pass_dce_loop): ...new function.
12335 (pass_cd_dce): Replace declaration with that of...
12336 (make_pass_cd_dce): ...new function.
12337 (pass_call_cdce): Replace declaration with that of...
12338 (make_pass_call_cdce): ...new function.
12339 (pass_merge_phi): Replace declaration with that of...
12340 (make_pass_merge_phi): ...new function.
12341 (pass_split_crit_edges): Replace declaration with that of...
12342 (make_pass_split_crit_edges): ...new function.
12343 (pass_pre): Replace declaration with that of...
12344 (make_pass_pre): ...new function.
12345 (pass_profile): Replace declaration with that of...
12346 (make_pass_profile): ...new function.
12347 (pass_strip_predict_hints): Replace declaration with that of...
12348 (make_pass_strip_predict_hints): ...new function.
12349 (pass_lower_complex_O0): Replace declaration with that of...
12350 (make_pass_lower_complex_O0): ...new function.
12351 (pass_lower_complex): Replace declaration with that of...
12352 (make_pass_lower_complex): ...new function.
12353 (pass_lower_vector): Replace declaration with that of...
12354 (make_pass_lower_vector): ...new function.
12355 (pass_lower_vector_ssa): Replace declaration with that of...
12356 (make_pass_lower_vector_ssa): ...new function.
12357 (pass_lower_omp): Replace declaration with that of...
12358 (make_pass_lower_omp): ...new function.
12359 (pass_diagnose_omp_blocks): Replace declaration with that of...
12360 (make_pass_diagnose_omp_blocks): ...new function.
12361 (pass_expand_omp): Replace declaration with that of...
12362 (make_pass_expand_omp): ...new function.
12363 (pass_expand_omp_ssa): Replace declaration with that of...
12364 (make_pass_expand_omp_ssa): ...new function.
12365 (pass_object_sizes): Replace declaration with that of...
12366 (make_pass_object_sizes): ...new function.
12367 (pass_strlen): Replace declaration with that of...
12368 (make_pass_strlen): ...new function.
12369 (pass_fold_builtins): Replace declaration with that of...
12370 (make_pass_fold_builtins): ...new function.
12371 (pass_stdarg): Replace declaration with that of...
12372 (make_pass_stdarg): ...new function.
12373 (pass_early_warn_uninitialized): Replace declaration with that of...
12374 (make_pass_early_warn_uninitialized): ...new function.
12375 (pass_late_warn_uninitialized): Replace declaration with that of...
12376 (make_pass_late_warn_uninitialized): ...new function.
12377 (pass_cse_reciprocals): Replace declaration with that of...
12378 (make_pass_cse_reciprocals): ...new function.
12379 (pass_cse_sincos): Replace declaration with that of...
12380 (make_pass_cse_sincos): ...new function.
12381 (pass_optimize_bswap): Replace declaration with that of...
12382 (make_pass_optimize_bswap): ...new function.
12383 (pass_optimize_widening_mul): Replace declaration with that of...
12384 (make_pass_optimize_widening_mul): ...new function.
12385 (pass_warn_function_return): Replace declaration with that of...
12386 (make_pass_warn_function_return): ...new function.
12387 (pass_warn_function_noreturn): Replace declaration with that of...
12388 (make_pass_warn_function_noreturn): ...new function.
12389 (pass_cselim): Replace declaration with that of...
12390 (make_pass_cselim): ...new function.
12391 (pass_phiopt): Replace declaration with that of...
12392 (make_pass_phiopt): ...new function.
12393 (pass_forwprop): Replace declaration with that of...
12394 (make_pass_forwprop): ...new function.
12395 (pass_phiprop): Replace declaration with that of...
12396 (make_pass_phiprop): ...new function.
12397 (pass_tree_ifcombine): Replace declaration with that of...
12398 (make_pass_tree_ifcombine): ...new function.
12399 (pass_dse): Replace declaration with that of...
12400 (make_pass_dse): ...new function.
12401 (pass_nrv): Replace declaration with that of...
12402 (make_pass_nrv): ...new function.
12403 (pass_rename_ssa_copies): Replace declaration with that of...
12404 (make_pass_rename_ssa_copies): ...new function.
12405 (pass_sink_code): Replace declaration with that of...
12406 (make_pass_sink_code): ...new function.
12407 (pass_fre): Replace declaration with that of...
12408 (make_pass_fre): ...new function.
12409 (pass_check_data_deps): Replace declaration with that of...
12410 (make_pass_check_data_deps): ...new function.
12411 (pass_copy_prop): Replace declaration with that of...
12412 (make_pass_copy_prop): ...new function.
12413 (pass_vrp): Replace declaration with that of...
12414 (make_pass_vrp): ...new function.
12415 (pass_uncprop): Replace declaration with that of...
12416 (make_pass_uncprop): ...new function.
12417 (pass_return_slot): Replace declaration with that of...
12418 (make_pass_return_slot): ...new function.
12419 (pass_reassoc): Replace declaration with that of...
12420 (make_pass_reassoc): ...new function.
12421 (pass_rebuild_cgraph_edges): Replace declaration with that of...
12422 (make_pass_rebuild_cgraph_edges): ...new function.
12423 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
12424 (make_pass_remove_cgraph_callee_edges): ...new function.
12425 (pass_build_cgraph_edges): Replace declaration with that of...
12426 (make_pass_build_cgraph_edges): ...new function.
12427 (pass_local_pure_const): Replace declaration with that of...
12428 (make_pass_local_pure_const): ...new function.
12429 (pass_tracer): Replace declaration with that of...
12430 (make_pass_tracer): ...new function.
12431 (pass_warn_unused_result): Replace declaration with that of...
12432 (make_pass_warn_unused_result): ...new function.
12433 (pass_diagnose_tm_blocks): Replace declaration with that of...
12434 (make_pass_diagnose_tm_blocks): ...new function.
12435 (pass_lower_tm): Replace declaration with that of...
12436 (make_pass_lower_tm): ...new function.
12437 (pass_tm_init): Replace declaration with that of...
12438 (make_pass_tm_init): ...new function.
12439 (pass_tm_mark): Replace declaration with that of...
12440 (make_pass_tm_mark): ...new function.
12441 (pass_tm_memopt): Replace declaration with that of...
12442 (make_pass_tm_memopt): ...new function.
12443 (pass_tm_edges): Replace declaration with that of...
12444 (make_pass_tm_edges): ...new function.
12445 (pass_split_functions): Replace declaration with that of...
12446 (make_pass_split_functions): ...new function.
12447 (pass_feedback_split_functions): Replace declaration with that of...
12448 (make_pass_feedback_split_functions): ...new function.
12449 (pass_strength_reduction): Replace declaration with that of...
12450 (make_pass_strength_reduction): ...new function.
12451 (pass_ipa_lower_emutls): Replace declaration with that of...
12452 (make_pass_ipa_lower_emutls): ...new function.
12453 (pass_ipa_function_and_variable_visibility): Replace declaration with
12455 (make_pass_ipa_function_and_variable_visibility): ...new function.
12456 (pass_ipa_tree_profile): Replace declaration with that of...
12457 (make_pass_ipa_tree_profile): ...new function.
12458 (pass_early_local_passes): Replace declaration with that of...
12459 (make_pass_early_local_passes): ...new function.
12460 (pass_ipa_whole_program_visibility): Replace declaration with that
12462 (make_pass_ipa_whole_program_visibility): ...new function.
12463 (pass_ipa_lto_gimple_out): Replace declaration with that of...
12464 (make_pass_ipa_lto_gimple_out): ...new function.
12465 (pass_ipa_increase_alignment): Replace declaration with that of...
12466 (make_pass_ipa_increase_alignment): ...new function.
12467 (pass_ipa_inline): Replace declaration with that of...
12468 (make_pass_ipa_inline): ...new function.
12469 (pass_ipa_free_lang_data): Replace declaration with that of...
12470 (make_pass_ipa_free_lang_data): ...new function.
12471 (pass_ipa_free_inline_summary): Replace declaration with that of...
12472 (make_pass_ipa_free_inline_summary): ...new function.
12473 (pass_ipa_cp): Replace declaration with that of...
12474 (make_pass_ipa_cp): ...new function.
12475 (pass_ipa_reference): Replace declaration with that of...
12476 (make_pass_ipa_reference): ...new function.
12477 (pass_ipa_pure_const): Replace declaration with that of...
12478 (make_pass_ipa_pure_const): ...new function.
12479 (pass_ipa_pta): Replace declaration with that of...
12480 (make_pass_ipa_pta): ...new function.
12481 (pass_ipa_lto_finish_out): Replace declaration with that of...
12482 (make_pass_ipa_lto_finish_out): ...new function.
12483 (pass_ipa_tm): Replace declaration with that of...
12484 (make_pass_ipa_tm): ...new function.
12485 (pass_ipa_profile): Replace declaration with that of...
12486 (make_pass_ipa_profile): ...new function.
12487 (pass_ipa_cdtor_merge): Replace declaration with that of...
12488 (make_pass_ipa_cdtor_merge): ...new function.
12489 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
12491 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12492 (pass_init_datastructures): Replace declaration with that of...
12493 (make_pass_init_datastructures): ...new function.
12494 (pass_fixup_cfg): Replace declaration with that of...
12495 (make_pass_fixup_cfg): ...new function.
12496 (pass_expand): Replace declaration with that of...
12497 (make_pass_expand): ...new function.
12498 (pass_instantiate_virtual_regs): Replace declaration with that of...
12499 (make_pass_instantiate_virtual_regs): ...new function.
12500 (pass_rtl_fwprop): Replace declaration with that of...
12501 (make_pass_rtl_fwprop): ...new function.
12502 (pass_rtl_fwprop_addr): Replace declaration with that of...
12503 (make_pass_rtl_fwprop_addr): ...new function.
12504 (pass_jump): Replace declaration with that of...
12505 (make_pass_jump): ...new function.
12506 (pass_jump2): Replace declaration with that of...
12507 (make_pass_jump2): ...new function.
12508 (pass_lower_subreg): Replace declaration with that of...
12509 (make_pass_lower_subreg): ...new function.
12510 (pass_cse): Replace declaration with that of...
12511 (make_pass_cse): ...new function.
12512 (pass_fast_rtl_dce): Replace declaration with that of...
12513 (make_pass_fast_rtl_dce): ...new function.
12514 (pass_ud_rtl_dce): Replace declaration with that of...
12515 (make_pass_ud_rtl_dce): ...new function.
12516 (pass_rtl_dce): Replace declaration with that of...
12517 (make_pass_rtl_dce): ...new function.
12518 (pass_rtl_dse1): Replace declaration with that of...
12519 (make_pass_rtl_dse1): ...new function.
12520 (pass_rtl_dse2): Replace declaration with that of...
12521 (make_pass_rtl_dse2): ...new function.
12522 (pass_rtl_dse3): Replace declaration with that of...
12523 (make_pass_rtl_dse3): ...new function.
12524 (pass_rtl_cprop): Replace declaration with that of...
12525 (make_pass_rtl_cprop): ...new function.
12526 (pass_rtl_pre): Replace declaration with that of...
12527 (make_pass_rtl_pre): ...new function.
12528 (pass_rtl_hoist): Replace declaration with that of...
12529 (make_pass_rtl_hoist): ...new function.
12530 (pass_rtl_store_motion): Replace declaration with that of...
12531 (make_pass_rtl_store_motion): ...new function.
12532 (pass_cse_after_global_opts): Replace declaration with that of...
12533 (make_pass_cse_after_global_opts): ...new function.
12534 (pass_rtl_ifcvt): Replace declaration with that of...
12535 (make_pass_rtl_ifcvt): ...new function.
12536 (pass_into_cfg_layout_mode): Replace declaration with that of...
12537 (make_pass_into_cfg_layout_mode): ...new function.
12538 (pass_outof_cfg_layout_mode): Replace declaration with that of...
12539 (make_pass_outof_cfg_layout_mode): ...new function.
12540 (pass_loop2): Replace declaration with that of...
12541 (make_pass_loop2): ...new function.
12542 (pass_rtl_loop_init): Replace declaration with that of...
12543 (make_pass_rtl_loop_init): ...new function.
12544 (pass_rtl_move_loop_invariants): Replace declaration with that of...
12545 (make_pass_rtl_move_loop_invariants): ...new function.
12546 (pass_rtl_unswitch): Replace declaration with that of...
12547 (make_pass_rtl_unswitch): ...new function.
12548 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
12549 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12550 (pass_rtl_doloop): Replace declaration with that of...
12551 (make_pass_rtl_doloop): ...new function.
12552 (pass_rtl_loop_done): Replace declaration with that of...
12553 (make_pass_rtl_loop_done): ...new function.
12554 (pass_web): Replace declaration with that of...
12555 (make_pass_web): ...new function.
12556 (pass_cse2): Replace declaration with that of...
12557 (make_pass_cse2): ...new function.
12558 (pass_df_initialize_opt): Replace declaration with that of...
12559 (make_pass_df_initialize_opt): ...new function.
12560 (pass_df_initialize_no_opt): Replace declaration with that of...
12561 (make_pass_df_initialize_no_opt): ...new function.
12562 (pass_reginfo_init): Replace declaration with that of...
12563 (make_pass_reginfo_init): ...new function.
12564 (pass_inc_dec): Replace declaration with that of...
12565 (make_pass_inc_dec): ...new function.
12566 (pass_stack_ptr_mod): Replace declaration with that of...
12567 (make_pass_stack_ptr_mod): ...new function.
12568 (pass_initialize_regs): Replace declaration with that of...
12569 (make_pass_initialize_regs): ...new function.
12570 (pass_combine): Replace declaration with that of...
12571 (make_pass_combine): ...new function.
12572 (pass_if_after_combine): Replace declaration with that of...
12573 (make_pass_if_after_combine): ...new function.
12574 (pass_ree): Replace declaration with that of...
12575 (make_pass_ree): ...new function.
12576 (pass_partition_blocks): Replace declaration with that of...
12577 (make_pass_partition_blocks): ...new function.
12578 (pass_match_asm_constraints): Replace declaration with that of...
12579 (make_pass_match_asm_constraints): ...new function.
12580 (pass_regmove): Replace declaration with that of...
12581 (make_pass_regmove): ...new function.
12582 (pass_split_all_insns): Replace declaration with that of...
12583 (make_pass_split_all_insns): ...new function.
12584 (pass_fast_rtl_byte_dce): Replace declaration with that of...
12585 (make_pass_fast_rtl_byte_dce): ...new function.
12586 (pass_lower_subreg2): Replace declaration with that of...
12587 (make_pass_lower_subreg2): ...new function.
12588 (pass_mode_switching): Replace declaration with that of...
12589 (make_pass_mode_switching): ...new function.
12590 (pass_sms): Replace declaration with that of...
12591 (make_pass_sms): ...new function.
12592 (pass_sched): Replace declaration with that of...
12593 (make_pass_sched): ...new function.
12594 (pass_ira): Replace declaration with that of...
12595 (make_pass_ira): ...new function.
12596 (pass_reload): Replace declaration with that of...
12597 (make_pass_reload): ...new function.
12598 (pass_clean_state): Replace declaration with that of...
12599 (make_pass_clean_state): ...new function.
12600 (pass_branch_prob): Replace declaration with that of...
12601 (make_pass_branch_prob): ...new function.
12602 (pass_value_profile_transformations): Replace declaration with that
12604 (make_pass_value_profile_transformations): ...new function.
12605 (pass_postreload_cse): Replace declaration with that of...
12606 (make_pass_postreload_cse): ...new function.
12607 (pass_gcse2): Replace declaration with that of...
12608 (make_pass_gcse2): ...new function.
12609 (pass_split_after_reload): Replace declaration with that of...
12610 (make_pass_split_after_reload): ...new function.
12611 (pass_branch_target_load_optimize1): Replace declaration with that
12613 (make_pass_branch_target_load_optimize1): ...new function.
12614 (pass_thread_prologue_and_epilogue): Replace declaration with that
12616 (make_pass_thread_prologue_and_epilogue): ...new function.
12617 (pass_stack_adjustments): Replace declaration with that of...
12618 (make_pass_stack_adjustments): ...new function.
12619 (pass_peephole2): Replace declaration with that of...
12620 (make_pass_peephole2): ...new function.
12621 (pass_if_after_reload): Replace declaration with that of...
12622 (make_pass_if_after_reload): ...new function.
12623 (pass_regrename): Replace declaration with that of...
12624 (make_pass_regrename): ...new function.
12625 (pass_cprop_hardreg): Replace declaration with that of...
12626 (make_pass_cprop_hardreg): ...new function.
12627 (pass_reorder_blocks): Replace declaration with that of...
12628 (make_pass_reorder_blocks): ...new function.
12629 (pass_branch_target_load_optimize2): Replace declaration with that
12631 (make_pass_branch_target_load_optimize2): ...new function.
12632 (pass_leaf_regs): Replace declaration with that of...
12633 (make_pass_leaf_regs): ...new function.
12634 (pass_split_before_sched2): Replace declaration with that of...
12635 (make_pass_split_before_sched2): ...new function.
12636 (pass_compare_elim_after_reload): Replace declaration with that of...
12637 (make_pass_compare_elim_after_reload): ...new function.
12638 (pass_sched2): Replace declaration with that of...
12639 (make_pass_sched2): ...new function.
12640 (pass_stack_regs): Replace declaration with that of...
12641 (make_pass_stack_regs): ...new function.
12642 (pass_stack_regs_run): Replace declaration with that of...
12643 (make_pass_stack_regs_run): ...new function.
12644 (pass_df_finish): Replace declaration with that of...
12645 (make_pass_df_finish): ...new function.
12646 (pass_compute_alignments): Replace declaration with that of...
12647 (make_pass_compute_alignments): ...new function.
12648 (pass_duplicate_computed_gotos): Replace declaration with that of...
12649 (make_pass_duplicate_computed_gotos): ...new function.
12650 (pass_variable_tracking): Replace declaration with that of...
12651 (make_pass_variable_tracking): ...new function.
12652 (pass_free_cfg): Replace declaration with that of...
12653 (make_pass_free_cfg): ...new function.
12654 (pass_machine_reorg): Replace declaration with that of...
12655 (make_pass_machine_reorg): ...new function.
12656 (pass_cleanup_barriers): Replace declaration with that of...
12657 (make_pass_cleanup_barriers): ...new function.
12658 (pass_delay_slots): Replace declaration with that of...
12659 (make_pass_delay_slots): ...new function.
12660 (pass_split_for_shorten_branches): Replace declaration with that of...
12661 (make_pass_split_for_shorten_branches): ...new function.
12662 (pass_split_before_regstack): Replace declaration with that of...
12663 (make_pass_split_before_regstack): ...new function.
12664 (pass_convert_to_eh_region_ranges): Replace declaration with that
12666 (make_pass_convert_to_eh_region_ranges): ...new function.
12667 (pass_shorten_branches): Replace declaration with that of...
12668 (make_pass_shorten_branches): ...new function.
12669 (pass_set_nothrow_function_flags): Replace declaration with that of...
12670 (make_pass_set_nothrow_function_flags): ...new function.
12671 (pass_dwarf2_frame): Replace declaration with that of...
12672 (make_pass_dwarf2_frame): ...new function.
12673 (pass_final): Replace declaration with that of...
12674 (make_pass_final): ...new function.
12675 (pass_rtl_seqabstr): Replace declaration with that of...
12676 (make_pass_rtl_seqabstr): ...new function.
12677 (pass_release_ssa_names): Replace declaration with that of...
12678 (make_pass_release_ssa_names): ...new function.
12679 (pass_early_inline): Replace declaration with that of...
12680 (make_pass_early_inline): ...new function.
12681 (pass_inline_parameters): Replace declaration with that of...
12682 (make_pass_inline_parameters): ...new function.
12683 (pass_update_address_taken): Replace declaration with that of...
12684 (make_pass_update_address_taken): ...new function.
12685 (pass_convert_switch): Replace declaration with that of...
12686 (make_pass_convert_switch): ...new function.
12687 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
12688 to a subclass of simple_ipa_opt_pass along with...
12689 (pass_data_ipa_tree_profile): ...new pass_data instance and...
12690 (make_pass_ipa_tree_profile): ...new function.
12691 * tree-sra.c (pass_sra_early): Convert from a global struct to a
12692 subclass of gimple_opt_pass along with...
12693 (pass_data_sra_early): ...new pass_data instance and...
12694 (make_pass_sra_early): ...new function.
12695 (pass_sra): Convert from a global struct to a subclass of
12696 gimple_opt_pass along with...
12697 (pass_data_sra): ...new pass_data instance and...
12698 (make_pass_sra): ...new function.
12699 (pass_early_ipa_sra): Convert from a global struct to a subclass of
12700 gimple_opt_pass along with...
12701 (pass_data_early_ipa_sra): ...new pass_data instance and...
12702 (make_pass_early_ipa_sra): ...new function.
12703 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
12704 subclass of gimple_opt_pass along with...
12705 (pass_data_ccp): ...new pass_data instance and...
12706 (make_pass_ccp): ...new function.
12707 (pass_fold_builtins): Convert from a global struct to a subclass of
12708 gimple_opt_pass along with...
12709 (pass_data_fold_builtins): ...new pass_data instance and...
12710 (make_pass_fold_builtins): ...new function.
12711 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
12712 subclass of gimple_opt_pass along with...
12713 (pass_data_copy_prop): ...new pass_data instance and...
12714 (make_pass_copy_prop): ...new function.
12715 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
12716 global struct to a subclass of gimple_opt_pass along with...
12717 (pass_data_rename_ssa_copies): ...new pass_data instance and...
12718 (make_pass_rename_ssa_copies): ...new function.
12719 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
12720 subclass of gimple_opt_pass along with...
12721 (pass_data_dce): ...new pass_data instance and...
12722 (make_pass_dce): ...new function.
12723 (pass_dce_loop): Convert from a global struct to a subclass of
12724 gimple_opt_pass along with...
12725 (pass_data_dce_loop): ...new pass_data instance and...
12726 (make_pass_dce_loop): ...new function.
12727 (pass_cd_dce): Convert from a global struct to a subclass of
12728 gimple_opt_pass along with...
12729 (pass_data_cd_dce): ...new pass_data instance and...
12730 (make_pass_cd_dce): ...new function.
12731 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
12732 subclass of gimple_opt_pass along with...
12733 (pass_data_dominator): ...new pass_data instance and...
12734 (make_pass_dominator): ...new function.
12735 (pass_phi_only_cprop): Convert from a global struct to a subclass of
12736 gimple_opt_pass along with...
12737 (pass_data_phi_only_cprop): ...new pass_data instance and...
12738 (make_pass_phi_only_cprop): ...new function.
12739 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
12740 subclass of gimple_opt_pass along with...
12741 (pass_data_dse): ...new pass_data instance and...
12742 (make_pass_dse): ...new function.
12743 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
12744 a subclass of gimple_opt_pass along with...
12745 (pass_data_forwprop): ...new pass_data instance and...
12746 (make_pass_forwprop): ...new function.
12747 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
12748 struct to a subclass of gimple_opt_pass along with...
12749 (pass_data_tree_ifcombine): ...new pass_data instance and...
12750 (make_pass_tree_ifcombine): ...new function.
12751 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
12752 subclass of gimple_opt_pass along with...
12753 (pass_data_ch): ...new pass_data instance and...
12754 (make_pass_ch): ...new function.
12755 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
12756 subclass of gimple_opt_pass along with...
12757 (pass_data_tree_loop): ...new pass_data instance and...
12758 (make_pass_tree_loop): ...new function.
12759 (pass_tree_loop_init): Convert from a global struct to a subclass of
12760 gimple_opt_pass along with...
12761 (pass_data_tree_loop_init): ...new pass_data instance and...
12762 (make_pass_tree_loop_init): ...new function.
12763 (pass_lim): Convert from a global struct to a subclass of
12764 gimple_opt_pass along with...
12765 (pass_data_lim): ...new pass_data instance and...
12766 (make_pass_lim): ...new function.
12767 (pass_tree_unswitch): Convert from a global struct to a subclass of
12768 gimple_opt_pass along with...
12769 (pass_data_tree_unswitch): ...new pass_data instance and...
12770 (make_pass_tree_unswitch): ...new function.
12771 (pass_predcom): Convert from a global struct to a subclass of
12772 gimple_opt_pass along with...
12773 (pass_data_predcom): ...new pass_data instance and...
12774 (make_pass_predcom): ...new function.
12775 (pass_vectorize): Convert from a global struct to a subclass of
12776 gimple_opt_pass along with...
12777 (pass_data_vectorize): ...new pass_data instance and...
12778 (make_pass_vectorize): ...new function.
12779 (pass_graphite): Convert from a global struct to a subclass of
12780 gimple_opt_pass along with...
12781 (pass_data_graphite): ...new pass_data instance and...
12782 (make_pass_graphite): ...new function.
12783 (pass_graphite_transforms): Convert from a global struct to a subclass
12784 of gimple_opt_pass along with...
12785 (pass_data_graphite_transforms): ...new pass_data instance and...
12786 (make_pass_graphite_transforms): ...new function.
12787 (pass_check_data_deps): Convert from a global struct to a subclass of
12788 gimple_opt_pass along with...
12789 (pass_data_check_data_deps): ...new pass_data instance and...
12790 (make_pass_check_data_deps): ...new function.
12791 (pass_iv_canon): Convert from a global struct to a subclass of
12792 gimple_opt_pass along with...
12793 (pass_data_iv_canon): ...new pass_data instance and...
12794 (make_pass_iv_canon): ...new function.
12795 (pass_scev_cprop): Convert from a global struct to a subclass of
12796 gimple_opt_pass along with...
12797 (pass_data_scev_cprop): ...new pass_data instance and...
12798 (make_pass_scev_cprop): ...new function.
12799 (pass_record_bounds): Convert from a global struct to a subclass of
12800 gimple_opt_pass along with...
12801 (pass_data_record_bounds): ...new pass_data instance and...
12802 (make_pass_record_bounds): ...new function.
12803 (pass_complete_unroll): Convert from a global struct to a subclass of
12804 gimple_opt_pass along with...
12805 (pass_data_complete_unroll): ...new pass_data instance and...
12806 (make_pass_complete_unroll): ...new function.
12807 (pass_complete_unrolli): Convert from a global struct to a subclass of
12808 gimple_opt_pass along with...
12809 (pass_data_complete_unrolli): ...new pass_data instance and...
12810 (make_pass_complete_unrolli): ...new function.
12811 (pass_parallelize_loops): Convert from a global struct to a subclass
12812 of gimple_opt_pass along with...
12813 (pass_data_parallelize_loops): ...new pass_data instance and...
12814 (make_pass_parallelize_loops): ...new function.
12815 (pass_loop_prefetch): Convert from a global struct to a subclass of
12816 gimple_opt_pass along with...
12817 (pass_data_loop_prefetch): ...new pass_data instance and...
12818 (make_pass_loop_prefetch): ...new function.
12819 (pass_iv_optimize): Convert from a global struct to a subclass of
12820 gimple_opt_pass along with...
12821 (pass_data_iv_optimize): ...new pass_data instance and...
12822 (make_pass_iv_optimize): ...new function.
12823 (pass_tree_loop_done): Convert from a global struct to a subclass of
12824 gimple_opt_pass along with...
12825 (pass_data_tree_loop_done): ...new pass_data instance and...
12826 (make_pass_tree_loop_done): ...new function.
12827 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
12828 struct to a subclass of gimple_opt_pass along with...
12829 (pass_data_cse_reciprocals): ...new pass_data instance and...
12830 (make_pass_cse_reciprocals): ...new function.
12831 (pass_cse_sincos): Convert from a global struct to a subclass of
12832 gimple_opt_pass along with...
12833 (pass_data_cse_sincos): ...new pass_data instance and...
12834 (make_pass_cse_sincos): ...new function.
12835 (pass_optimize_bswap): Convert from a global struct to a subclass of
12836 gimple_opt_pass along with...
12837 (pass_data_optimize_bswap): ...new pass_data instance and...
12838 (make_pass_optimize_bswap): ...new function.
12839 (pass_optimize_widening_mul): Convert from a global struct to a
12840 subclass of gimple_opt_pass along with...
12841 (pass_data_optimize_widening_mul): ...new pass_data instance and...
12842 (make_pass_optimize_widening_mul): ...new function.
12843 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
12844 subclass of gimple_opt_pass along with...
12845 (pass_data_phiopt): ...new pass_data instance and...
12846 (make_pass_phiopt): ...new function.
12847 (pass_cselim): Convert from a global struct to a subclass of
12848 gimple_opt_pass along with...
12849 (pass_data_cselim): ...new pass_data instance and...
12850 (make_pass_cselim): ...new function.
12851 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
12852 subclass of gimple_opt_pass along with...
12853 (pass_data_phiprop): ...new pass_data instance and...
12854 (make_pass_phiprop): ...new function.
12855 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
12856 subclass of gimple_opt_pass along with...
12857 (pass_data_pre): ...new pass_data instance and...
12858 (make_pass_pre): ...new function.
12859 (pass_fre): Convert from a global struct to a subclass of
12860 gimple_opt_pass along with...
12861 (pass_data_fre): ...new pass_data instance and...
12862 (make_pass_fre): ...new function.
12863 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
12864 subclass of gimple_opt_pass along with...
12865 (pass_data_reassoc): ...new pass_data instance and...
12866 (make_pass_reassoc): ...new function.
12867 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
12868 subclass of gimple_opt_pass along with...
12869 (pass_data_sink_code): ...new pass_data instance and...
12870 (make_pass_sink_code): ...new function.
12871 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
12872 subclass of gimple_opt_pass along with...
12873 (pass_data_strlen): ...new pass_data instance and...
12874 (make_pass_strlen): ...new function.
12875 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
12876 struct to a subclass of gimple_opt_pass along with...
12877 (pass_data_build_alias): ...new pass_data instance and...
12878 (make_pass_build_alias): ...new function.
12879 (pass_build_ealias): Convert from a global struct to a subclass of
12880 gimple_opt_pass along with...
12881 (pass_data_build_ealias): ...new pass_data instance and...
12882 (make_pass_build_ealias): ...new function.
12883 (pass_ipa_pta): Convert from a global struct to a subclass of
12884 simple_ipa_opt_pass along with...
12885 (pass_data_ipa_pta): ...new pass_data instance and...
12886 (make_pass_ipa_pta): ...new function.
12887 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
12888 subclass of gimple_opt_pass along with...
12889 (pass_data_uncprop): ...new pass_data instance and...
12890 (make_pass_uncprop): ...new function.
12891 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
12892 global struct to a subclass of gimple_opt_pass along with...
12893 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
12894 (make_pass_late_warn_uninitialized): ...new function.
12895 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
12896 to a subclass of gimple_opt_pass along with...
12897 (pass_data_init_datastructures): ...new pass_data instance and...
12898 (make_pass_init_datastructures): ...new function.
12899 (pass_early_warn_uninitialized): Convert from a global struct to a
12900 subclass of gimple_opt_pass along with...
12901 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
12902 (make_pass_early_warn_uninitialized): ...new function.
12903 (pass_update_address_taken): Convert from a global struct to a
12904 subclass of gimple_opt_pass along with...
12905 (pass_data_update_address_taken): ...new pass_data instance and...
12906 (make_pass_update_address_taken): ...new function.
12907 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
12908 struct to a subclass of gimple_opt_pass along with...
12909 (pass_data_release_ssa_names): ...new pass_data instance and...
12910 (make_pass_release_ssa_names): ...new function.
12911 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
12912 subclass of gimple_opt_pass along with...
12913 (pass_data_stdarg): ...new pass_data instance and...
12914 (make_pass_stdarg): ...new function.
12915 * tree-switch-conversion.c (pass_convert_switch): Convert from a
12916 global struct to a subclass of gimple_opt_pass along with...
12917 (pass_data_convert_switch): ...new pass_data instance and...
12918 (make_pass_convert_switch): ...new function.
12919 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
12920 to a subclass of gimple_opt_pass along with...
12921 (pass_data_tail_recursion): ...new pass_data instance and...
12922 (make_pass_tail_recursion): ...new function.
12923 (pass_tail_calls): Convert from a global struct to a subclass of
12924 gimple_opt_pass along with...
12925 (pass_data_tail_calls): ...new pass_data instance and...
12926 (make_pass_tail_calls): ...new function.
12927 * tree-vect-generic.c (pass_lower_vector): Convert from a global
12928 struct to a subclass of gimple_opt_pass along with...
12929 (pass_data_lower_vector): ...new pass_data instance and...
12930 (make_pass_lower_vector): ...new function.
12931 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
12932 gimple_opt_pass along with...
12933 (pass_data_lower_vector_ssa): ...new pass_data instance and...
12934 (make_pass_lower_vector_ssa): ...new function.
12935 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
12936 to a subclass of gimple_opt_pass along with...
12937 (pass_data_slp_vectorize): ...new pass_data instance and...
12938 (make_pass_slp_vectorize): ...new function.
12939 (pass_ipa_increase_alignment): Convert from a global struct to a
12940 subclass of simple_ipa_opt_pass along with...
12941 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
12942 (make_pass_ipa_increase_alignment): ...new function.
12943 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
12944 gimple_opt_pass along with...
12945 (pass_data_vrp): ...new pass_data instance and...
12946 (make_pass_vrp): ...new function.
12947 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
12948 subclass of simple_ipa_opt_pass along with...
12949 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
12950 (make_pass_ipa_free_lang_data): ...new function.
12951 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
12952 gimple_opt_pass along with...
12953 (pass_data_tsan): ...new pass_data instance and...
12954 (make_pass_tsan): ...new function.
12955 (pass_tsan_O0): Convert from a global struct to a subclass of
12956 gimple_opt_pass along with...
12957 (pass_data_tsan_O0): ...new pass_data instance and...
12958 (make_pass_tsan_O0): ...new function.
12959 * var-tracking.c (pass_variable_tracking): Convert from a global
12960 struct to a subclass of rtl_opt_pass along with...
12961 (pass_data_variable_tracking): ...new pass_data instance and...
12962 (make_pass_variable_tracking): ...new function.
12963 * web.c (pass_web): Convert from a global struct to a subclass of
12964 rtl_opt_pass along with...
12965 (pass_data_web): ...new pass_data instance and...
12966 (make_pass_web): ...new function.
12967 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
12968 declaration with that of...
12969 (make_pass_mode_switch_use): ...new function.
12970 (pass_resolve_sw_modes): Replace declaration with that of...
12971 (make_pass_resolve_sw_modes): ...new function.
12972 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
12973 from a global struct to a subclass of rtl_opt_pass along with...
12974 (pass_data_mode_switch_use): ...new pass_data instance and...
12975 (make_pass_mode_switch_use): ...new function.
12976 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
12977 from a global struct to a subclass of rtl_opt_pass along with...
12978 (pass_data_resolve_sw_modes): ...new pass_data instance and...
12979 (make_pass_resolve_sw_modes): ...new function.
12980 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
12981 struct to a subclass of rtl_opt_pass along with...
12982 (pass_data_insert_vzeroupper): ...new pass_data instance and...
12983 (make_pass_insert_vzeroupper): ...new function.
12984 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
12985 global struct to a subclass of rtl_opt_pass along with...
12986 (pass_data_work_around_errata): ...new pass_data instance and...
12987 (make_pass_work_around_errata): ...new function.
12988 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
12989 struct to a subclass of rtl_opt_pass along with...
12990 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
12991 (make_pass_mips_machine_reorg2): ...new function.
12993 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12995 * passes.c (pass_manager::operator new): New.
12997 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12999 Handwritten part of conversion of passes to C++ classes.
13001 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13002 (toplev.o): Add dep on PASS_MANAGER_H.
13003 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13004 of early local pases to reflect this moving from a global to a
13005 member of gcc::pass_manager.
13006 (cgraph_add_new_function): Likewise.
13007 * lto-cgraph.c (lto_output_node): Update for conversion of
13008 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13009 * passes.c (opt_pass::clone): New.
13010 (opt_pass::gate): New.
13011 (opt_pass::execute): New.
13012 (opt_pass::opt_pass): New.
13013 (pass_manager::execute_early_local_passes): New.
13014 (pass_manager::execute_pass_mode_switching): new.
13015 (finish_optimization_passes): Convert to...
13016 (pass_manager::finish_optimization_passes): ...this.
13017 (finish_optimization_passes): Update for conversion of passes to
13019 (register_dump_files_1): Use has_gate since we cannot portably
13020 check a vtable entry against NULL.
13021 (dump_one_pass): Likewise.
13022 (ipa_write_summaries_2): Likewise.
13023 (ipa_write_optimization_summaries_1): Likewise.
13024 (ipa_read_summaries_1): Likewise.
13025 (ipa_read_optimization_summaries_1): Likewise.
13026 (execute_ipa_stmt_fixups): Likewise.
13027 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13028 pass-creation functions rather than wiring up globals, and
13029 storing the results in fields of pass_manager generated using
13030 pass-instances.def.
13031 (pass_manager::dump_profile_report): Update for conversion of
13032 passes to C++ classes.
13033 (pass_manager::execute_ipa_summary_passes): Likewise.
13034 (execute_one_ipa_transform_pass): Likewise.
13035 (execute_one_pass): Use has_gate and has_execute since we cannot
13036 portably check a vtable entry against NULL.
13037 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13038 (pass_manager): Use pass-instances.def to add fields for the
13039 various pass instances.
13040 * toplev.c (finalize): Update for move of
13041 finish_optimization_passes to a method of gcc::pass_manager.
13042 * toplev.h (finish_optimization_passes): Move to method of class
13044 * tree-pass.h (struct pass_data): New.
13045 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13046 (opt_pass::gate): Convert to virtual function.
13047 (opt_pass::~opt_pass): New.
13048 (opt_pass::clone): New.
13049 (opt_pass::execute): Convert to virtual function.
13050 (opt_pass::opt_pass): New.
13051 (opt_pass::ctxt_): new.
13052 (gimple_opt_pass): Convert to subclass of opt_pass.
13053 (gimple_opt_pass::gimple_opt_pass): New.
13054 (rtl_opt_pass): Convert to subclass of opt_pass.
13055 (rtl_opt_pass::rtl_opt_pass): New.
13056 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13057 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13058 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13059 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13060 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13061 invocation of pass_mode_switching to reflect this moving from a
13062 global to a member of gcc::pass_manager.
13063 (ix86_option_override): Rework how pass_insert_vzeroupper is
13064 added to the pass_manager to reflect autogenerated changes.
13065 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13067 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13069 PR rtl-optimization/57708
13070 * recog.c (peep2_find_free_register): Validate all regs in a
13073 2013-08-05 Jan Hubicka <jh@suse.cz>
13076 * cgraph.c (verify_cgraph_node): Accept local flags from other
13078 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13079 (function_and_variable_visibility): Likewise.
13080 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13082 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13084 * graph.c (init_graph_slim_pretty_print): Remove.
13085 (print_graph_cfg): Do not call it. Use local pretty printer.
13086 (start_graph_dump): Likewise.
13088 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13090 * gimple-pretty-print.c (buffer): Remove.
13091 (initialized): Likewise.
13092 (maybe_init_pretty_print): Likewise.
13093 (print_gimple_stmt): Do not call it. Use non-static local
13094 pretty_printer variable.
13095 (print_gimple_expr): Likewise.
13096 (print_gimple_seq): Likewise.
13097 (gimple_dump_bb): Likewise.
13099 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13101 * asan.c (asan_pp): Remove.
13102 (asan_pp_initialized): Likewise.
13103 (asan_pp_initialize): Likewise.
13104 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13105 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13106 (asan_add_global): Likewise.
13108 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13110 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13111 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13112 * pretty-print.c (pp_formatted_text_data): Likewise.
13113 (pp_indent): Rename from pp_base_indent.
13114 (pp_format): Rename from pp_base_format.
13115 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13116 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13117 (pp_flush): Rename from pp_base_flush.
13118 (pp_set_line_maximum_length): Rename from
13119 pp_base_set_line_maximum_length.
13120 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13121 (pp_set_prefix): Rename from pp_base_set_prefix.
13122 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13123 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13124 (pp_append_text): Rename from pp_base_append_text.
13125 (pp_formatted_text): Rename from pp_base_formatted_text.
13126 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13127 (pp_remaining_character_count_for_line): Rename from
13128 pp_base_remaining_character_count_for_line.
13129 (pp_newline): Rename from pp_base_newline.
13130 (pp_character): Rename from pp_base_character.
13131 (pp_string): Rename from pp_base_string.
13132 (pp_maybe_space): Rename from pp_base_maybe_space.
13133 * asan.c (asan_pp_string): Adjust.
13134 (asan_emit_stack_protection): Likewise.
13135 (asan_add_global): Likewise.
13136 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13137 * tree-mudflap.c (mf_varname_tree): Likewise.
13138 * tree-pretty-print.c (pp_tree_identifier): Rename from
13139 pp_base_tree_identifier.
13140 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13141 Declare as function.
13143 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13145 * pretty-print.h (pp_bar_bar): New.
13146 (pp_ampersand_ampersand): Likewise.
13147 (pp_less_equal): Likewise.
13148 (pp_greater_equal): Likewise.
13149 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13150 printer functions instead of pp_string or operators and punctuators.
13151 (dump_gimple_call): Likewise.
13152 (dump_gimple_omp_for): Likewise.
13153 (dump_gimple_transaction): Likewise.
13154 (dump_gimple_phi): Likewise.
13155 (pp_gimple_stmt_1): Likewise.
13156 * sched-vis.c (print_insn): Likewise.
13157 * tree-mudflap.c (mf_varname_tree): Likewise.
13158 * tree-pretty-print.c (dump_block_node): Likewise.
13159 (dump_generic_node): Likewise.
13161 2013-08-02 Jan Hubicka <jh@suse.cz>
13163 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13165 * lto-streamer-out.c (tree_is_indexable): Results decls and
13166 parm decls are not indexable.
13167 (DFS_write_tree_body): Do not follow args and results.
13168 (hash_tree): Likewise.
13169 (output_functions): Rearrange so struct function is needed
13170 only when real body is output; be able to also ouptut abstract
13171 functions; output DECL_ARGUMENTS and DECL_RESULT.
13172 (lto_output): When not in WPA, ale store abstract functions.
13173 (write_symbol): Do not care about RESULT_DECL.
13174 (output_symbol_p): Handle correctly sbtract decls.
13175 * lto-streamer-in.c (input_function): Rearrange so struct
13176 function can be NULL at entry; allow streaming of
13177 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13178 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13179 sanity check during LTO.
13180 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13181 RESULT_DECl and DECL_ARGUMENTS.
13182 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
13185 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13187 * pretty-print.h (pp_underscore): New.
13189 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
13190 printer functions instead of pp_character.
13191 (dump_binary_rhs): Likewise.
13192 (dump_ternary_rhs): Likewise.
13193 (dump_gimple_call_args): Likewise.
13194 (pp_points_to_solution): Likewise.
13195 (dump_gimple_call): Likewise.
13196 (dump_gimple_switch): Likewise.
13197 (dump_gimple_cond): Likewise.
13198 (dump_gimple_bind): Likewise.
13199 (dump_gimple_try): Likewise.
13200 (dump_gimple_omp_for): Likewise.
13201 (dump_gimple_omp_continue): Likewise.
13202 (dump_gimple_omp_single): Likewise.
13203 (dump_gimple_omp_sections): Likewise.
13204 (dump_gimple_omp_block): Likewise.
13205 (dump_gimple_omp_critical): Likewise.
13206 (dump_gimple_transaction): Likewise.
13207 (dump_gimple_asm): Likewise.
13208 (dump_gimple_phi): Likewise.
13209 (dump_gimple_omp_parallel): Likewise.
13210 (dump_gimple_omp_task): Likewise.
13211 (dump_gimple_omp_atomic_load): Likewise.
13212 (dump_gimple_omp_atomic_store): Likewise.
13213 (dump_gimple_mem_ops): Likewise.
13214 (pp_gimple_stmt_1): Likewise.
13215 (pp_cfg_jump): Likewise.
13216 (dump_implicit_edges): Likewise.
13217 (gimple_dump_bb_for_graph): Likewise.
13218 * graph.c (draw_cfg_node): Likewise.
13219 * langhooks.c (lhd_print_error_function): Likewise.
13220 * sched-vis.c (print_exp): Likewise.
13221 (print_value): Likewise.
13222 (print_pattern): Likewise.
13223 (print_insn): Likewise.
13224 (rtl_dump_bb_for_graph): Likewise.
13225 * tree-pretty-print.c (dump_function_declaration): Likewise.
13226 (dump_array_domain): Likewise.
13227 (dump_omp_clause): Likewise.
13228 (dump_location): Likewise.
13229 (dump_generic_node): Likewise.
13230 (print_struct_decl): Likewise.
13231 * diagnostic.c (diagnostic_show_locus): Use pp_space.
13233 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13235 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
13236 candidate table when replacing a candidate statement.
13237 (replace_rhs_if_not_dup): Likewise.
13238 (replace_one_candidate): Likewise.
13240 2013-08-02 Jan Hubicka <jh@suse.cz>
13241 Martin Liska <marxin.liska@gmail.com>
13243 * cgraphunit.c (add_new_function): Fix logic when adding from
13245 (assemble_thunk): Rename to ...
13246 (expand_thunk); .. this one; export; get it working with
13247 general functions; make produced gimple valid.
13248 * cgraph.h (expand_thunk): Declare.
13250 2013-08-02 Jan Hubicka <jh@suse.cz>
13252 * ipa-cp.c (gather_context_independent_values): Use
13253 ipa_get_param_move_cost.
13254 (get_replacement_map): Remove PARAM; move parameter folding
13256 (create_specialized_node): Update.
13257 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
13258 assert that we have gimple body; update move_cost.
13259 (count_formal_params): Assert that we have gimple body.
13260 (ipa_dump_param): New function.
13261 (ipa_alloc_node_params): Break out from ...
13262 (ipa_initialize_node_params): ... here.
13263 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
13264 (ipa_write_node_info): Stream move costs.
13265 (ipa_read_node_info): Read move costs.
13266 (ipa_update_after_lto_read): Do not recompute node params.
13267 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
13268 (ipa_get_param): Check we are not in WPA.
13269 (ipa_get_param_move_cost): New.
13270 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
13271 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
13272 parm numbers to be present.
13274 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13276 PR rtl-optimization/58048
13277 * lra-constraints.c (process_alt_operands): Don't check asm
13278 operand on register.
13280 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
13282 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
13283 the implied StoreLoad barrier for atomic operations if before.
13285 2013-08-02 Jan Hubicka <jh@suse.cz>
13286 Martin Liska <marxin.liska@gmail.com>
13288 * cgraph.c (cgraph_function_body_availability): Do not check
13290 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
13291 symtab_node_availability): Declare.
13292 * ipa.c (can_replace_by_local_alias): New.
13293 (function_and_variable_visibility): Use it.
13294 * symtab.c (symtab_for_node_and_aliases,
13295 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
13297 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13299 PR rtl-optimization/57963
13300 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
13301 (lra_constraints): Use them.
13303 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
13305 * config/arm/types.md (define_attr "type"): Add "load_acq"
13307 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
13309 (cortex_a53_store1): Likewise.
13311 2013-08-01 Jan Hubicka <jh@suse.cz>
13313 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
13314 partitions are not needed.
13316 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
13318 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
13319 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
13320 MAYBE_NON_Q_CLASS_P where appropriate.
13322 2013-08-01 Jan Hubicka <jh@suse.cz>
13324 * cgraph.h (release_function_body): Declare.
13325 * tree.c (free_lang_data_in_decl): Free, parameters and return values
13326 of unused delcarations.
13328 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13330 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
13331 RTL form when subtracting a constant.
13333 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13335 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
13336 Generate canonical plus rtx with negated immediate instead of minus
13338 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
13340 2013-08-01 Jan Hubicka <jh@suse.cz>
13342 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
13343 (cgraph_release_function_body): Likewise.
13344 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
13345 * cgraph.h (cgrpah_node): Rename abstract_and_needed
13346 to used_as_abstract_origin.
13347 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
13348 symbols used as abstract origins.
13349 * cgraphunit.c (analyze_functions): Update.
13350 * ipa.c (symtab_remove_unreachable_nodes): Recompute
13351 used_as_abstract_origin.
13352 * tree-inline.c (tree_function_versioning): Update
13353 used_as_abstract_origin; be ready for DECL_RESULT and
13354 DECL_ARGUMENTS to be NULL.
13356 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
13357 for abstract functions.
13358 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
13361 2013-08-01 Jan Hubicka <jh@suse.cz>
13363 * profile.c (compute_value_histograms): Fix thinko.
13365 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13367 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
13368 aarch-common-protos.h to extra_headers.
13369 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
13370 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
13371 * config/aarch64/t-aarch64 (aarch-common.o): Define.
13373 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13375 * config/aarch64/aarch64.md (define_attr "type"): Delete.
13376 Include "../arm/types.md". Define "type" attribute for all patterns.
13377 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
13380 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
13382 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
13383 to support power8 load fusion.
13384 (fusion_gpr_mem_load): Likewise.
13386 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
13388 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
13389 declarations for power8 load fusion.
13390 (emit_fusion_gpr_load): Likewise.
13392 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
13393 tuning for power8, turn on fusion mode by default. Turn on sign
13394 extending fusion mode if normal fusion mode is on, and we are at
13396 (fusion_gpr_load_p): New function, return true if we can fuse an
13397 addis instruction with a dependent load to a GPR.
13398 (emit_fusion_gpr_load): Emit the instructions for power8 load
13401 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
13402 (VSX load fusion peepholes): New peepholes to fuse together an
13403 addi instruction with a VSX load instruction.
13405 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
13406 peepholes to fuse an addis instruction with a load to a GPR base
13407 register. If we are supporting sign extending fusions, convert
13408 sign extending loads to zero extending loads and add an explicit
13411 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
13413 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
13414 aarch-common-protos.h to extra_headers.
13415 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
13416 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
13417 (arm_early_store_addr_dep): Likewise.
13418 (arm_no_early_alu_shift_dep): Likewise.
13419 (arm_no_early_alu_shift_value_dep): Likewise.
13420 (arm_no_early_mul_dep): Likewise.
13421 (arm_no_early_store_addr_dep): Likewise.
13422 (arm_mac_accumulator_is_mul_result): Likewise.
13423 (arm_mac_accumulator_is_result): Likewise.
13424 * config/arm/aarch-common.c: ... here. New file.
13425 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
13427 (arm_early_store_addr_dep): Likewise.
13428 (arm_no_early_alu_shift_dep): Likewise.
13429 (arm_no_early_alu_shift_value_dep): Likewise.
13430 (arm_no_early_mul_dep): Likewise.
13431 (arm_no_early_store_addr_dep): Likewise.
13432 (arm_mac_accumulator_is_mul_result): Likewise.
13433 (arm_mac_accumulator_is_result): Likewise.
13434 * config/arm/aarch-common-protos.h: ... here. New file.
13435 * config/arm/t-arm (aarch-common.o): Define.
13437 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
13439 * config/arm/arm.md: Include new file "types.md".
13440 (define_attr "type"): Move from here to ...
13441 (define_attr "mul32"): Likewise.
13442 (define_attr "mul64"): Likewise.
13443 * config/arm/types.md: ... here. New file.
13445 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
13447 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
13448 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
13450 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13452 * gen-pass-instances.awk: Fix offset of substr().
13454 2013-07-31 David Malcolm <dmalcolm@redhat.com>
13456 * Makefile.in (pass-instances.def): New.
13457 (passes.o): Replace dependency on passes.def with one on
13460 * gen-pass-instances.awk: New.
13462 * passes.c (pass_manager::pass_manager): Use pass-instances.def
13463 rather than passes.def, updating local definition of NEXT_PASS
13464 macro to add an extra NUM parameter (currently unused).
13466 2013-07-30 David Malcolm <dmalcolm@redhat.com>
13468 * Makefile.in (PASS_MANAGER_H): New.
13469 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
13470 (passes.o): Likewise.
13471 (statistics.o): Likewise.
13472 (cgraphunit.o): Likewise.
13473 (context.o): Depend on PASS_MANAGER_H.
13475 * pass_manager.h: New.
13477 * cgraphunit.c (cgraph_add_new_function): Update for moves
13478 of globals to fields of pass_manager.
13479 (analyze_function): Likewise.
13480 (expand_function): Likewise.
13481 (ipa_passes): Likewise.
13482 (compile): Likewise.
13484 * context.c (context::context): New.
13485 * context.h (context::context): New.
13486 (context::get_passes): New.
13487 (context::passes_): New.
13489 * lto-cgraph.c (input_node): Update for moves of globals to
13490 fields of pass_manager.
13492 * passes.c (all_passes): Remove, in favor of a field of the
13493 same name within the new class pass_manager.
13494 (all_small_ipa_passes): Likewise.
13495 (all_lowering_passes): Likewise.
13496 (all_regular_ipa_passes): Likewise.
13497 (all_late_ipa_passes): Likewise.
13498 (all_lto_gen_passes): Likewise.
13499 (passes_by_id): Likewise.
13500 (passes_by_id_size): Likewise.
13501 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
13502 the new class pass_manager.
13503 (set_pass_for_id): Convert to...
13504 (pass_manager::set_pass_for_id): ...method.
13505 (get_pass_for_id): Convert to...
13506 (pass_manager::get_pass_for_id): ...method.
13507 (register_one_dump_file): Move body of implementation into...
13508 (pass_manager::register_one_dump_file): ...here.
13509 (register_dump_files_1): Convert to...
13510 (pass_manager::register_dump_files_1): ...method.
13511 (register_dump_files): Convert to...
13512 (pass_manager::register_dump_files): ...method.
13513 (create_pass_tab): Update for moves of globals to fields of
13515 (dump_passes): Move body of implementation into...
13516 (pass_manager::dump_passes): ...here.
13517 (register_pass): Move body of implementation into...
13518 (pass_manager::register_pass): ...here.
13519 (init_optimization_passes): Convert into...
13520 (pass_manager::pass_manager): ...constructor for new
13521 pass_manager class, and initialize the pass_lists array.
13522 (check_profile_consistency): Update for moves of globals to
13523 fields of pass_manager.
13524 (dump_profile_report): Move body of implementation into...
13525 (pass_manager::dump_profile_report): ...here.
13526 (ipa_write_summaries_1): Update for moves of pass lists from
13527 being globals to fields of pass_manager.
13528 (ipa_write_optimization_summaries): Likewise.
13529 (ipa_read_summaries): Likewise.
13530 (ipa_read_optimization_summaries): Likewise.
13531 (execute_all_ipa_stmt_fixups): Likewise.
13533 * statistics.c (statistics_fini): Update for moves of globals to
13534 fields of pass_manager.
13536 * toplev.c (general_init): Replace call to
13537 init_optimization_passes with construction of the pass_manager
13540 * tree-pass.h (all_passes): Remove, in favor of a field of the
13541 same name within the new class pass_manager.
13542 (all_small_ipa_passes): Likewise.
13543 (all_lowering_passes): Likewise.
13544 (all_regular_ipa_passes): Likewise.
13545 (all_lto_gen_passes): Likewise.
13546 (all_late_ipa_passes): Likewise.
13547 (passes_by_id): Likewise.
13548 (passes_by_id_size): Likewise.
13549 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
13550 the new class pass_manager.
13551 (get_pass_for_id): Remove.
13553 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
13555 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
13558 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
13560 * arm.md (mulhi3): New expand pattern.
13562 2013-07-30 Jan Hubicka <jh@suse.cz>
13563 Martin Liska <marxin.liska@gmail.com>
13565 * profile.c (compute_value_histograms): Do not ICE when
13566 there is mismatch only on some counters.
13568 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13570 PR rtl-optimization/57637
13571 * function.c (move_insn_for_shrink_wrap): Also check the
13572 GEN set of the LIVE problem for the liveness analysis
13573 if it exists, otherwise give up.
13575 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13577 PR tree-optimization/57993
13578 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
13579 replaced statement in the candidate table.
13580 (phi_add_costs): Return infinite cost when the hidden basis does
13581 not dominate all phis on which the candidate is dependent.
13582 (replace_one_candidate): Record replaced statement in the
13585 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
13587 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
13588 (ashlv2si3): New expander.
13589 (*ashlv2si3_i): New define_insn_and_split.
13590 * predicates.md (float_operation): Allow patterns with three
13591 basic sub-patterns.
13593 PR rtl-optimization/58021
13594 * mode-switching.c (create_pre_exit): Always split off preceding
13595 insns if we are not at the basic block head.
13597 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
13599 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
13600 (UCLIBC_DYNAMIC_LINKER): New macro.
13601 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
13603 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
13604 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
13606 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
13607 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
13608 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
13609 for SF and DF modes. Use ieee_quad_format for TF mode.
13610 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
13611 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
13612 (mips_option_override): Handle `-mnan=legacy'.
13613 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
13614 `-mabs=2008' and `-mnan=2008'.
13615 (OPTION_DEFAULT_SPECS): Add "nan" default.
13616 (ASM_SPEC): Handle `-mnan='.
13617 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
13618 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
13619 comment accordingly.
13620 (neg<mode>2): Likewise.
13621 * config/mips/mips.opt (mabs, mnan): New options.
13622 * doc/install.texi (Configuration): Document `--with-nan=' option.
13623 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
13625 (MIPS Options): Document them.
13626 * config.gcc <mips*-*-*>: Handle `--with-nan='.
13627 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
13628 * configure: Regenerate.
13629 * config.in: Regenerate.
13631 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
13633 * config/i386/i386.md (float post-reload splitters): Do not check
13634 for subregs of SSE registers.
13636 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
13637 H.J. Lu <hongjiu.lu@intel.com>
13641 * config/i386/i386.md (post-reload splitter
13642 to avoid partial SSE reg dependency stalls): New pattern.
13644 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
13646 * config/s390/s390.md ("movcc"): Swap load and store instructions.
13648 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
13650 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
13651 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
13653 2013-07-26 Cary Coutant <ccoutant@google.com>
13655 * dwarf2out.c (die_checksum_ordered): Don't include template
13656 instantiations in signature.
13657 (is_template_parameter): New function.
13658 (is_template_instantiation): New function.
13659 (generate_skeleton_bottom_up): Don't include template instantiations
13661 (generate_skeleton): Likewise.
13662 (break_out_comdat_types): Move recursive call to break out nested
13664 (prune_unused_types_mark_generic_parms_dies): Call
13665 is_template_parameter.
13667 2013-07-26 Ian Bolton <ian.bolton@arm.com>
13669 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
13670 uses vector registers.
13671 * config/aarch64/iterators.md: Add attributes rtn and vas.
13673 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13674 Richard Earnshaw <richard.earnshaw@arm.com>
13676 * combine.c (simplify_comparison): Re-canonicalize operands
13678 * config/arm/arm.md (movcond_addsi): New splitter.
13680 2013-07-25 Sterling Augustine <saugustine@google.com>
13682 * dwarf2out.c (size_of_pubnames): Move code to...
13683 (include_pubname_in_output): ...here. New.
13684 (want_pubnames): Rearrange.
13685 (output_pubnames): Call include_pubname_in_output. Move assertion.
13687 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
13689 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
13691 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
13694 * doc/extend.texi: Remove obsolete builtins. Fix
13695 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
13697 2013-07-25 Jan Hubicka <jh@suse.cz>
13699 * cgraph.c (release_function_body): Break out from ...
13700 (cgraph_release_function_body): ... this one; also release DECL_RESULT
13701 and DECL_ARGUMENTS.
13702 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
13703 old_tree in the map.
13704 (create_specialized_node): Update.
13705 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
13707 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
13708 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
13709 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
13710 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
13713 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13715 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
13716 addsi3_carryin_alt2_<optab>): Correct output template.
13718 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13720 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
13721 Adjust for arm_restrict_it.
13722 Remove trailing whitespace.
13724 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
13726 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
13727 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
13729 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
13731 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
13733 PR rtl-optimization/57960
13734 * lra-constraints.c (process_alt_operands): Use the right mode
13735 when checking strict_low.
13737 2013-07-25 Jan Hubicka <jh@suse.cz>
13739 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
13740 * cgraph.c (cgraph_remove_node): Do not release function body
13741 when in cgraph streaming.
13742 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
13743 in other partitions are not considered reachable; fix handling of
13746 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13748 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
13750 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13755 * config/arm/arm.md ("*sibcall_insn): Replace use of
13756 Ss with US. Adjust output for v5 and v4t.
13757 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
13759 * config/arm/constraints.md ("Ss"): Rename to US.
13761 2013-07-25 Terry Guo <terry.guo@arm.com>
13763 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
13764 shift_add/shift_sub0/shift_sub1 RTXs.
13766 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
13767 Anton Blanchard <anton@au1.ibm.com>
13769 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
13770 (altivec_vpks<VI_char>ss): Likewise.
13771 (altivec_vpks<VI_char>us): Likewise.
13772 (altivec_vpku<VI_char>us): Likewise.
13773 (altivec_vpku<VI_char>um): Likewise.
13775 2013-07-24 David Malcolm <dmalcolm@redhat.com>
13777 Introduce context class.
13779 * Makefile.in (CONTEXT_H): New.
13780 (OBJS): Add context.o.
13781 (toplev.o): Add CONTEXT_H to dependencies.
13784 * toplev.c (general_init): Create the singleton gcc::context instance.
13790 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
13792 PR rtl-optimization/57968
13793 * mode-switching.c (create_pre_exit): Allow instructions that
13794 don't set a return register to need a non-exit mode.
13796 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13797 Anton Blanchard <anton@au1.ibm.com>
13799 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
13800 operands to vperm for little endian.
13801 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
13802 of lvsl to create the control mask for a vperm for little endian.
13804 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13805 Anton Blanchard <anton@au1.ibm.com>
13807 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
13808 two operands for little-endian.
13810 2013-07-23 Steve Ellcey <sellcey@mips.com>
13812 * config/mips/mips.c (mips_case_values_threshold): New.
13813 (TARGET_CASE_VALUES_THRESHOLD): Define.
13815 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13816 Anton Blanchard <anton@au1.ibm.com>
13818 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
13819 selection of field for vector splat in little endian mode.
13821 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
13823 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
13824 expanders to rs6000.md.
13825 (ior<mode>3): Likewise.
13826 (and<mode>3): Likewise.
13827 (one_cmpl<mode>2): Likewise.
13828 (nor<mode>3): Likewise.
13829 (andc<mode>3): Likewise.
13830 (eqv<mode>3): Likewise.
13831 (nand<mode>3): Likewise.
13832 (orc<mode>3): Likewise.
13834 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
13837 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
13838 to split multi-word logical operations.
13839 (rs6000_split_logical_di): Likewise.
13840 (rs6000_split_logical): Likewise.
13842 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
13843 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
13844 and allow TImode operations in 32-bit.
13845 (vsx_and<mode>3_64bit): Likewise.
13846 (vsx_ior<mode>3_32bit): Likewise.
13847 (vsx_ior<mode>3_64bit): Likewise.
13848 (vsx_xor<mode>3_32bit): Likewise.
13849 (vsx_xor<mode>3_64bit): Likewise.
13850 (vsx_one_cmpl<mode>2_32bit): Likewise.
13851 (vsx_one_cmpl<mode>2_64bit): Likewise.
13852 (vsx_nor<mode>3_32bit): Likewise.
13853 (vsx_nor<mode>3_64bit): Likewise.
13854 (vsx_andc<mode>3_32bit): Likewise.
13855 (vsx_andc<mode>3_64bit): Likewise.
13856 (vsx_eqv<mode>3_32bit): Likewise.
13857 (vsx_eqv<mode>3_64bit): Likewise.
13858 (vsx_nand<mode>3_32bit): Likewise.
13859 (vsx_nand<mode>3_64bit): Likewise.
13860 (vsx_orc<mode>3_32bit): Likewise.
13861 (vsx_orc<mode>3_64bit): Likewise.
13863 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
13864 logical types in GPRs.
13866 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
13867 logical insns to rs6000.md, and allow TImode operations in
13869 (altivec_ior<mode>3): Likewise.
13870 (altivec_xor<mode>3): Likewise.
13871 (altivec_one_cmpl<mode>2): Likewise.
13872 (altivec_nor<mode>3): Likewise.
13873 (altivec_andc<mode>3): Likewise.
13875 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
13876 attributes for moving the 128-bit logical operations into
13878 (BOOL_REGS_OUTPUT): Likewise.
13879 (BOOL_REGS_OP1): Likewise.
13880 (BOOL_REGS_OP2): Likewise.
13881 (BOOL_REGS_UNARY): Likewise.
13882 (BOOL_REGS_AND_CR0): Likewise.
13883 (one_cmpl<mode>2): Add support for DI logical operations on
13884 32-bit, splitting the operations to 32-bit.
13885 (anddi3): Likewise.
13886 (iordi3): Likewise.
13887 (xordi3): Likewise.
13888 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
13889 changes to combine the 32/64-bit code, allow logical operations on
13890 TI mode in 32-bit, and to use similar match_operator patterns like
13891 scalar mode uses. Combine the Altivec and VSX code for logical
13892 operations, and move it here.
13893 (ior<mode>3, 128-bit types): Likewise.
13894 (xor<mode>3, 128-bit types): Likewise.
13895 (one_cmpl<mode>3, 128-bit types): Likewise.
13896 (nor<mode>3, 128-bit types): Likewise.
13897 (andc<mode>3, 128-bit types): Likewise.
13898 (eqv<mode>3, 128-bit types): Likewise.
13899 (nand<mode>3, 128-bit types): Likewise.
13900 (orc<mode>3, 128-bit types): Likewise.
13901 (and<mode>3_internal): Likewise.
13902 (bool<mode>3_internal): Likewise.
13903 (boolc<mode>3_internal1): Likewise.
13904 (boolc<mode>3_internal2): Likewise.
13905 (boolcc<mode>3_internal1): Likewise.
13906 (boolcc<mode>3_internal2): Likewise.
13907 (eqv<mode>3_internal1): Likewise.
13908 (eqv<mode>3_internal2): Likewise.
13909 (one_cmpl1<mode>3_internal): Likewise.
13911 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
13913 * config/microblaze/microblaze.c (microblaze_expand_prologue):
13914 Rename flag_stack_usage to flag_stack_usage_info.
13916 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
13918 * config/microblaze/sync.md: New file.
13919 * config/microblaze/microblaze.md: Include sync.md
13920 * config/microblaze/microblaze.c: Add print_operand 'y'.
13921 * config/microblaze/constraints.md: Add memory_contraint
13922 'Q' which is a single register.
13924 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
13926 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
13928 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
13930 * reload.c (find_reloads): Exit loop once we find this operand
13931 cannot be reloaded somehow for this alternative.
13933 * reload.c (find_reloads): Exit loop once we find a hard register.
13935 * rtlanal.c (computed_jump_p): Exit loop once we find label
13938 * i386.c (ix86_pad_returns): Exit loop after setting replace.
13940 * cfgloopmanip.c (remove_path): Exit loop after setting
13943 * gensupport.c (subst_dup): Avoid loop if code is not
13944 MATCH_DUP nor MATCH_OP_DUP.
13946 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
13948 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
13950 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13952 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
13953 true for SP_REGNUM if mode == ptr_mode.
13954 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
13955 with value R0_REGNUM + 31.
13957 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13959 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
13960 pad pointer-typed argument downward.
13962 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13964 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
13965 and __ILP32__ when the ILP32 model is in use.
13967 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13969 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
13970 (aarch64_load_symref_appropriately): In the case of
13971 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
13972 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
13973 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
13974 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
13975 if 'mode' doesn't equal to 'ptr_mode'.
13976 (aarch64_output_mi_thunk): Add an assertion on the alignment of
13977 'vcall_offset'; change to call aarch64_emit_move differently depending
13978 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
13979 to calculate the upper bound of 'vcall_offset'.
13980 (aarch64_cannot_force_const_mem): Change to also return true if
13982 (aarch64_legitimize_reload_address): In the case of large
13983 displacements, add new local variable 'xmode' and an assertion
13984 based on it; change to use 'xmode' to generate the new rtx and
13986 (aarch64_asm_trampoline_template): Change to generate the template
13987 differently depending on TARGET_ILP32 or not; change to use
13988 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
13989 (aarch64_trampoline_size): Removed.
13990 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
13991 and replace immediate literals with it. Change to use 'ptr_mode'
13992 instead of 'DImode' and call convert_memory_address if the mode
13993 of 'fnaddr' doesn't equal to 'ptr_mode'.
13994 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
13996 (aarch64_elf_asm_destructor): Likewise.
13997 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
13998 on TARGET_ILP32 instead of aarch64_trampoline_size.
13999 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14000 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14001 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14002 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14003 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14004 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14005 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14006 depending on the value of 'mode'.
14007 (add_losym_<mode>): New.
14008 (ldr_got_small_<mode>): New, based on ldr_got_small.
14009 (ldr_got_small): Remove.
14010 (ldr_got_small_sidi): New.
14011 * config/aarch64/iterators.md (P): New.
14012 (PTR): Change to 'ptr_mode' in the condition.
14014 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14016 * config.gcc (aarch64*-*-*): Support --with-abi.
14017 (aarch64*-*-elf): Support --with-multilib-list.
14018 (aarch64*-*-linux*): Likewise.
14019 (supported_defaults): Add abi to aarch64*-*-*.
14020 * configure.ac: Mention AArch64 for --with-multilib-list.
14021 * configure: Re-generated.
14022 * config/aarch64/biarchilp32.h: New file.
14023 * config/aarch64/biarchlp64.h: New file.
14024 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14026 (MULTILIB_DEFAULTS): Ditto.
14027 (DRIVER_SELF_SPECS): Ditto.
14028 (ASM_SPEC): Update to also substitute -mabi.
14029 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14030 file whose name depends on -mabi= and -mbig-endian.
14031 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14033 (POINTER_SIZE): New define.
14034 (POINTERS_EXTEND_UNSIGNED): Ditto.
14035 (enum aarch64_abi_type): New enumeration tag.
14036 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14037 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14038 (TARGET_ILP32): New define.
14039 * config/aarch64/aarch64.opt (mabi): New.
14040 (aarch64_abi): New.
14041 (ilp32, lp64): New values for -mabi.
14042 * config/aarch64/t-aarch64 (comma): New define.
14043 (MULTILIB_OPTIONS): Ditto.
14044 (MULTILIB_DIRNAMES): Ditto.
14045 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14046 * doc/invoke.texi: Document -mabi for AArch64.
14048 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14050 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14052 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14053 Anton Blanchard <anton@au1.ibm.com>
14055 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14056 endianness when selecting field to splat.
14058 2013-07-22 Eric Christopher <echristo@gmail.com>
14060 * dwarf2out.c (die_odr_checksum): New function to use
14061 CHECKSUM_ macros and ULEB128 for DIE tag.
14062 (generate_type_signature): Use.
14064 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14066 * config.gcc (sparc*-*-*): Accept leon3 processor.
14067 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14068 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14069 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14070 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14071 (mfix-ut699): Adjust comment.
14072 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14073 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14074 (CPP_CPU_SPEC): Likewise.
14075 (ASM_CPU_SPEC): Likewise.
14076 * config/sparc/sparc.c (leon3_cost): New constant.
14077 (sparc_option_override): Add leon3 support.
14078 (mem_ref): New function.
14079 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14080 (sparc_do_work_around_errata): Look into the instruction in the delay
14081 slot and adjust accordingly. Add fix for the data cache nullify issues
14082 of the UT699. Change insertion position for the NOP.
14083 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14084 (leon3_load): New reservation.
14085 (leon_store): Bump latency to 2.
14086 (grfpu): New automaton.
14087 (grfpu_alu): New unit.
14088 (grfpu_ds): Likewise.
14089 (leon_fp_alu): Adjust.
14090 (leon_fp_mult): Delete.
14091 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14092 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14093 * config/sparc/sparc.md (cpu): Add leon3.
14094 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14095 (swapsi): Likewise.
14096 (atomic_test_and_set): Likewise.
14097 (ldstub): Likewise.
14099 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14101 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14102 default for R5900 targets.
14103 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14104 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14105 * config/mips/mips.c (mips_option_override): Report an error for
14106 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14107 for -march=r5900 -mhard-float.
14109 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14111 * df-problems.c (can_move_insns_across): Exit loop once we
14112 find a non-fixed, non-global register.
14114 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14117 * omega.c (omega_eliminate_red): Break after setting red_found.
14118 (omega_problem_has_red_equations): Similarly after setting found.
14119 (omega_query_variable): Similarly after setting coupled.
14121 2013-07-22 Marek Polacek <polacek@redhat.com>
14123 * gimplify.c: Don't include gimple.h twice.
14125 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14127 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14128 instead of TARGET_THUMB1.
14129 (Pz): New constraint.
14130 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14132 (compare_negsi_si): Likewise.
14133 (compare_addsi2_op0): Likewise.
14134 (compare_addsi2_op1): Likewise.
14135 (addsi3_carryin_<optab>): Likewise.
14136 (addsi3_carryin_alt2_<optab>): Likewise.
14137 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14138 for arm_restrict_it.
14139 (subsi3_carryin): Likewise.
14140 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14141 (minmax_arithsi): Disable for arm_restrict_it.
14142 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14143 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14144 (satsi_<SAT:code>_shift): Likewise.
14145 (arm_shiftsi3): Add alternative for 16-bit encoding.
14146 (arm32_movhf): Disable for arm_restrict_it.
14147 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14148 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14150 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14152 * config/arm/arm.md (attribute "insn"): Delete.
14153 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14154 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14155 (not_shiftsi): Update for attribute change.
14156 (not_shiftsi_compare0): Likewise.
14157 (not_shiftsi_compare0_scratch): Likewise.
14158 (arm_one_cmplsi2): Likewise.
14159 (thumb1_one_cmplsi2): Likewise.
14160 (notsi_compare0): Likewise.
14161 (notsi_compare0_scratch): Likewise.
14162 (thumb1_movdi_insn): Likewise.
14163 (arm_movsi_insn): Likewise.
14164 (movhi_insn_arch4): Likewise.
14165 (movhi_bytes): Likewise.
14166 (arm_movqi_insn): Likewise.
14167 (thumb1_movqi_insn): Likewise.
14168 (arm32_movhf): Likewise.
14169 (thumb1_movhf): Likewise.
14170 (arm_movsf_soft_insn): Likewise.
14171 (thumb1_movsf_insn): Likewise.
14172 (thumb_movdf_insn): Likewise.
14173 (movsicc_insn): Likewise.
14174 (movsfcc_soft_insn): Likewise.
14175 (and_scc): Likewise.
14176 (cond_move): Likewise.
14177 (if_move_not): Likewise.
14178 (if_not_move): Likewise.
14179 (if_shift_move): Likewise.
14180 (if_move_shift): Likewise.
14181 (if_shift_shift): Likewise.
14182 (if_not_arith): Likewise.
14183 (if_arith_not): Likewise.
14184 (cond_move_not): Likewise.
14185 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
14186 (neon_mov<mode>): Likewise.
14187 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
14188 (thumb2_movsi_vfp): Likewise.
14189 (movsf_vfp): Likewise.
14190 (thumb2_movsf_vfp): Likewise.
14191 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14193 (cortexa7_older_only): Likewise.
14194 (cortexa7_younger): Likewise.
14195 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
14196 (1020alu_shift_op): Likewise.
14197 (1020alu_shift_reg_op): Likewise.
14198 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
14199 (alu_shift_op): Likewise.
14200 (alu_shift_reg_op): Likewise.
14201 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
14202 (11_alu_shift_op): Likewise.
14203 (11_alu_shift_reg_op): Likewise.
14204 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
14205 (9_alu_shift_reg_op): Likewise.
14206 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14208 (cortex_a15_alu_shift): Likewise.
14209 (cortex_a15_alu_shift_reg): Likewise.
14210 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
14211 (cortex_a5_alu_shift): Likewise.
14212 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
14214 (cortex_a53_alu_shift): Likewise.
14215 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14217 (cortex_a7_alu_reg): Likewise.
14218 (cortex_a7_alu_shift): Likewise.
14219 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14220 (cortex_a8_alu_shift): Likewise.
14221 (cortex_a8_alu_shift_reg): Likewise.
14222 (cortex_a8_mov): Likewise.
14223 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
14224 (cortex_a9_dp_shift): Likewise.
14225 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
14226 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
14227 (cortex_r4_mov): Likewise.
14228 (cortex_r4_alu_shift): Likewise.
14229 (cortex_r4_alu_shift_reg): Likewise.
14230 * config/arm/fa526.md (526_alu_op): Update for attribute change.
14231 (526_alu_shift_op): Likewise.
14232 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
14233 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
14234 (626te_alu_shift_op): Likewise.
14235 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
14236 (726te_alu_op): Likewise.
14237 (726te_alu_shift_op): Likewise.
14238 (726te_alu_shift_reg_op): Likewise.
14239 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
14240 (mp626_alu_shift_op): Likewise.
14241 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
14242 (pj4_alu_e1_conds): Likewise.
14243 (pj4_alu): Likewise.
14244 (pj4_alu_conds): Likewise.
14245 (pj4_shift): Likewise.
14246 (pj4_shift_conds): Likewise.
14247 (pj4_alu_shift): Likewise.
14248 (pj4_alu_shift_conds): Likewise.
14250 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14252 * config/arm/predicates.md (shiftable_operator_strict_it):
14254 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
14255 Disable cond_exec version for arm_restrict_it.
14256 (thumb2_smaxsi3): Convert to generate cond_exec.
14257 (thumb2_sminsi3): Likewise.
14258 (thumb32_umaxsi3): Likewise.
14259 (thumb2_uminsi3): Likewise.
14260 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
14261 (thumb2_neg_abssi2): Likewise.
14262 (thumb2_mov_scc): Add alternative for 16-bit encoding.
14263 (thumb2_movsicc_insn): Adjust alternatives.
14264 (thumb2_mov_negscc): Disable for arm_restrict_it.
14265 (thumb2_mov_negscc_strict_it): New pattern.
14266 (thumb2_mov_notscc_strict_it): New pattern.
14267 (thumb2_mov_notscc): Disable for arm_restrict_it.
14268 (thumb2_ior_scc): Likewise.
14269 (thumb2_ior_scc_strict_it): New pattern.
14270 (thumb2_cond_move): Adjust for arm_restrict_it.
14271 (thumb2_cond_arith): Disable for arm_restrict_it.
14272 (thumb2_cond_arith_strict_it): New pattern.
14273 (thumb2_cond_sub): Adjust for arm_restrict_it.
14274 (thumb2_movcond): Likewise.
14275 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
14276 (thumb2_zero_extendhisi2_v6): Likewise.
14277 (thumb2_zero_extendqisi2_v6): Likewise.
14278 (orsi_notsi_si): Likewise.
14279 (orsi_not_shiftsi_si): Likewise.
14281 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
14283 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
14284 instruction sequence is 1 byte shorter.
14286 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
14288 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
14289 it is not needed after split.
14291 2013-07-20 Iain Sandoe <iain@codesourcery.com>
14294 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
14295 second label for nonlocal goto receivers. Don't output pic base labels
14296 unless we're producing PIC; mark that action unreachable().
14297 (ix86_save_reg): If the function contains a nonlocal label, save the
14299 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
14300 * config/darwin.c (emitted_pic_label_num): New GTY.
14301 (update_pic_label_number_if_needed): New.
14302 (machopic_output_function_base_name): Adjust for nonlocal receiver
14304 (machopic_should_output_picbase_label): New.
14305 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
14306 (nonlocal_goto_receiver): New insn and split.
14308 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
14310 * config/aarch64/aarch64-builtins.c
14311 (aarch64_fold_builtin): Fold abs in all modes.
14312 * config/aarch64/aarch64-simd-builtins.def
14313 (abs): Enable for all modes.
14314 * config/aarch64/arm_neon.h
14315 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
14316 (vabs_f64): Add missing intrinsic.
14318 2013-07-19 Ian Bolton <ian.bolton@arm.com>
14320 * config/aarch64/arm_neon.h (vabs_s64): New function
14322 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
14325 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
14326 * config/avr/avr.md (adjust_len): Add `round'.
14327 * config/avr/avr-protos.h (avr_out_round): New prototype.
14328 (avr_out_plus): Add `out_label' argument.
14329 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
14330 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
14331 Handle the case where `insn' is just a pattern.
14332 (avr_out_bitop): Handle the case where `insn' is just a pattern.
14333 (avr_out_round): New function.
14334 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
14336 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
14338 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14339 Add check for flag_stack_usage to handle -fstack-usage support
14341 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
14343 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
14344 interaction for new Power8 flags and VSX.
14346 2013-07-18 Sriraman Tallam <tmsriram@google.com>
14348 PR middle-end/57698
14349 * tree-inline.c (expand_call_inline): Emit errors during
14350 early_inlining only if optimization is not turned on.
14352 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14356 * passes.c (init_optimization_passes): Move the construction of
14357 the pass hierarchy into a new passes.def file.
14359 * Makefile.in (passes.o): Add dependency on passes.def.
14361 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14363 * passes.c (init_optimization_passes): Introduce macros for
14364 constructing the tree of passes (INSERT_PASSES_AFTER,
14365 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
14366 TERMINATE_PASS_LIST).
14368 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
14369 Wei Mi <wmi@google.com>
14371 PR rtl-optimization/57878
14372 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
14374 (reload_pseudo_compare_func): Check nregs first for reload
14377 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14379 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
14381 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
14383 * read-rtl.c (validate_const_int): Once an invalid character is
14384 seen, quit the loop.
14386 * gengtype.c (write_roots): Similarly once we find the "deletable"
14387 or "if_marked" option.
14389 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14391 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
14392 "xtab" and "sat". Move value "clz" from here to ...
14393 (attriubte "type"): ... here.
14394 (satsi_<SAT:code>): Delete "insn" attribute.
14395 (satsi_<SAT:code>_shift): Likewise.
14396 (arm_zero_extendqisi2addsi): Likewise.
14397 (arm_extendqisi2addsi): Likewise.
14398 (clzsi2): Update for attribute changes.
14399 (rbitsi2): Likewise.
14400 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
14402 (arm_usatsihi): Likewise.
14403 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14405 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14407 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
14408 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
14409 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
14410 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
14411 in alphabetical order.
14412 (attribute "core_cycles"): Update for attribute changes.
14413 (arm_addsi3): Likewise.
14414 (addsi3_compare0): Likewise.
14415 (addsi3_compare0_scratch): Likewise.
14416 (addsi3_compare_op1): Likewise.
14417 (addsi3_compare_op2): Likewise.
14418 (compare_addsi2_op0): Likewise.
14419 (compare_addsi2_op1): Likewise.
14420 (addsi3_carryin_shift_<optab>): Likewise.
14421 (subsi3_carryin_shift): Likewise.
14422 (rsbsi3_carryin_shift): Likewise.
14423 (arm_subsi3_insn): Likewise.
14424 (subsi3_compare0): Likewise.
14425 (subsi3_compare): Likewise.
14426 (arm_andsi3_insn): Likewise.
14427 (thumb1_andsi3_insn): Likewise.
14428 (andsi3_compare0): Likewise.
14429 (andsi3_compare0_scratch): Likewise.
14430 (zeroextractsi_compare0_scratch
14431 (andsi_not_shiftsi_si): Likewise.
14432 (iorsi3_insn): Likewise.
14433 (iorsi3_compare0): Likewise.
14434 (iorsi3_compare0_scratch): Likewise.
14435 (arm_xorsi3): Likewise.
14436 (thumb1_xorsi3_insn): Likewise.
14437 (xorsi3_compare0): Likewise.
14438 (xorsi3_compare0_scratch): Likewise.
14439 (satsi_<SAT:code>_shift): Likewise.
14441 (arm_shiftsi3): Likewise.
14442 (shiftsi3_compare0): Likewise.
14443 (not_shiftsi): Likewise.
14444 (not_shiftsi_compare0): Likewise.
14445 (not_shiftsi_compare0_scratch): Likewise.
14446 (arm_one_cmplsi2): Likewise.
14447 (thumb_one_complsi2): Likewise.
14448 (notsi_compare0): Likewise.
14449 (notsi_compare0_scratch): Likewise.
14450 (thumb1_zero_extendhisi2): Likewise.
14451 (arm_zero_extendhisi2): Likewise.
14452 (arm_zero_extendhisi2_v6): Likewise.
14453 (arm_zero_extendhisi2addsi): Likewise.
14454 (thumb1_zero_extendqisi2): Likewise.
14455 (thumb1_zero_extendqisi2_v6): Likewise.
14456 (arm_zero_extendqisi2): Likewise.
14457 (arm_zero_extendqisi2_v6): Likewise.
14458 (arm_zero_extendqisi2addsi): Likewise.
14459 (thumb1_extendhisi2): Likewise.
14460 (arm_extendhisi2): Likewise.
14461 (arm_extendhisi2_v6): Likewise.
14462 (arm_extendqisi): Likewise.
14463 (arm_extendqisi_v6): Likewise.
14464 (arm_extendqisi2addsi): Likewise.
14465 (thumb1_extendqisi2): Likewise.
14466 (thumb1_movdi_insn): Likewise.
14467 (arm_movsi_insn): Likewise.
14468 (movsi_compare0): Likewise.
14469 (movhi_insn_arch4): Likewise.
14470 (movhi_bytes): Likewise.
14471 (arm_movqi_insn): Likewise.
14472 (thumb1_movqi_insn): Likewise.
14473 (arm32_movhf): Likewise.
14474 (thumb1_movhf): Likewise.
14475 (arm_movsf_soft_insn): Likewise.
14476 (thumb1_movsf_insn): Likewise.
14477 (movdf_soft_insn): Likewise.
14478 (thumb_movdf_insn): Likewise.
14479 (arm_cmpsi_insn): Likewise.
14480 (cmpsi_shiftsi): Likewise.
14481 (cmpsi_shiftsi_swp): Likewise.
14482 (arm_cmpsi_negshiftsi_si): Likewise.
14483 (movsicc_insn): Likewise.
14484 (movsfcc_soft_insn): Likewise.
14485 (arith_shiftsi): Likewise.
14486 (arith_shiftsi_compare0
14487 (arith_shiftsi_compare0_scratch
14488 (sub_shiftsi): Likewise.
14489 (sub_shiftsi_compare0
14490 (sub_shiftsi_compare0_scratch
14491 (and_scc): Likewise.
14492 (cond_move): Likewise.
14493 (if_plus_move): Likewise.
14494 (if_move_plus): Likewise.
14495 (if_move_not): Likewise.
14496 (if_not_move): Likewise.
14497 (if_shift_move): Likewise.
14498 (if_move_shift): Likewise.
14499 (if_shift_shift): Likewise.
14500 (if_not_arith): Likewise.
14501 (if_arith_not): Likewise.
14502 (cond_move_not): Likewise.
14503 (thumb1_ashlsi3): Set type attribute.
14504 (thumb1_ashrsi3): Likewise.
14505 (thumb1_lshrsi3): Likewise.
14506 (thumb1_rotrsi3): Likewise.
14507 (shiftsi3_compare0_scratch): Likewise.
14508 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
14509 (neon_mov<mode>): Likewise.
14510 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
14512 (thumb2_movsi_insn): Likewise.
14513 (thumb2_cmpsi_neg_shiftsi): Likewise.
14514 (thumb2_extendqisi_v6): Likewise.
14515 (thumb2_zero_extendhisi2_v6): Likewise.
14516 (thumb2_zero_extendqisi2_v6): Likewise.
14517 (thumb2_shiftsi3_short): Likewise.
14518 (thumb2_addsi3_compare0_scratch): Likewise.
14519 (orsi_not_shiftsi_si): Likewise.
14520 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
14521 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
14523 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
14524 (1020alu_shift_op): Likewise.
14525 (1020alu_shift_reg_op): Likewise.
14526 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
14527 (alu_shift_op): Likewise.
14528 (alu_shift_reg_op): Likewise.
14529 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
14530 (11_alu_shift_op): Likewise.
14531 (11_alu_shift_reg_op): Likewise.
14532 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
14533 (9_alu_shift_reg_op): Likewise.
14534 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14536 (cortex_a15_alu_shift): Likewise.
14537 (cortex_a15_alu_shift_reg): Likewise.
14538 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
14540 (cortex_a5_alu_shift): Likewise.
14541 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
14543 (cortex_a53_alu_shift): Likewise.
14544 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14546 (cortex_a7_alu_reg): Likewise.
14547 (cortex_a7_alu_shift): Likewise.
14548 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
14550 (cortex_a8_alu_shift): Likewise.
14551 (cortex_a8_alu_shift_reg): Likewise.
14552 (cortex_a8_mov): Likewise.
14553 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
14554 (cortex_a9_dp_shift): Likewise.
14555 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
14557 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
14559 (cortex_r4_mov): Likewise.
14560 (cortex_r4_alu_shift): Likewise.
14561 (cortex_r4_alu_shift_reg): Likewise.
14562 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
14563 (526_alu_shift_op): Likewise.
14564 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
14565 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
14566 (626te_alu_shift_op): Likewise.
14567 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
14568 (726te_alu_op): Likewise.
14569 (726te_alu_shift_op): Likewise.
14570 (726te_alu_shift_reg_op): Likewise.
14571 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
14572 (mp626_alu_shift_op): Likewise.
14573 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
14574 (pj4_alu_e1_conds): Likewise.
14575 (pj4_alu): Likewise.
14576 (pj4_alu_conds): Likewise.
14577 (pj4_shift): Likewise.
14578 (pj4_shift_conds): Likewise.
14579 (pj4_alu_shift): Likewise.
14580 (pj4_alu_shift_conds): Likewise.
14581 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14583 (cortexa7_older_only): Likewise.
14584 (cortexa7_younger): Likewise.
14586 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14588 * ipa-pure-const.c (generate_summary): Rename to...
14589 (pure_const_generate_summary): ... this.
14591 2013-07-17 Iain Sandoe <iain@codesourcery.com>
14593 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
14595 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
14598 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
14601 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14603 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
14604 enabled without -march=zEC12.
14605 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
14608 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
14610 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
14611 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
14612 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
14613 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
14614 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
14615 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
14616 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
14618 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
14619 and ISA_HAS_NMADD3_NMSUB3.
14620 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
14621 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
14622 (nmsub4<mode>, nmsub3<mode>): Likewise.
14623 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
14625 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
14627 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
14628 TARGET_MIPS5400 checking.
14630 2013-07-16 Jakub Jelinek <jakub@redhat.com>
14631 Peter Bergner <bergner@vnet.ibm.com>
14633 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
14634 registers in the comment.
14635 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
14636 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
14637 rather than FIRST_PSEUDO_REGISTERS.
14639 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
14641 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
14642 enable extra ISA flags with TARGET_HTM.
14644 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
14646 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
14649 2013-07-15 Cong Hou <congh@google.com>
14651 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
14652 in compare function for sorting.
14654 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
14656 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
14657 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
14658 * config/rs6000/rs6000.opt: Add -mhtm option.
14659 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
14660 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
14661 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
14662 __HTM__ if the HTM instructions are available.
14663 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
14664 htm_spr_reg_operand): New define_predicates.
14665 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
14666 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
14668 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
14669 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
14670 HTM builtin functions.
14671 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
14672 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
14673 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
14674 (rs6000_builtin_mask_calculate): Likewise.
14675 (rs6000_option_override_internal): Likewise.
14676 (bdesc_htm): Add new HTM builtin support.
14677 (htm_spr_num): New function.
14678 (htm_spr_regno): Likewise.
14679 (rs6000_htm_spr_icode): Likewise.
14680 (htm_expand_builtin): Likewise.
14681 (htm_init_builtins): Likewise.
14682 (rs6000_expand_builtin): Add support for HTM builtin functions.
14683 (rs6000_init_builtins): Likewise.
14684 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
14686 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
14687 (TARGET_HTM, MASK_HTM): Define macros.
14688 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
14689 (FIXED_REGISTERS): Likewise.
14690 (CALL_USED_REGISTERS): Likewise.
14691 (CALL_REALLY_USED_REGISTERS): Likewise.
14692 (REG_ALLOC_ORDER): Likewise.
14693 (enum reg_class): Likewise.
14694 (REG_CLASS_NAMES): Likewise.
14695 (REG_CLASS_CONTENTS): Likewise.
14696 (REGISTER_NAMES): Likewise.
14697 (ADDITIONAL_REGISTER_NAMES): Likewise.
14698 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
14699 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
14700 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
14701 * config/rs6000/htm.md: New file.
14702 * config/rs6000/htmintrin.h: New file.
14703 * config/rs6000/htmxlintrin.h: New file.
14705 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
14707 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
14708 Define SYMBOL_TINY_GOT, update comment.
14709 * config/aarch64/aarch64.c
14710 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
14711 (aarch64_expand_mov_immediate): Likewise.
14712 (aarch64_print_operand): Likewise.
14713 (aarch64_classify_symbol): Likewise.
14714 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
14715 (ldr_got_tiny): Define.
14717 2013-07-13 Tobias Grosser <tobias@grosser.es>
14719 PR tree-optimization/54094
14720 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
14721 scheduling dimension for the parallelism check from the polyhedral
14722 information in the AST.
14723 * graphite-dependences.c (carries_deps): Do not assume the schedule is
14726 2013-07-13 Jason Merrill <jason@redhat.com>
14728 * print-tree.c (debug_vec_tree): Use debug_raw.
14729 (debug_raw (vec<tree, va_gc> &)): New.
14730 (debug_raw (vec<tree, va_gc> *)): New.
14731 * tree.h: Declare them.
14733 2013-07-13 Bin Cheng <bin.cheng@arm.com>
14735 * ifcvt.c (ifcvt_after_combine): New static variable.
14736 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
14738 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
14739 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
14740 rest_of_handle_if_after_reload): Pass new argument for if_convert.
14742 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
14744 * config/mips/mips.c (mips_expand_call): Remove empty statement.
14746 2013-07-12 Michael Matz <matz@suse.de>
14748 PR middle-end/55771
14749 * convert.c (convert_to_real): Reject non-float inner types.
14751 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
14753 * config/aarch64/aarch64-protos.h
14754 (aarch64_simd_immediate_valid_for_move): Remove.
14755 * config/aarch64/aarch64.c (simd_immediate_info): New member.
14756 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
14758 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
14760 2013-07-11 Steve Ellcey <sellcey@mips.com>
14762 * config/mips/mips.c (mips_conditional_register_usage): Do not
14763 use t[0-7] registers in MIPS16 mode when optimizing for size.
14765 2013-07-11 Sriraman Tallam <tmsriram@google.com>
14767 * config/i386/i386.c (dispatch_function_versions): Fix array
14768 indexing of function_version_info to match actual_versions.
14770 2013-07-11 Teresa Johnson <tejohnson@google.com>
14772 * vec.h (struct va_gc): Move release out-of-line.
14773 (va_gc::release): Call ggc_free on released vec.
14775 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
14777 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
14778 Require GOT register as additional operand in UNSPEC.
14779 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
14780 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
14781 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
14782 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
14783 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
14784 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
14785 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
14787 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
14790 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
14791 name seen by assembler/linker rather if available.
14793 2013-07-11 Andreas Schwab <schwab@suse.de>
14795 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
14797 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
14799 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
14801 2013-07-10 Joseph Myers <joseph@codesourcery.com>
14803 * doc/tm.texi.in: Move hook documentation to ....
14804 * target.def: ... here.
14806 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
14807 text on @hook line.
14808 * doc/tm.texi: Regenerate.
14810 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
14813 * doc/invoke.texi: Document -Wconditionally-supported.
14815 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
14818 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
14821 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
14824 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
14825 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
14826 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
14827 Remove duplicate devices.
14828 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
14829 * config/avr/t-multilib: Regenerate.
14830 * config/avr/avr-tables.opt: Regenerate.
14831 * doc/avr-mmcu.texi: Regenerate.
14833 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
14836 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
14838 2013-07-10 Graham Stott <graham.stott@btinternet.com>
14840 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
14841 the cost of MULT when optimizing for size.
14843 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14845 * config/cr16/cr16-protos.h: Don't include target.h.
14847 2013-07-09 Joseph Myers <joseph@codesourcery.com>
14849 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
14850 adjust register size for TDmode and TFmode for VSX registers.
14852 2013-07-08 Kai Tietz <ktietz@redhat.com>
14855 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
14856 hook_bool_const_tree_true.
14858 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14860 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
14861 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
14862 * config/s390/s390.md: Define FPR*_REGNUM constants.
14863 Fix FPR2_REGNUM constant (18 -> 17).
14864 ("*trunc<BFP:mode><DFP_ALL:mode>2")
14865 ("*trunc<DFP_ALL:mode><BFP:mode>2")
14866 ("trunc<BFP:mode><DFP_ALL:mode>2")
14867 ("trunc<DFP_ALL:mode><BFP:mode>2")
14868 ("*extend<BFP:mode><DFP_ALL:mode>2")
14869 ("*extend<DFP_ALL:mode><BFP:mode>2")
14870 ("extend<BFP:mode><DFP_ALL:mode>2")
14871 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
14874 2013-07-08 Graham Stott <graham.stott@btinternet.com>
14876 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
14878 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14880 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
14881 and cfun_fpr_bit_p to cfun_fpr_save_p.
14882 (s390_frame_area, s390_register_info, s390_frame_info)
14883 (s390_emit_prologue, s390_emit_epilogue)
14884 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
14886 * config/s390/s390.h: Define *_REGNUM macros for floating point
14889 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
14891 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
14893 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
14895 PR rtl-optimization/57786
14896 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
14897 and break out of the loop when it is set to false.
14899 2013-07-08 Jakub Jelinek <jakub@redhat.com>
14902 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
14903 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
14904 (const_int 63)) 0)).
14905 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
14906 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
14907 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
14909 PR rtl-optimization/57829
14910 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
14911 mask bits outside of mode are just sign-extension from mode to HWI.
14913 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
14915 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
14916 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
14917 adjust_address instead of change_address to keep info about alignment.
14918 (emit_strmov): Remove.
14919 (emit_memmov): New function.
14920 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
14921 (expand_movmem_epilogue): Likewise and return updated rtx for
14923 (expand_constant_movmem_prologue): Likewise and return updated rtx for
14924 destination and source.
14925 (decide_alignment): Refactor, handle vector_loop.
14926 (ix86_expand_movmem): Likewise.
14927 (ix86_expand_setmem): Likewise.
14928 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
14930 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
14932 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
14933 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
14935 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
14937 * config/i386/sse.md (sse_movlhps): Change alternative 3
14938 of operand 2 to "m".
14940 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
14943 * config/i386/sse.md (iptr): New mode attribute.
14944 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
14945 (<sse>_vm<plusminus_insn><mode>3): Ditto.
14946 (<sse>_vmmul<mode>3): Ditto.
14947 (<sse>_vmdiv<mode>3): Ditto.
14948 (sse_vmrcpv4sf2): Ditto.
14949 (<sse>_vmsqrt<mode>2): Ditto.
14950 (sse_vmrsqrtv4sf2): Ditto.
14951 (<sse>_vm<code><mode>3): Ditto.
14952 (avx_vmcmp<mode>3): Ditto.
14953 (<sse>_vmmaskcmp<mode>3): Ditto.
14954 (<sse>_comi): Ditto.
14955 (<sse>_ucomi): Ditto.
14956 (*xop_vmfrcz_<mode>): Ditto.
14957 (*fmai_fmadd_<mode>): Ditto.
14958 (*fmai_fmsub_<mode>): Ditto.
14959 (*fmai_fnmadd_<mode>): Ditto.
14960 (*fmai_fnmsub_<mode>): Ditto.
14961 (*fma4i_vmfmadd_<mode>): Ditto.
14962 (*fma4i_vmfmsub_<mode>): Ditto.
14963 (*fma4i_vmfnmadd_<mode>): Ditto.
14964 (*fma4i_vmfnmsub_<mode>): Ditto.
14965 (*xop_vmfrcz_<mode>): Ditto.
14966 (sse_cvtps2pi): Ditto.
14967 (sse_cvttps2pi): Ditto.
14968 (sse_cvtss2si): Ditto.
14969 (sse_cvtss2si_2): Ditto.
14970 (sse_cvtss2siq_2): Ditto.
14971 (sse_cvttss2si): Ditto.
14972 (sse_cvttss2siq): Ditto.
14973 (sse_cvtsd2si): Ditto.
14974 (sse_cvtsd2si_2): Ditto.
14975 (sse_cvtsd2siq_2): Ditto.
14976 (sse_cvttsd2si): Ditto.
14977 (sse_cvttsd2siq): Ditto.
14978 (sse_cvtsd2ss): Ditto.
14979 (sse_cvtss2sd): Ditto.
14980 (avx2_pbroadcast<mode>): Ditto.
14981 (avx2_pbroadcast<mode>_1): Ditto.
14982 (*avx_vperm_broadcast_v4sf): Ditto.
14984 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
14985 (sse_movlhps): Ditto.
14986 (sse_storehps): Ditto.
14987 (sse_loadhps): Ditto.
14988 (sse_storelps): Ditto.
14989 (sse_loadlps): Ditto.
14990 (*vec_concatv4sf): Ditto.
14991 (*vec_interleave_highv2df): Ditto.
14992 (*vec_interleave_lowv2df): Ditto.
14993 (*vec_extractv2df_1_sse): Ditto.
14994 (*vec_extractv2df_0_sse): Ditto.
14995 (sse2_storelpd): Ditto.
14996 (sse2_loadlpd): Ditto.
14997 (sse2_movsd): Ditto.
14998 (*vec_concatv4si): Ditto.
14999 (vec_concatv2di): Ditto.
15001 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15002 for Intel asm dialect.
15003 (mmx_punpcklwd): Ditto.
15004 (mmx_punpckldq): Ditto.
15006 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15007 for intel assembler dialect.
15009 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15012 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15013 for BUILT_IN_C{LZ,LRSB}*.
15014 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15015 * tree-vrp.c (extract_range_basic): Handle
15016 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15017 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15018 fall thru to code calling set_value*.
15019 * builtins.c (expand_builtin): Remove *IMAX cases.
15020 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15021 if width is bigger than 2*HWI.
15023 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15025 PR rtl-optimization/55342
15026 * lra-int.h (lra_subreg_reload_pseudos): New.
15027 * lra.c: Add undoing optional reloads to the block diagram.
15028 (lra_subreg_reload_pseudos): New.
15029 (lra_optional_reload_pseudos): Change comments.
15030 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15031 lra_optional_reload_pseudos after undo transformations.
15032 * lra-assigns.c (pseudo_prefix_title): New.
15033 (lra_setup_reg_renumber): Use it.
15034 (spill_for): Ditto. Check subreg reload pseudos too.
15035 (assign_by_spills): Consider subreg reload pseudos too.
15036 * lra-constraints.c (simplify_operand_subreg): Use
15037 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15038 (curr_insn_transform): Recognize and do optional reloads.
15039 (undo_optional_reloads): New.
15040 (lra_undo_inheritance): Call undo_optional_reloads.
15042 2013-07-05 Thomas Quinot <quinot@adacore.com>
15044 * tree-complex.c (expand_complex_operations_1): Fix typo.
15046 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15048 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15049 (tune_params): New member 'const vec_costs'.
15050 * config/aarch64/aarch64.c (generic_vector_cost): New.
15051 (generic_tunings): New member 'generic_vector_cost'.
15052 (aarch64_builtin_vectorization_cost): New.
15053 (aarch64_add_stmt_cost): New.
15054 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15055 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15057 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15060 * config/i386/predicates.md (vsib_address_operand): Disallow
15061 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15063 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15065 PR middle-end/55030
15066 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15067 expand_builtin_setjmp_receiver.
15068 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15069 with NULL for the label parameter.
15070 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15071 the frame-pointer. Adjust comments.
15072 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15073 only if LABEL is non-NULL.
15075 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15077 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15078 (ARM_ABI_AAPCS64): Ditto.
15080 (ARM_DEFAULT_ABI): Ditto.
15082 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15084 * config/aarch64/aarch64-builtins.c
15085 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15086 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15088 * config/aarch64/aarch64-simd.md
15089 (aarch64_ld1<VALL:mode>): New.
15090 (aarch64_st1<VALL:mode>): Likewise.
15091 * config/aarch64/arm_neon.h
15092 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15094 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15096 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15098 (ix86_option_override_internal):Turn on all -mavx target flags by
15099 default as they are dependent on AVX anyway.
15101 2013-07-02 Cary Coutant <ccoutant@google.com>
15103 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15104 deterministic hash.
15105 (loc_checksum_ordered): Likewise.
15106 (hash_loc_operands): Remove inline keyword.
15108 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15110 PR tree-optimization/57741
15111 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15112 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15113 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15114 Allow REAL_CST step_exprs if flag_associative_math.
15115 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15117 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15119 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15121 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15123 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15125 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15127 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15129 (iorsi3_insn): Likewise.
15130 (arm_xorsi3): Likewise.
15132 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15134 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15136 (attribute "type"): ... here, and prefix with "wmmx_".
15137 (attribute "core_cycles"): Update for attribute changes.
15138 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15139 (tbcstv4hi): Likewise.
15140 (tbcstv2si): Likewise.
15141 (iwmmxt_iordi3): Likewise.
15142 (iwmmxt_xordi3): Likewise.
15143 (iwmmxt_anddi3): Likewise.
15144 (iwmmxt_nanddi3): Likewise.
15145 (iwmmxt_arm_movdi): Likewise.
15146 (iwmmxt_movsi_insn): Likewise.
15147 (mov<mode>_internal): Likewise.
15148 (and<mode>3_iwmmxt): Likewise.
15149 (ior<mode>3_iwmmxt): Likewise.
15150 (xor<mode>3_iwmmxt): Likewise.
15151 (add<mode>3_iwmmxt): Likewise.
15152 (ssaddv8qi3): Likewise.
15153 (ssaddv4hi3): Likewise.
15154 (ssaddv2si3): Likewise.
15155 (usaddv8qi3): Likewise.
15156 (usaddv4hi3): Likewise.
15157 (usaddv2si3): Likewise.
15158 (sub<mode>3_iwmmxt): Likewise.
15159 (sssubv8qi3): Likewise.
15160 (sssubv4hi3): Likewise.
15161 (sssubv2si3): Likewise.
15162 (ussubv8qi3): Likewise.
15163 (ussubv4hi3): Likewise.
15164 (ussubv2si3): Likewise.
15165 (mulv4hi3_iwmmxt): Likewise.
15166 (smulv4hi3_highpart): Likewise.
15167 (umulv4hi3_highpart): Likewise.
15168 (iwmmxt_wmacs): Likewise.
15169 (iwmmxt_wmacsz): Likewise.
15170 (iwmmxt_wmacu): Likewise.
15171 (iwmmxt_wmacuz): Likewise.
15172 (iwmmxt_clrdi): Likewise.
15173 (iwmmxt_clrv8qi): Likewise.
15174 (iwmmxt_clr4hi): Likewise.
15175 (iwmmxt_clr2si): Likewise.
15176 (iwmmxt_uavgrndv8qi3): Likewise.
15177 (iwmmxt_uavgrndv4hi3): Likewise.
15178 (iwmmxt_uavgv8qi3): Likewise.
15179 (iwmmxt_uavgv4hi3): Likewise.
15180 (iwmmxt_tinsrb): Likewise.
15181 (iwmmxt_tinsrh): Likewise.
15182 (iwmmxt_tinsrw): Likewise.
15183 (iwmmxt_textrmub): Likewise.
15184 (iwmmxt_textrmsb): Likewise.
15185 (iwmmxt_textrmuh): Likewise.
15186 (iwmmxt_textrmsh): Likewise.
15187 (iwmmxt_textrmw): Likewise.
15188 (iwmxxt_wshufh): Likewise.
15189 (eqv8qi3): Likewise.
15190 (eqv4hi3): Likewise.
15191 (eqv2si3): Likewise.
15192 (gtuv8qi3): Likewise.
15193 (gtuv4hi3): Likewise.
15194 (gtuv2si3): Likewise.
15195 (gtv8qi3): Likewise.
15196 (gtv4hi3): Likewise.
15197 (gtv2si3): Likewise.
15198 (smax<mode>3_iwmmxt): Likewise.
15199 (umax<mode>3_iwmmxt): Likewise.
15200 (smin<mode>3_iwmmxt): Likewise.
15201 (umin<mode>3_iwmmxt): Likewise.
15202 (iwmmxt_wpackhss): Likewise.
15203 (iwmmxt_wpackwss): Likewise.
15204 (iwmmxt_wpackdss): Likewise.
15205 (iwmmxt_wpackhus): Likewise.
15206 (iwmmxt_wpackwus): Likewise.
15207 (iwmmxt_wpackdus): Likewise.
15208 (iwmmxt_wunpckihb): Likewise.
15209 (iwmmxt_wunpckihh): Likewise.
15210 (iwmmxt_wunpckihw): Likewise.
15211 (iwmmxt_wunpckilb): Likewise.
15212 (iwmmxt_wunpckilh): Likewise.
15213 (iwmmxt_wunpckilw): Likewise.
15214 (iwmmxt_wunpckehub): Likewise.
15215 (iwmmxt_wunpckehuh): Likewise.
15216 (iwmmxt_wunpckehuw): Likewise.
15217 (iwmmxt_wunpckehsb): Likewise.
15218 (iwmmxt_wunpckehsh): Likewise.
15219 (iwmmxt_wunpckehsw): Likewise.
15220 (iwmmxt_wunpckelub): Likewise.
15221 (iwmmxt_wunpckeluh): Likewise.
15222 (iwmmxt_wunpckeluw): Likewise.
15223 (iwmmxt_wunpckelsb): Likewise.
15224 (iwmmxt_wunpckelsh): Likewise.
15225 (iwmmxt_wunpckelsw): Likewise.
15226 (ror<mode>3): Likewise.
15227 (ashr<mode>3_iwmmxt): Likewise.
15228 (lshr<mode>3_iwmmxt): Likewise.
15229 (ashl<mode>3_iwmmxt): Likewise.
15230 (ror<mode>3_di): Likewise.
15231 (ashr<mode>3_di): Likewise.
15232 (lshr<mode>3_di): Likewise.
15233 (ashl<mode>3_di): Likewise.
15234 (iwmmxt_wmadds): Likewise.
15235 (iwmmxt_wmaddu): Likewise.
15236 (iwmmxt_tmia): Likewise.
15237 (iwmmxt_tmiaph): Likewise.
15238 (iwmmxt_tmiabb): Likewise.
15239 (iwmmxt_tmiatb): Likewise.
15240 (iwmmxt_tmiabt): Likewise.
15241 (iwmmxt_tmiatt): Likewise.
15242 (iwmmxt_tmovmskb): Likewise.
15243 (iwmmxt_tmovmskh): Likewise.
15244 (iwmmxt_tmovmskw): Likewise.
15245 (iwmmxt_waccb): Likewise.
15246 (iwmmxt_wacch): Likewise.
15247 (iwmmxt_waccw): Likewise.
15248 (iwmmxt_waligni): Likewise.
15249 (iwmmxt_walignr): Likewise.
15250 (iwmmxt_walignr0): Likewise.
15251 (iwmmxt_walignr1): Likewise.
15252 (iwmmxt_walignr2): Likewise.
15253 (iwmmxt_walignr3): Likewise.
15254 (iwmmxt_wsadb): Likewise.
15255 (iwmmxt_wsadh): Likewise.
15256 (iwmmxt_wsadbz): Likewise.
15257 (iwmmxt_wsadhz): Likewise.
15258 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
15259 (iwmmxt_wabsdiffb): Likewise.
15260 (iwmmxt_wabsdiffh): Likewise.
15261 (iwmmxt_wabsdiffw): Likewise.
15262 (iwmmxt_waddsubhx): Likewise
15263 (iwmmxt_wsubaddhx): Likewise.
15264 (addc<mode>3): Likewise.
15265 (iwmmxt_avg4): Likewise.
15266 (iwmmxt_avg4r): Likewise.
15267 (iwmmxt_wmaddsx): Likewise.
15268 (iwmmxt_wmaddux): Likewise.
15269 (iwmmxt_wmaddsn): Likewise.
15270 (iwmmxt_wmaddun): Likewise.
15271 (iwmmxt_wmulwsm): Likewise.
15272 (iwmmxt_wmulwum): Likewise.
15273 (iwmmxt_wmulsmr): Likewise.
15274 (iwmmxt_wmulumr): Likewise.
15275 (iwmmxt_wmulwsmr): Likewise.
15276 (iwmmxt_wmulwumr): Likewise.
15277 (iwmmxt_wmulwl): Likewise.
15278 (iwmmxt_wqmulm): Likewise.
15279 (iwmmxt_wqmulwm): Likewise.
15280 (iwmmxt_wqmulmr): Likewise.
15281 (iwmmxt_wqmulwmr): Likewise.
15282 (iwmmxt_waddbhusm): Likewise.
15283 (iwmmxt_waddbhusl): Likewise.
15284 (iwmmxt_wqmiabb): Likewise.
15285 (iwmmxt_wqmiabt): Likewise.
15286 (iwmmxt_wqmiatb): Likewise.
15287 (iwmmxt_wqmiatt): Likewise.
15288 (iwmmxt_wqmiabbn): Likewise.
15289 (iwmmxt_wqmiabtn): Likewise.
15290 (iwmmxt_wqmiatbn): Likewise.
15291 (iwmmxt_wqmiattn): Likewise.
15292 (iwmmxt_wmiabb): Likewise.
15293 (iwmmxt_wmiabt): Likewise.
15294 (iwmmxt_wmiatb): Likewise.
15295 (iwmmxt_wmiatt): Likewise.
15296 (iwmmxt_wmiabbn): Likewise.
15297 (iwmmxt_wmiabtn): Likewise.
15298 (iwmmxt_wmiatbn): Likewise.
15299 (iwmmxt_wmiattn): Likewise.
15300 (iwmmxt_wmiawbb): Likewise.
15301 (iwmmxt_wmiawbt): Likewise.
15302 (iwmmxt_wmiawtb): Likewise.
15303 (iwmmxt_wmiawtt): Likewise.
15304 (iwmmxt_wmiawbbn): Likewise.
15305 (iwmmxt_wmiawbtn): Likewise.
15306 (iwmmxt_wmiawtbn): Likewise.
15307 (iwmmxt_wmiawttn): Likewise.
15308 (iwmmxt_wmerge): Likewise.
15309 (iwmmxt_tandc<mode>3): Likewise.
15310 (iwmmxt_torc<mode>3): Likewise.
15311 (iwmmxt_torvsc<mode>3): Likewise.
15312 (iwmmxt_textrc<mode>3): Likewise.
15313 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
15314 (wmmxt_pack): Likewise.
15315 (wmmxt_mult_c1): Likewise.
15316 (wmmxt_mult_c2): Likewise.
15317 (wmmxt_alu_c1): Likewise.
15318 (wmmxt_alu_c2): Likewise.
15319 (wmmxt_alu_c3): Likewise.
15320 (wmmxt_transfer_c1): Likewise.
15321 (wmmxt_transfer_c2): Likewise.
15322 (wmmxt_transfer_c3): Likewise.
15323 (marvell_f_iwmmxt_wstr): Likewise.
15324 (marvell_f_iwmmxt_wldr): Likewise.
15326 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
15328 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
15330 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15333 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15334 * lra-constraints.c (need_for_split_p): Check call used hard regs
15335 living through calls.
15337 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
15338 call used regs for call insn.
15340 2013-06-28 Jakub Jelinek <jakub@redhat.com>
15343 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
15344 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
15346 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15348 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
15350 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15352 * lra-constraints.c (need_for_split_p): Check call used hard regs
15353 living through calls.
15355 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
15358 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
15359 to tie with any other modes. Eliminate Altivec vector mode tests,
15360 since these are a subset of ALTIVEC or VSX vector modes. Simplify
15361 code, to return 0 if testing MODE2 for a condition, if we've
15362 already tested MODE1 for the same condition.
15364 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15366 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
15369 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15371 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15372 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
15374 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15376 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
15378 (aarch64_symbolic_constant_p): Remove.
15379 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
15380 static. Fix line length and white space.
15381 (aarch64_symbolic_constant_p): Remove.
15382 * config/aarch64/predicates.md (aarch64_valid_symref):
15383 Use aarch64_classify_symbol_expression.
15385 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15387 * config/arm/constraints.md (Ts): New constraint.
15388 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
15390 (compare_scc): Use "Ts" constraint for operand 0.
15391 (ior_scc_scc): Likewise.
15392 (and_scc_scc): Likewise.
15393 (and_scc_scc_nodom): Likewise.
15394 (ior_scc_scc_cmp): Likewise for operand 7.
15395 (and_scc_scc_cmp): Likewise.
15396 * config/arm/thumb2.md (thumb2_movsi_insn):
15397 Add alternatives for 16-bit encodings.
15398 (thumb2_movhi_insn): Likewise.
15399 (thumb2_movsicc_insn): Likewise.
15400 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
15401 (thumb2_negscc): Use "Ts" constraint.
15402 Move mvn instruction outside cond_exec block.
15403 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
15404 for 16-bit encodings.
15406 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15408 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
15410 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
15411 (mulsi3subsi): Likewise.
15412 (mulsidi3adddi): Likewise.
15413 (mulsidi3_v6): Likewise.
15414 (umulsidi3_v6): Likewise.
15415 (umulsidi3adddi_v6): Likewise.
15416 (smulsi3_highpart_v6): Likewise.
15417 (umulsi3_highpart_v6): Likewise.
15418 (mulhisi3tb): Likewise.
15419 (mulhisi3bt): Likewise.
15420 (mulhisi3tt): Likewise.
15421 (maddhisi4): Likewise.
15422 (maddhisi4tb): Likewise.
15423 (maddhisi4tt): Likewise.
15424 (maddhidi4): Likewise.
15425 (maddhidi4tb): Likewise.
15426 (maddhidi4tt): Likewise.
15427 (zeroextractsi_compare0_scratch): Likewise.
15428 (insv_zero): Likewise.
15429 (insv_t2): Likewise.
15430 (anddi_notzesidi_di): Likewise.
15431 (anddi_notsesidi_di): Likewise.
15432 (andsi_notsi_si): Likewise.
15433 (iordi_zesidi_di): Likewise.
15434 (xordi_zesidi_di): Likewise.
15435 (andsi_iorsi3_notsi): Likewise.
15436 (smax_0): Likewise.
15437 (smax_m1): Likewise.
15438 (smin_0): Likewise.
15439 (not_shiftsi): Likewise.
15440 (unaligned_loadsi): Likewise.
15441 (unaligned_loadhis): Likewise.
15442 (unaligned_loadhiu): Likewise.
15443 (unaligned_storesi): Likewise.
15444 (unaligned_storehi): Likewise.
15445 (extv_reg): Likewise.
15446 (extzv_t2): Likewise.
15447 (divsi3): Likewise.
15448 (udivsi3): Likewise.
15449 (arm_zero_extendhisi2addsi): Likewise.
15450 (arm_zero_extendqisi2addsi): Likewise.
15451 (compareqi_eq0): Likewise.
15452 (arm_extendhisi2_v6): Likewise.
15453 (arm_extendqisi2addsi): Likewise.
15454 (arm_movt): Likewise.
15455 (thumb2_ldrd): Likewise.
15456 (thumb2_ldrd_base): Likewise.
15457 (thumb2_ldrd_base_neg): Likewise.
15458 (thumb2_strd): Likewise.
15459 (thumb2_strd_base): Likewise.
15460 (thumb2_strd_base_neg): Likewise.
15461 (arm_negsi2): Add alternative for 16-bit encoding.
15462 (arm_one_cmplsi2): Likewise.
15464 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15466 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
15467 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
15468 (movdfcc): Likewise.
15469 * config/arm/vfp.md (*thumb2_movsf_vfp):
15470 Disable predication for arm_restrict_it.
15471 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
15472 (*thumb2_movdfcc_vfp): Likewise.
15473 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
15474 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
15475 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
15476 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
15477 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
15478 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
15479 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
15480 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
15481 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
15482 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
15483 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
15484 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
15485 Disable predication for arm_restrict_it.
15487 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
15489 * config/i386/bmiintrin.h (_bextr_u32): New.
15490 (_bextr_u64): Ditto.
15492 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
15494 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
15495 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
15496 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
15497 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
15498 * config/mips/n32-elf.h: ...this new file.
15500 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
15503 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
15504 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
15506 2013-06-27 Catherine Moore <clm@codesourcery.com>
15508 * config/mips/mips-tables.opt: Regenerate.
15509 * config/mips/mips-cpus.def: Add m14ke and m14kec.
15510 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
15511 * doc/invoke.texi: Add -m14kc.
15513 2013-06-27 Jakub Jelinek <jakub@redhat.com>
15516 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
15517 constraints of operand 1 and 2.
15520 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
15521 to match RTL canonicalization. Swap predicates and
15522 constraints of operand 1 and 2.
15524 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
15526 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
15527 Process OP_INOUT regs for splitting too.
15529 2013-06-27 Jakub Jelinek <jakub@redhat.com>
15531 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
15532 decl before the loop, initialize to NULL.
15533 (vectorizable_load): Initialize ptr_incr to NULL.
15535 2013-06-27 Martin Jambor <mjambor@suse.cz>
15538 * ipa-ref.h (ipa_maybe_record_reference): Declare.
15539 * ipa-ref.c (ipa_maybe_record_reference): New function.
15540 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
15541 * ipa-cp.c (create_specialized_node): Record potential references from
15543 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
15545 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
15547 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
15548 parameter 'mode' of type 'enum machine_mode mode'; change to pass
15549 'mode' to force_reg.
15550 (aarch64_add_offset): Update calls to aarch64_force_temporary.
15551 (aarch64_expand_mov_immediate): Likewise.
15553 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
15555 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
15556 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
15558 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15560 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
15561 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
15562 (struct machine_function): Add tbegin_p.
15563 (s390_canonicalize_comparison): Fold CC mode compares to
15564 conditional jump if possible.
15565 (s390_emit_jump): Return the emitted jump.
15566 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
15567 Handle CCRAWmode compares.
15568 (s390_option_override): Default to -mhtm if available.
15569 (s390_reg_clobbered_rtx): Handle floating point regs as well.
15570 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
15571 FPRs instead of df_regs_ever_live_p.
15572 (s390_optimize_nonescaping_tx): New function.
15573 (s390_init_frame_layout): Extend clobbered_regs array to cover
15575 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
15576 (s390_expand_tbegin): New function.
15577 (enum s390_builtin): New enum definition.
15578 (code_for_builtin): New array definition.
15579 (s390_init_builtins): New function.
15580 (s390_expand_builtin): New function.
15581 (TARGET_INIT_BUILTINS): Define.
15582 (TARGET_EXPAND_BUILTIN): Define.
15583 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
15584 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
15585 (s390_alc_comparison): Likewise.
15586 * config/s390/s390-modes.def: Add CCRAWmode.
15587 * config/s390/s390.h (processor_flags): Add PF_TX.
15588 (TARGET_CPU_HTM): Define macro.
15589 (TARGET_HTM): Define macro.
15590 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
15591 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
15592 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
15593 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
15594 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
15595 (TBEGIN_MASK, TBEGINC_MASK): New constants.
15596 ("*cc_to_int"): Move up.
15597 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
15598 constants other than 0.
15599 ("*ccraw_to_int"): New insn and splitter definition.
15600 ("tbegin", "tbegin_nofloat", "tbegin_retry")
15601 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
15602 ("tx_assist"): New expander.
15603 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
15604 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
15605 * config/s390/s390.opt: Add -mhtm option.
15606 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
15607 * config/s390/htmxlintrin.h: New file.
15608 * config/s390/htmintrin.h: New file.
15609 * config/s390/s390intrin.h: New file.
15610 * doc/extend.texi: Document htm builtins.
15611 * config.gcc: Add the new header files to extra_headers.
15613 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
15615 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
15616 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
15618 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
15619 Pat Haugen <pthaugen@us.ibm.com>
15620 Peter Bergner <bergner@vnet.ibm.com>
15622 * config/rs6000/power8.md: New.
15623 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
15624 setting for power8 entry.
15625 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
15626 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
15627 test for Power4/Power5 only.
15628 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
15630 (force_new_group): Adjust comment.
15631 * config/rs6000/rs6000.md: Include power8.md.
15633 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
15635 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
15636 * config/arm/arm-protos.h (arm_max_conditional_execute): New
15638 (tune_params): Update comment.
15639 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
15640 (arm_max_conditional_execute): New function.
15641 (thumb2_final_prescan_insn): Use max_insn_skipped and
15642 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
15644 2013-06-25 Jakub Jelinek <jakub@redhat.com>
15646 PR tree-optimization/57705
15647 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
15648 SSA_NAME step, provided that it is not defined inside the loop.
15649 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
15650 (get_initial_def_for_induction): Handle SSA_NAME IV step.
15652 2013-06-25 Martin Jambor <mjambor@suse.cz>
15654 PR middle-end/57670
15655 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
15656 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
15658 (ipa_note_param_call): Initialize member_ptr flag.
15659 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
15660 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
15661 (ipa_write_indirect_edge_info): Stream member_ptr flag.
15662 (ipa_read_indirect_edge_info): Likewise.
15664 2013-06-25 Richard Biener <rguenther@suse.de>
15666 PR middle-end/56977
15667 * passes.c (init_optimization_passes): Move pass_fold_builtins
15668 and pass_dce earlier with -Og.
15670 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
15672 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
15673 <BIT_FIELD_REF>: Remove trailing TAB.
15674 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
15677 2013-06-24 Martin Jambor <mjambor@suse.cz>
15679 PR tree-optimization/57358
15680 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
15681 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
15682 (ipa_analyze_params_uses): Generate pessimistic info when true.
15684 2013-06-24 Martin Jambor <mjambor@suse.cz>
15686 PR tree-optimization/57539
15687 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
15688 global.inlined_to of the new node to it. All callers changed.
15689 * ipa-inline-transform.c (clone_inlined_nodes): New variable
15690 inlining_into, pass it to cgraph_clone_node.
15691 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
15692 ipa_free_edge_args_substructures.
15693 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
15694 rdesc linked list. Do not assert rdesc edges have inlined caller.
15695 Assert we have found an rdesc in the rdesc list.
15697 2013-06-24 Richard Biener <rguenther@suse.de>
15699 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
15700 (pointer_set_lookup): Declare.
15701 (class pointer_map): New template class implementing a
15702 generic pointer to T map.
15703 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
15704 pointer_map<T>::contains, pointer_map<T>::insert,
15705 pointer_map<T>::traverse): New functions.
15706 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
15707 (pointer_set_lookup): New function.
15708 (pointer_set_contains): Use pointer_set_lookup.
15709 (pointer_set_insert): Likewise.
15710 (insert_aux): Remove.
15711 (struct pointer_map_t): Embed a pointer_set_t.
15712 (pointer_map_create): Adjust.
15713 (pointer_map_destroy): Likewise.
15714 (pointer_map_contains): Likewise.
15715 (pointer_map_insert): Likewise.
15716 (pointer_map_traverse): Likewise.
15717 * tree-streamer.h (struct streamer_tree_cache_d): Use a
15718 pointer_map<unsigned> instead of a pointer_map_t.
15719 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
15720 (streamer_tree_cache_lookup): Likewise.
15721 (streamer_tree_cache_create): Likewise.
15722 (streamer_tree_cache_delete): Likewise.
15723 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
15724 pointer_map<unsigned> instead of a pointer_map_t.
15725 (lto_init_tree_ref_encoder): Adjust.
15726 (lto_destroy_tree_ref_encoder): Likewise.
15727 * lto-section-out.c (lto_output_decl_index): Likewise.
15728 (lto_record_function_out_decl_state): Likewise.
15729 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
15731 2013-06-24 Richard Biener <rguenther@suse.de>
15733 PR tree-optimization/57488
15734 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
15736 2013-06-24 Alan Modra <amodra@gmail.com>
15738 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
15739 (gen_easy_altivec_constant): Likewise.
15740 * config/rs6000/predicates.md (easy_vector_constant_add_self,
15741 easy_vector_constant_msb): Likewise.
15743 2013-06-23 Jakub Jelinek <jakub@redhat.com>
15746 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
15747 add missing return true.
15749 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
15752 * config/sh/predicates.md (general_extend_operand): Invoke
15753 general_movsrc_operand for memory operands.
15754 (general_movsrc_operand): Allow reg+reg addressing, do not use
15755 general_operand for memory operands.
15757 2013-06-23 Sriraman Tallam <tmsriram@google.com>
15759 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
15760 when current target options does not apply.
15761 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
15762 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
15763 * config/i386/bmiintrin.h: Pass appropriate target
15764 attributes to header.
15765 * config/i386/mmintrin.h: Ditto.
15766 * config/i386/nmmintrin.h: Ditto.
15767 * config/i386/avx2intrin.h: Ditto.
15768 * config/i386/fxsrintrin.h: Ditto.
15769 * config/i386/tbmintrin.h: Ditto.
15770 * config/i386/xsaveintrin.h: Ditto.
15771 * config/i386/f16cintrin.h: Ditto.
15772 * config/i386/xtestintrin.h: Ditto.
15773 * config/i386/xsaveoptintrin.h: Ditto.
15774 * config/i386/bmi2intrin.h: Ditto.
15775 * config/i386/lzcntintrin.h: Ditto.
15776 * config/i386/smmintrin.h: Ditto.
15777 * config/i386/wmmintrin.h: Ditto.
15778 * config/i386/x86intrin.h: Remove all header include guards.
15779 * config/i386/prfchwintrin.h: Ditto.
15780 * config/i386/pmmintrin.h: Ditto.
15781 * config/i386/tmmintrin.h: Ditto.
15782 * config/i386/xmmintrin.h: Ditto.
15783 * config/i386/popcntintrin.h: Ditto.
15784 * config/i386/rdseedintrin.h: Ditto.
15785 * config/i386/ammintrin.h: Ditto.
15786 * config/i386/emmintrin.h: Ditto.
15787 * config/i386/immintrin.h: Remove all header include guards.
15788 * config/i386/fma4intrin.h: Ditto.
15789 * config/i386/lwpintrin.h: Ditto.
15790 * config/i386/xopintrin.h: Ditto.
15791 * config/i386/ia32intrin.h: Ditto.
15792 * config/i386/avxintrin.h: Ditto.
15793 * config/i386/rtmintrin.h: Ditto.
15794 * config/i386/fmaintrin.h: Ditto.
15795 * config/i386/mm3dnow.h: Ditto.
15797 2013-06-22 Sriraman Tallam <tmsriram@google.com>
15799 * common/config/i386/i386-common.c: Handle LZCNT.
15801 2013-06-22 Andi Kleen <ak@linux.intel.com>
15803 * doc/extend.texi: Use __atomic_store_n instead of
15804 __atomic_store in HLE example.
15806 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
15808 * config/sh/sh.c: Remove <cstdlib> workaround.
15810 2013-06-21 Andi Kleen <ak@linux.intel.com>
15812 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
15814 2013-06-21 Andi Kleen <ak@linux.intel.com>
15816 * doc/extend.texi: Document that __atomic_clear and
15817 __atomic_test_and_set should only be used with bool.
15819 2013-06-20 Jan Hubicka <jh@suse.cz>
15821 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
15822 types_same_for_odr.
15823 * tree.c (decls_same_for_odr): New function.
15824 (same_for_edr): New function.
15825 (types_same_for_odr): New function.
15826 (get_binfo_at_offset): Use it.
15827 * tree.h (types_same_for_odr): Declare.
15829 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
15830 Jason Merrill <jason@redhat.com>
15832 * system.h: Include <cstdlib> as well as <stdlib.h>.
15834 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
15837 * config/i386/i386.c (construct_container): Report error if
15838 long double is used with disabled x87 float returns.
15840 2013-06-20 Jan Hubicka <jh@suse.cz>
15842 * lto-cgraph.c (input_symtab): Do not set cgraph state.
15844 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
15846 PR rtl-optimization/57425
15847 PR rtl-optimization/57569
15848 * alias.c (write_dependence_p): Remove parameters mem_mode and
15849 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
15850 Changed all callers.
15851 (canon_anti_dependence): Get comments and semantics in sync.
15852 Add parameter mem_canonicalized. Changed all callers.
15853 * rtl.h (canon_anti_dependence): Update prototype.
15855 2013-06-20 Richard Biener <rguenther@suse.de>
15857 * data-streamer-in.c (streamer_read_uhwi): Optimize single
15858 byte case, inline streamer_read_uchar and defer section
15861 2013-06-20 Richard Biener <rguenther@suse.de>
15863 PR tree-optimization/57584
15864 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
15865 SSA names into the expanded expression that take part in
15866 abnormal coalescing.
15868 2013-06-19 Sharad Singhai <singhai@google.com>
15870 * gcov.c (print_usage): Handle new option.
15871 (process_args): Ditto.
15872 (get_gcov_intermediate_filename): New function.
15873 (output_intermediate_file): New function.
15874 (output_gcov_file): New function
15875 (generate_results): Handle new option.
15876 (release_function): Relase demangled name.
15877 (read_graph_file): Handle demangled name.
15878 (output_lines): Ditto.
15879 * doc/gcov.texi: Document gcov intermediate format.
15881 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
15884 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
15885 (lra_emit_add): Use the functions. Add comment about Y as an
15888 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
15891 * collect2.c (collect_atexit): New.
15892 (collect_exit): Delete.
15893 (main): Register collect_atexit with atexit.
15894 (collect_wait): Change collect_exit to exit.
15896 * collect2.h (collect_exit): Delete.
15897 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
15899 2013-06-19 Wei Mi <wmi@google.com>
15901 PR rtl-optimization/57518
15902 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
15903 if regno is used in paradoxical subreg.
15904 (update_equiv_regs): Check pdx_subregs[regno] before
15905 set a reg to be equivalent with a mem.
15907 2013-06-19 Matthias Klose <doko@ubuntu.com>
15910 * file-find.h (find_a_file): Add a mode parameter.
15911 * file-find.c (find_a_file): Likewise.
15912 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
15913 with X_OK for the executables.
15914 * collect2.c (main): Call find_a_file with X_OK.
15916 2013-06-19 Steve Ellcey <sellcey@mips.com>
15919 * config/mips/mips.md (casesi_internal_mips16_<mode>):
15920 Use NEXT_INSN instead of next_real_insn.
15922 2013-06-19 Jan Hubicka <jh@suse.cz>
15924 * cgraph.h (const_value_known_p): Replace by ...
15925 (ctor_for_folding): .. this one.
15926 * cgraphunit.c (process_function_and_variable_attributes): Use it.
15927 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
15928 * expr.c (expand_expr_real_1): Likewise.
15929 (string_constant): Likewise.
15930 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
15931 * ipa.c (process_references): Likewise.
15932 (symtab_remove_unreachable_nodes): Likewise.
15933 * ipa-inline-analysis.c (param_change_prob): Likewise.
15934 * gimple-fold.c (canonicalize_constructor_val): Likewise.
15935 (get_base_constructor): Likwise.
15936 * varpool.c (varpool_remove_node): Likewise.
15937 (varpool_remove_initializer): LIkewise.
15938 (dump_varpool_node): LIkwise.
15939 (const_value_known_p): Rewrite to ...
15940 (ctor_for_folding): ... this one.
15942 2013-06-19 Jakub Jelinek <jakub@redhat.com>
15945 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
15946 PERSONALITY in $PATH derived prefixes.
15948 2013-06-19 Jeff Law <law@redhat.com>
15950 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
15953 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
15954 (simplify_bitwise_binary): Use it to simpify certain binary ops on
15957 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
15959 * config/arm/vfp.md: Move VFP instruction classification documentation
15961 * config/arm/arm.md: ... here. Update instruction classification
15964 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
15966 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
15967 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
15968 pattern. Use more efficient sequences on ARMv5 and Thumb2.
15970 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
15973 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
15974 with NEXT_INSN. Use tablejump_p to check for jump table data
15977 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
15980 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
15981 that now in C++ the value is correct per the C++ standards.
15983 2013-06-19 Richard Biener <rguenther@suse.de>
15985 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
15986 for global context.
15988 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15991 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15994 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
15995 with next_active_insn.
15997 2013-06-18 Sriraman Tallam <tmsriram@google.com>
15999 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16000 functions are inlined during failures to flag an error.
16001 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16002 in early inline pass.
16004 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16006 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16009 2013-06-18 Julian Brown <julian@codesourcery.com>
16011 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16012 Permit virtual register pre-reload if !strict.
16013 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16015 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16017 * config/arm/neon.md (movmisalign<mode>): Use
16018 neon_perm_struct_or_reg_operand instead of
16019 neon_struct_or_register_operand.
16020 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16021 neon_permissive_struct_operand instead of neon_struct_operand.
16022 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16023 neon_vector_mem_operand.
16024 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16025 neon_vector_mem_operand.
16026 (neon_permissive_struct_operand): New.
16027 (neon_struct_or_register_operand): Rename to...
16028 (neon_perm_struct_or_reg_operand): This. Adjust call to
16029 neon_vector_mem_operand.
16031 2013-06-18 Richard Biener <rguenther@suse.de>
16033 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16034 * lto-streamer.h: Include pointer-set.h.
16035 (struct lto_decl_slot): Remove.
16036 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16037 Remove next_index entry.
16038 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16039 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16040 (lto_init_tree_ref_encoder): Adjust.
16041 (lto_destroy_tree_ref_encoder): Likewise.
16042 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16043 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16044 (lto_output_decl_index): Adjust.
16045 (lto_new_out_decl_state): Likewise.
16046 (lto_record_function_out_decl_state): Likewise.
16047 * lto-streamer-out.c (copy_function): Likewise.
16049 2013-06-18 Richard Biener <rguenther@suse.de>
16051 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16052 * cgraphunit.c: Include cfgloop.h.
16053 (init_lowered_empty_function): Initialize the loop tree.
16054 (assemble_thunk): Insert new BBs into loops.
16056 2013-06-18 Richard Biener <rguenther@suse.de>
16058 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16059 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16060 the map from cache entry to cache index optional.
16061 (streamer_tree_cache_replace_tree): Adjust accordingly.
16062 (streamer_tree_cache_append): Likewise.
16063 (streamer_tree_cache_delete): Likewise.
16064 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16065 streamer cache map from cache entry to cache index.
16066 * lto-streamer-out.c (create_output_block): Adjust.
16068 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16070 * config/arm/arm.md (attribute "insn"): Move multiplication and
16071 division attributes to...
16072 (attribute "type"): ... here. Remove mult.
16073 (attribute "mul32"): New attribute.
16074 (attribute "mul64"): Add umaal.
16075 (*arm_mulsi3): Update attributes.
16076 (*arm_mulsi3_v6): Likewise.
16077 (*thumb_mulsi3): Likewise.
16078 (*thumb_mulsi3_v6): Likewise.
16079 (*mulsi3_compare0): Likewise.
16080 (*mulsi3_compare0_v6): Likewise.
16081 (*mulsi_compare0_scratch): Likewise.
16082 (*mulsi_compare0_scratch_v6): Likewise.
16083 (*mulsi3addsi): Likewise.
16084 (*mulsi3addsi_v6): Likewise.
16085 (*mulsi3addsi_compare0): Likewise.
16086 (*mulsi3addsi_compare0_v6): Likewise.
16087 (*mulsi3addsi_compare0_scratch): Likewise.
16088 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16089 (*mulsi3subsi): Likewise.
16090 (*mulsidi3adddi): Likewise.
16091 (*mulsi3addsi_v6): Likewise.
16092 (*mulsidi3adddi_v6): Likewise.
16093 (*mulsidi3_nov6): Likewise.
16094 (*mulsidi3_v6): Likewise.
16095 (*umulsidi3_nov6): Likewise.
16096 (*umulsidi3_v6): Likewise.
16097 (*umulsidi3adddi): Likewise.
16098 (*umulsidi3adddi_v6): Likewise.
16099 (*smulsi3_highpart_nov6): Likewise.
16100 (*smulsi3_highpart_v6): Likewise.
16101 (*umulsi3_highpart_nov6): Likewise.
16102 (*umulsi3_highpart_v6): Likewise.
16103 (mulhisi3): Likewise.
16104 (*mulhisi3tb): Likewise.
16105 (*mulhisi3bt): Likewise.
16106 (*mulhisi3tt): Likewise.
16107 (maddhisi4): Likewise.
16108 (*maddhisi4tb): Likewise.
16109 (*maddhisi4tt): Likewise.
16110 (maddhidi4): Likewise.
16111 (*maddhidi4tb): Likewise.
16112 (*maddhidi4tt): Likewise.
16113 (divsi3): Likewise.
16114 (udivsi3): Likewise.
16115 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16116 (thumb2_mulsi_short_compare0): Likewise.
16117 (thumb2_mulsi_short_compare0_scratch): Likewise.
16118 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16119 (1020mult2): Likewise.
16120 (1020mult3): Likewise.
16121 (1020mult4): Likewise.
16122 (1020mult5): Likewise.
16123 (1020mult6): Likewise.
16124 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16126 (cortex_a15_mult64): Likewise.
16127 (cortex_a15_sdiv): Likewise.
16128 (cortex_a15_udiv): Likewise.
16129 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16135 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16136 (pj4_ir_div): Likewise.
16137 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16138 (11_mult2): Likewise.
16139 (11_mult3): Likewise.
16140 (11_mult4): Likewise.
16141 (11_mult5): Likewise.
16142 (11_mult6): Likewise.
16143 (11_mult7): Likewise.
16144 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16145 (cortex_a8_mla): Likewise.
16146 (cortex_a8_mull): Likewise.
16147 (cortex_a8_smulwy): Likewise.
16148 (cortex_a8_smlald): Likewise.
16149 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16150 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16151 (cortex_r4_mul_3): Likewise.
16152 (cortex_r4_mla_4): Likewise.
16153 (cortex_r4_mla_3): Likewise.
16154 (cortex_r4_smlald): Likewise.
16155 (cortex_r4_mull): Likewise.
16156 (cortex_r4_sdiv): Likewise.
16157 (cortex_r4_udiv): Likewise.
16158 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16159 (cortex_a7_idiv): Likewise.
16160 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16161 (9_mult2): Likewise.
16162 (9_mult3): Likewise.
16163 (9_mult4): Likewise.
16164 (9_mult5): Likewise.
16165 (9_mult6): Likewise.
16166 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16167 (cortex_a53_sdiv): Likewise.
16168 (cortex_a53_udiv): Likewise.
16169 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16170 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16171 (mp626_mult2): Likewise.
16172 (mp626_mult3): Likewise.
16173 (mp626_mult4): Likewise.
16174 * config/arm/fa526.md (526_mult1): Update attribute change.
16175 (526_mult2): Likewise.
16176 * config/arm/arm-generic.md (mult): Update attribute change.
16177 (mult_ldsched_strongarm): Likewise.
16178 (mult_ldsched): Likewise.
16179 (multi_cycle): Likewise.
16180 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16181 * config/arm/fa606te.md (606te_mult1): Update attribute change.
16182 (606te_mult2): Likewise.
16183 (606te_mult3): Likewise.
16184 (606te_mult4): Likewise.
16185 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
16186 (cortex_a9_mac16): Likewise.
16187 (cortex_a9_multiply): Likewise.
16188 (cortex_a9_mac): Likewise.
16189 (cortex_a9_multiply_long): Likewise.
16190 * config/arm/fa626te.md (626te_mult1): Update attribute change.
16191 (626te_mult2): Likewise.
16192 (626te_mult3): Likewise.
16193 (626te_mult4): Likewise.
16195 2013-06-18 Richard Biener <rguenther@suse.de>
16198 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
16200 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16203 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16204 with next_active_insn.
16206 2013-06-18 Alan Modra <amodra@gmail.com>
16208 * config/rs6000/rs6000.h (enum data_align): New.
16209 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
16210 (DATA_ABI_ALIGNMENT): Define.
16211 (CONSTANT_ALIGNMENT): Correct comment.
16212 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
16213 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
16215 2013-06-17 David Malcolm <dmalcolm@redhat.com>
16217 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
16218 ATTRIBUTE_UNUSED marking.
16220 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
16222 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
16223 alternative and update.
16224 (aarch64_dup_lanedi): Delete.
16225 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
16226 * config/aarch64/aarch64-simd-builtins.def: Update.
16228 2013-06-17 Richard Biener <rguenther@suse.de>
16230 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
16231 (lto_input_scc): Declare.
16232 (lto_input_tree_1): Likewise.
16233 (struct lto_stats_d): Add num_tree_bodies_output and
16234 num_pickle_refs_output.
16235 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
16236 (lto_read_tree_1): Split out from ...
16237 (lto_read_tree): ... this.
16238 (lto_input_scc): New function.
16239 (lto_input_tree_1): Split out from ...
16240 (lto_input_tree): ... this. Handle LTO_tree_scc.
16241 (lto_data_in_create): Create the streamer cache without hashes.
16242 * lto-streamer-out.c (create_output_block): Create the streamer
16243 cache with hashes when not doing WPA.
16244 (lto_write_tree_1): Split out from ...
16245 (lto_write_tree): ... this.
16246 (get_symbol_initial_value): New function.
16247 (lto_output_tree_1): Split out from ...
16248 (lto_output_tree): ... this. Write trees as series of SCCs
16249 using a DFS walk via DFS_write_tree.
16250 (struct sccs, struct scc_entry): New types.
16251 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
16252 (DFS_write_tree_body): New function.
16253 (DFS_write_tree): Likewise.
16254 (hash_tree): Likewise.
16255 (scc_entry_compare): Likewise.
16256 (hash_scc): Likewise.
16257 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
16258 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
16259 TREE_CHAIN as regular reference.
16260 (streamer_read_integer_cst): Remove.
16261 (streamer_get_pickled_tree): Adjust.
16262 * tree-streamer-out.c (streamer_write_chain): Disable streaming
16263 of DECL_EXTERNALs in BLOCK_VARS for now.
16264 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
16266 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
16267 Add hash value argument and record that if hashes are recorded
16269 (streamer_tree_cache_insert_1): Adjust.
16270 (streamer_tree_cache_insert): Likewise.
16271 (streamer_tree_cache_insert_at): Rename to ...
16272 (streamer_tree_cache_replace_tree): ... this and adjust.
16273 (streamer_tree_cache_append): Adjust.
16274 (record_common_node): Likewise.
16275 (streamer_tree_cache_create): Add argument whether to
16276 record hash values together with trees.
16277 (streamer_tree_cache_delete): Adjust.
16278 * tree-streamer.h (struct streamer_tree_cache_d): Add
16280 (streamer_read_integer_cst): Remove.
16281 (streamer_tree_cache_insert): Adjust.
16282 (streamer_tree_cache_append): Likewise.
16283 (streamer_tree_cache_insert_at): Rename to ...
16284 (streamer_tree_cache_replace_tree): ... this and adjust.
16285 (streamer_tree_cache_create): Add argument whether to record hashes.
16286 (streamer_tree_cache_get): Rename to ...
16287 (streamer_tree_cache_get_tree): ... this.
16288 (streamer_tree_cache_get_hash): New function.
16289 * tree.c (cache_integer_cst): New function.
16290 * tree.h (cache_integer_cst): Declare.
16291 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
16292 * lto-symtab.c (lto_varpool_replace_node): Only release
16293 DECL_INITIAL of non-prevailing decls.
16294 * varpool.c (varpool_remove_initializer): Do not release
16295 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
16297 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16299 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
16300 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
16301 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
16302 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
16303 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
16304 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
16305 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
16306 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
16307 instead of TARGET_64BIT.
16308 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
16309 Require ISA_HAS_<D>DIV.
16311 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
16313 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
16314 (mips*-*-linux*): Move default with_llsc setting to where other
16316 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
16318 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
16319 Likewise. Remove default with_tune setting. Move default float
16320 setting to its own block. Handle with_llsc in the same block as above.
16322 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
16324 PR rtl-optimization/57425
16325 PR rtl-optimization/57569
16326 * alias.c (write_dependence_p): Add new parameters mem_mode,
16327 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
16328 Changed all callers.
16329 (canon_anti_dependence): New function.
16330 * cse.c (check_dependence): Use canon_anti_dependence.
16331 * cselib.c (cselib_invalidate_mem): Likewise.
16332 * rtl.h (canon_anti_dependence): Declare.
16334 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16336 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
16337 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
16338 ".set mips3" for 64-bit targets.
16340 2013-06-15 Dehao Chen <dehao@google.com>
16342 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
16343 * gimple-low.c (gimple_check_call_matching_types): Likewise.
16344 (gimple_check_call_args): Likewise.
16345 * value-prof.c (check_ic_target): Likewise.
16346 * ipa-inline.c (early_inliner): Likewise.
16347 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
16348 * cgraph.c (cgraph_create_edge_1): Likewise.
16349 (cgraph_make_edge_direct): Likewise.
16351 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
16354 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
16355 rs6000_output_move_128bit to handle emitting quad memory
16356 operations. Set attribute length to 8 bytes.
16358 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
16360 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
16362 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
16363 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
16364 (aarch64_<su>mlsl<mode>): Likewise.
16366 2013-06-14 Mike Stump <mikestump@comcast.net>
16368 * Makefile.in (TARGET_H): Add insn-codes.h.
16370 2013-06-14 Alan Modra <amodra@gmail.com>
16372 PR middle-end/57134
16373 PR middle-end/57586
16374 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
16375 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
16376 bitfield expansion when EXPAND_MEMORY.
16377 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
16379 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
16381 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
16382 test for clearing quad memory on 32-bit later.
16384 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
16386 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
16387 (fold_negate_expr): Likewise.
16388 (fold_real_zero_addition_p): Handle vectors.
16389 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
16391 2013-06-14 Alan Modra <amodra@gmail.com>
16393 * varasm.c (force_const_mem): Revert 2013-06-07 change.
16395 2013-06-13 Jan Hubicka <jh@suse.cz>
16397 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16398 Local comdats are not externally visible.
16399 * symtab.c (dump_symtab_base): Dump externally visible.
16400 (verify_symtab_base): Verify back links in the symtab hash.
16402 2013-06-13 Bin Cheng <bin.cheng@arm.com>
16404 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
16405 CONVERT_EXPR as equal nodes.
16407 2013-06-13 Bin Cheng <bin.cheng@arm.com>
16409 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
16411 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
16413 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
16414 Generalize to complex and vector.
16415 * tree.c (build_all_ones_cst): New function.
16416 * tree.h (build_all_ones_cst): Declare it.
16418 2013-06-13 Alan Modra <amodra@gmail.com>
16420 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
16421 * config/rs6000/rs6000.md (signbittf2): New insn.
16422 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
16423 (abstf2_internal, cmptf_internal2): Likewise.
16424 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
16426 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
16427 Pat Haugen <pthaugen@us.ibm.com>
16428 Peter Bergner <bergner@vnet.ibm.com>
16430 * config/rs6000/rs6000.c (emit_load_locked): Add support for
16431 power8 byte, half-word, and quad-word atomic instructions.
16432 (emit_store_conditional): Likewise.
16433 (rs6000_expand_atomic_compare_and_swap): Likewise.
16434 (rs6000_expand_atomic_op): Likewise.
16436 * config/rs6000/sync.md (larx): Add new modes for power8.
16438 (AINT): New mode iterator to include TImode as well as normal
16439 integer modes on power8.
16440 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
16441 that VSX registers are not considered. Use AINT mode iterator
16442 instead of INT1 to allow inclusion of quad word atomic operations
16444 (load_locked<mode>): Likewise.
16445 (store_conditional<mode>): Likewise.
16446 (atomic_compare_and_swap<mode>): Likewise.
16447 (atomic_exchange<mode>): Likewise.
16448 (atomic_nand<mode>): Likewise.
16449 (atomic_fetch_<fetchop_name><mode>): Likewise.
16450 (atomic_nand_fetch<mode>): Likewise.
16451 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
16453 (ATOMIC): On power8, add QImode, HImode modes.
16454 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
16455 modes that promote to SImode.
16456 (load_lockedti): Convert TImode arguments to PTImode, so that we
16457 get a guaranteed even/odd register pair.
16458 (load_lockedpti): Likewise.
16459 (store_conditionalti): Likewise.
16460 (store_conditionalpti): Likewise.
16462 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
16463 atomic load/store instructions.
16466 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
16468 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
16470 (insn_count): New attribute, with most cases extracted from...
16471 (length): ...here. Redefine most cases in terms of insn_count.
16472 (single_insn): Delete.
16473 (can_delay): Use insn_count to check for single instructions.
16474 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
16475 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
16476 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
16477 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
16478 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
16479 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
16480 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
16481 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
16482 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
16483 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
16484 rather than "length".
16485 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
16486 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
16487 Use "insn_count" rather than "length".
16488 * config/mips/mips-dsp.md
16489 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
16490 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
16493 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
16495 PR tree-optimization/57361
16496 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
16498 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
16500 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
16502 (aarch64_simd_combine<mode>): New instruction expansion.
16503 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
16504 function prototype.
16505 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
16506 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
16508 2013-06-12 Jan Hubicka <jh@suse.cz>
16510 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
16511 decl has when in streaming stage.
16512 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
16513 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
16515 2013-06-12 Roland Stigge <stigge@antcom.de>
16518 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
16520 2013-06-12 Jakub Jelinek <jakub@redhat.com>
16522 PR tree-optimization/57537
16523 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
16524 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
16526 2013-06-12 Richard Biener <rguenther@suse.de>
16528 * data-streamer.h (streamer_write_char_stream): CSE
16529 obs->current_pointer.
16530 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
16531 streamer_write_char_stream manually and optimize the resulting loop.
16532 (streamer_write_hwi_stream): Likewise.
16534 2013-06-12 Jan Hubicka <jh@suse.cz>
16536 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
16537 * cgraph.h (varpool_create_empty_node): Declare.
16538 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
16540 * symtab.c (symtab_unregister_node): Be lax about missin entries
16542 (symtab_get_node): Update comment.
16543 * varpool.c (varpool_create_empty_node): Break out from ...
16544 (varpool_node_for_decl): ... here.
16545 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
16547 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
16549 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
16550 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
16551 part. Use straight-line flow at the end.
16552 <COMPONENT_REF>: Remove superfluous else.
16553 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
16555 2013-06-12 Jakub Jelinek <jakub@redhat.com>
16558 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
16559 target hook even for !TREE_PUBLIC decls. If no resolution info
16560 is available, return false for common and external decls.
16562 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16564 * config/rl78/constraints.md (U): New constraint.
16565 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
16568 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
16571 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
16572 to allow returning address to AT_PLATFORM name.
16574 2013-06-11 Jan Hubicka <jh@suse.cz>
16576 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
16577 * cgraph.h (symtab_node_base): Add weakref flag.
16578 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
16579 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
16580 (output_weakrefs): Use weakref flag.
16581 * fold-const.c (simple_operand_p): Handle WEAK.
16582 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
16583 * ipa.c (varpool_externally_visible_p): Drop weakref.
16584 (function_and_variable_visibility): Update comment; fix weakref
16585 sanity checks; do not clear DECL_WEAK on them.
16586 * lto-cgraph.c (lto_output_node): update.
16587 (lto_output_varpool_node): Update.
16588 (input_overwrite_node): Update.
16589 (input_node): Update.
16590 (input_varpool_node): Update.
16591 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
16592 (lto_symtab_merge_symbols): Add sanity check.
16593 (lto_symtab_prevailing_decl): Do not special case weakrefs.
16594 * passes.c (rest_of_decl_compilation): Set static flag, too.
16595 * symtab.c (dump_symtab_base): Dump weakref.
16596 (verify_symtab_base): Sanity check weakrefs.
16597 (symtab_make_decl_local): Remove duplicated code.
16598 (symtab_alias_ultimate_target): Simplify.
16599 * varpool.c (varpool_create_variable_alias): Set weakref flag.
16601 2013-06-11 Tom de Vries <tom@codesourcery.com>
16603 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
16604 sequence_vect == NULL.
16606 2013-06-11 DJ Delorie <dj@redhat.com>
16608 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
16609 (rl78_unwind_word_mode): New.
16611 2013-06-11 David Malcolm <dmalcolm@redhat.com>
16613 * final.c (debug_prefix_maps): Make static.
16615 2013-06-11 David Malcolm <dmalcolm@redhat.com>
16617 * function.c (initial_trampoline): Remove stray copy.
16619 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
16621 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
16623 2013-06-11 Martin Jambor <mjambor@suse.cz>
16625 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
16626 within bounds at the beginning of the function.
16628 2013-06-11 Alan Modra <amodra@gmail.com>
16630 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
16632 (get_named_section): Don't NULL !DECL_P decl.
16634 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
16636 * doc/invoke.texi (core-avx2): Document.
16638 (atom): Updated with MOVBE.
16640 2013-06-11 Richard Biener <rguenther@suse.de>
16642 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
16644 2013-06-11 Anton Blanchard <anton@samba.org>
16646 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
16647 correct shift value in little-endian mode.
16649 2013-06-11 Jakub Jelinek <jakub@redhat.com>
16652 * varasm.c (get_variable_align): Move #endif to the right place.
16654 2013-06-10 Cary Coutant <ccoutant@google.com>
16656 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
16657 for hash so that hash table traversal order is deterministic.
16659 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
16660 Pat Haugen <pthaugen@us.ibm.com>
16661 Peter Bergner <bergner@vnet.ibm.com>
16663 * config/rs6000/vector.md (GPR move splitter): Do not split moves
16664 of vectors in GPRS if they are direct moves or quad word load or
16667 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
16669 (direct_move_p): Likewise.
16670 (quad_load_store_p): Likewise.
16672 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
16673 classes into bins based on the physical register type.
16674 (reg_class_to_reg_type): Likewise.
16675 (IS_STD_REG_TYPE): Likewise.
16676 (IS_FP_VECT_REG_TYPE): Likewise.
16677 (reload_fpr_gpr): Arrays to determine what insn to use if we can
16678 use direct move instructions.
16679 (reload_gpr_vsx): Likewise.
16680 (reload_vsx_gpr): Likewise.
16681 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
16682 information that is a simplification of register classes. Also
16683 precalculate direct move reload helpers.
16684 (direct_move_p): New function to return true if the operation can
16685 be done as a direct move instruciton.
16686 (quad_load_store_p): New function to return true if the operation
16687 is a quad memory operation.
16688 (rs6000_legitimize_address): If quad memory, only allow register
16689 indirect for TImode addresses.
16690 (rs6000_legitimate_address_p): Likewise.
16691 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
16692 (rs6000_reload_register_type): Likewise.
16693 (register_to_reg_type): Return register type.
16694 (rs6000_secondary_reload_simple_move): New helper function for
16695 secondary reload and secondary memory needed to identify anything
16696 that is a simple move, and does not need reloading.
16697 (rs6000_secondary_reload_direct_move): New helper function for
16698 secondary reload to identify cases that can be done with several
16699 instructions via the direct move instructions.
16700 (rs6000_secondary_reload_move): New helper function for secondary
16701 reload to identify moves between register types that can be done.
16702 (rs6000_secondary_reload): Add support for quad memory operations
16703 and for direct move.
16704 (rs6000_secondary_memory_needed): Likewise.
16705 (rs6000_debug_secondary_memory_needed): Change argument names.
16706 (rs6000_output_move_128bit): New function to return the move to
16707 use for 128-bit moves, including knowing about the various
16708 limitations of quad memory operations.
16710 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
16711 memory operations. call rs6000_output_move_128bit for the actual
16712 instruciton(s) to generate.
16713 (vsx_movti_64bit): Likewise.
16715 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
16716 (UNSPEC_P8V_MTVSRWZ): Likewise.
16717 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
16718 (UNSPEC_P8V_MTVSRD): Likewise.
16719 (UNSPEC_P8V_XXPERMDI): Likewise.
16720 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
16721 (UNSPEC_FUSION_GPR): Likewise.
16722 (FMOVE128_GPR): New iterator for direct move.
16723 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
16724 (f32_sv): Likewise.
16725 (f32_dm): Likewise.
16726 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
16727 loads and direct move instructions.
16728 (zero_extendsidi2_lfiwzx): Likewise.
16729 (extendsidi2_lfiwax): Likewise.
16730 (extendsidi2_nocell): Likewise.
16731 (floatsi<mode>2_lfiwax): Likewise.
16732 (lfiwax): Likewise.
16733 (floatunssi<mode>2_lfiwzx): Likewise.
16734 (lfiwzx): Likewise.
16735 (fix_trunc<mode>_stfiwx): Likewise.
16736 (fixuns_trunc<mode>_stfiwx): Likewise.
16737 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
16738 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
16739 (parity<mode>2_cmpb): Set length/type attr.
16740 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
16741 for 'mr.' to fast_compare.
16742 (bpermd_<mode>): Change type attr to popcnt.
16743 (p8_fmrgow_<mode>): New insns for power8 direct move support.
16744 (p8_mtvsrwz_1): Likewise.
16745 (p8_mtvsrwz_2): Likewise.
16746 (reload_fpr_from_gpr<mode>): Likewise.
16747 (p8_mtvsrd_1): Likewise.
16748 (p8_mtvsrd_2): Likewise.
16749 (p8_xxpermdi_<mode>): Likewise.
16750 (reload_vsx_from_gpr<mode>): Likewise.
16751 (reload_vsx_from_gprsf): Likewise.
16752 (p8_mfvsrd_3_<mode>): LIkewise.
16753 (reload_gpr_from_vsx<mode>): Likewise.
16754 (reload_gpr_from_vsxsf): Likewise.
16755 (p8_mfvsrd_4_disf): Likewise.
16756 (multi-word GPR splits): Do not split direct moves or quad memory
16759 2013-06-10 David Malcolm <dmalcolm@redhat.com>
16761 * tree-into-ssa.c (interesting_blocks): Make static.
16763 2013-06-10 Jakub Jelinek <jakub@redhat.com>
16766 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
16767 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
16768 Use DATA_ABI_ALIGNMENT for that case instead if defined.
16769 (get_variable_align): New function.
16770 (get_variable_section, emit_bss, emit_common,
16771 assemble_variable_contents, place_block_symbol): Use
16772 get_variable_align instead of DECL_ALIGN.
16773 (assemble_noswitch_variable): Add align argument, use it
16774 instead of DECL_ALIGN.
16775 (assemble_variable): Adjust caller. Use get_variable_align
16776 instead of DECL_ALIGN.
16777 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
16779 (DATA_ABI_ALIGNMENT): Define.
16780 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
16781 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
16782 opt is false, only return the psABI mandated alignment increase.
16783 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
16784 (DATA_ABI_ALIGNMENT): ... this.
16785 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
16786 (DATA_ABI_ALIGNMENT): ... this.
16787 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
16788 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
16789 (DATA_ABI_ALIGNMENT): ... this.
16790 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
16791 * doc/tm.texi: Regenerated.
16793 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
16795 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
16796 cmp_code to construct REG_EQUAL note.
16798 2013-06-09 Jakub Jelinek <jakub@redhat.com>
16801 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
16802 that operands[2] doesn't overlap with operands[0].
16804 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
16805 Jan Hubicka <jh@suse.cz>
16807 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
16808 hack to mark symbols as used.
16810 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
16812 PR rtl-optimization/57559
16813 * lra-constraints.c (process_alt_operands): Don't discourage
16814 memory with known offset for offsetable memory constraint.
16815 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
16817 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
16819 * varasm.c (struct oc_local_state): Reorder fields.
16820 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
16821 and adjust accordingly.
16822 (output_constructor): Reorder initialization code and adjust call to
16823 output_constructor_bitfield.
16825 2013-06-07 Jan Hubicka <jh@suse.cz>
16827 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
16829 2013-06-07 David Malcolm <dmalcolm@redhat.com>
16831 * tree-object-size.c (unknown): Make const.
16833 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16835 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
16836 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
16837 for last alternative in the cpu_facility attribute.
16839 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16842 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
16843 (xordi3): Change operand 2 constraint to arm_xordi_operand.
16844 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
16845 * config/arm/constraints.md (Dg): New constraint.
16846 * config/arm/neon.md (xordi3_neon): Remove.
16847 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
16848 * config/arm/predicates.md (arm_xordi_operand): New predicate.
16850 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16852 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
16853 Clean up alternatives.
16855 2013-06-07 Alan Modra <amodra@gmail.com>
16857 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
16860 2013-06-07 Alan Modra <amodra@gmail.com>
16862 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
16864 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16866 * config/arm/constraints.md (Df): New constraint.
16867 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
16868 Correct length attribute for last two alternatives.
16870 2013-06-07 Alan Modra <amodra@gmail.com>
16872 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
16873 override user -mfp-in-toc.
16874 (offsettable_ok_by_alignment): Consider just the current access
16875 rather than the whole object, unless BLKmode. Handle
16876 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
16877 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
16878 for -mcmodel=medium.
16879 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
16880 override user -mfp-in-toc or -msum-in-toc. Default to
16881 -mno-fp-in-toc for -mcmodel=medium.
16883 2013-06-06 DJ Delorie <dj@redhat.com>
16885 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
16886 TARGET_VALID_POINTER_MODE.
16888 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
16889 Pat Haugen <pthaugen@us.ibm.com>
16890 Peter Bergner <bergner@vnet.ibm.com>
16892 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
16893 Document new power8 builtins.
16895 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
16896 condition code register, to allow 128-bit logical operations to be
16897 done in the VSX or GPR registers.
16898 (nor<mode>3): Use the canonical form for nor.
16899 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
16900 vclz*, and vpopcnt* vector instructions.
16901 (nand<mode>3): Likewise.
16902 (orc<mode>3): Likewise.
16903 (clz<mode>2): LIkewise.
16904 (popcount<mode>2): Likewise.
16906 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
16907 that only the GPRs are recognized.
16909 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16910 support for new power8 builtins.
16912 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
16914 (xscvdpspn): Likewise.
16920 (vpopcnt): Likewise.
16921 (vpopcntb): Likewise.
16922 (vpopcnth): Likewise.
16923 (vpopcntw): Likewise.
16924 (vpopcntd): Likewise.
16926 (vmrgew): Likewise.
16927 (vmrgow): Likewise.
16929 (eqv_v16qi3): Likewise.
16930 (eqv_v8hi3): Likewise.
16931 (eqv_v4si3): Likewise.
16932 (eqv_v2di3): Likewise.
16933 (eqv_v4sf3): Likewise.
16934 (eqv_v2df3): Likewise.
16936 (nand_v16qi3): Likewise.
16937 (nand_v8hi3): Likewise.
16938 (nand_v4si3): Likewise.
16939 (nand_v2di3): Likewise.
16940 (nand_v4sf3): Likewise.
16941 (nand_v2df3): Likewise.
16943 (orc_v16qi3): Likewise.
16944 (orc_v8hi3): Likewise.
16945 (orc_v4si3): Likewise.
16946 (orc_v2di3): Likewise.
16947 (orc_v4sf3): Likewise.
16948 (orc_v2df3): Likewise.
16950 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
16951 allow power8 quad mode in 64-bit.
16952 (rs6000_builtin_vectorized_function): Add support to vectorize
16953 ISA 2.07 count leading zeros, population count builtins.
16954 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
16955 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
16956 (builtin_function_type): Add vgbbd builtin function which takes an
16958 (altivec_expand_vec_perm_const): Add support for new power8 merge
16961 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
16962 that does not include TImdoe for use with 32-bit.
16963 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
16965 (UNSPEC_VSX_CVDPSPN): Likewise.
16966 (vsx_xscvdpspn): Likewise.
16967 (vsx_xscvspdpn): Likewise.
16968 (vsx_xscvdpspn_scalar): Likewise.
16969 (vsx_xscvspdpn_directmove): Likewise.
16970 (vsx_and<mode>3): Split logical operations into 32-bit and
16971 64-bit. Add support to do logical operations on TImode as well as
16972 VSX vector types. Allow logical operations to be done in either
16973 VSX registers or in general purpose registers in 64-bit mode. Add
16974 splitters if GPRs were used. For AND, add clobber of CCmode to
16975 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
16977 (vsx_and<mode>3_32bit): Likewise.
16978 (vsx_and<mode>3_64bit): Likewise.
16979 (vsx_ior<mode>3): Likewise.
16980 (vsx_ior<mode>3_32bit): Likewise.
16981 (vsx_ior<mode>3_64bit): Likewise.
16982 (vsx_xor<mode>3): Likewise.
16983 (vsx_xor<mode>3_32bit): Likewise.
16984 (vsx_xor<mode>3_64bit): Likewise.
16985 (vsx_one_cmpl<mode>2): Likewise.
16986 (vsx_one_cmpl<mode>2_32bit): Likewise.
16987 (vsx_one_cmpl<mode>2_64bit): Likewise.
16988 (vsx_nor<mode>3): Likewise.
16989 (vsx_nor<mode>3_32bit): Likewise.
16990 (vsx_nor<mode>3_64bit): Likewise.
16991 (vsx_andc<mode>3): Likewise.
16992 (vsx_andc<mode>3_32bit): Likewise.
16993 (vsx_andc<mode>3_64bit): Likewise.
16994 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
16995 and xxlorc instructions.
16996 (vsx_eqv<mode>3_64bit): Likewise.
16997 (vsx_nand<mode>3_32bit): Likewise.
16998 (vsx_nand<mode>3_64bit): Likewise.
16999 (vsx_orc<mode>3_32bit): Likewise.
17000 (vsx_orc<mode>3_64bit): Likewise.
17002 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17004 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17006 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17007 (p8_vmrgow): Likewise.
17008 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17009 GPRs to be split under VSX.
17010 (p8v_clz<mode>2): Add power8 count leading zero support.
17011 (p8v_popcount<mode>2): Add power8 population count support.
17012 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17015 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17018 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17020 (vec_nand): Likewise.
17021 (vec_vclz): Likewise.
17022 (vec_vclzb): Likewise.
17023 (vec_vclzd): Likewise.
17024 (vec_vclzh): Likewise.
17025 (vec_vclzw): Likewise.
17026 (vec_vgbbd): Likewise.
17027 (vec_vmrgew): Likewise.
17028 (vec_vmrgow): Likewise.
17029 (vec_vpopcnt): Likewise.
17030 (vec_vpopcntb): Likewise.
17031 (vec_vpopcntd): Likewise.
17032 (vec_vpopcnth): Likewise.
17033 (vec_vpopcntw): Likewise.
17035 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17037 PR rtl-optimization/57468
17038 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17041 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17043 PR rtl-optimization/57459
17044 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17045 type when setting live regs.
17047 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17049 * config/s390/s390.opt (mlra): New option.
17050 * config/s390/s390.c (s390_decompose_address): Check displacement
17051 for all registers for LRA.
17052 (s390_secondary_reload): Don't used secondary reloads for LRA.
17053 (s390_lra_p): New function.
17054 (TARGET_LRA_P): Define.
17055 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17056 of attribute cpu_facility to zarch for the last alternative.
17057 (*cmpmem_short): Ditto.
17059 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17061 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17062 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17063 (arm_expand_prologue): Likewise.
17065 2013-06-06 Teresa Johnson <tejohnson@google.com>
17068 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17069 as this is now done by redirect_edge_and_branch_force.
17070 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17071 barriers, and fix interaction with splitting.
17072 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17073 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17074 reflect changes made in the routine.
17075 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17076 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17077 since this is called in cfglayout mode, and replace partition fixup
17078 with assert as that is now done by force_nonfallthru_and_redirect.
17079 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17080 already be marked with region crossing note.
17081 (insert_section_boundary_note): Make non-static, gate on flag
17082 has_bb_partition, rewrite to also check for multiple partitions.
17083 (rest_of_handle_reorder_blocks): Remove call to
17084 insert_section_boundary_note, now done later during free_cfg.
17085 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17086 * bb-reorder.h (insert_section_boundary_note): Declare.
17087 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17088 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17089 invoke insert_section_boundary_note.
17090 (try_redirect_by_replacing_jump): Remove unnecessary
17091 check for region crossing note.
17092 (fixup_partition_crossing): New function.
17093 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17094 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17095 in non-cfglayout mode.
17096 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17097 remove old code that tried to do this. Emit barrier correctly
17098 when we are in cfglayout mode.
17099 (last_bb_in_partition): New function.
17100 (rtl_split_edge): Correctly fixup partition boundaries.
17101 (commit_one_edge_insertion): Remove old code that tried to
17102 fixup region crossing edge since this is now handled in
17103 split_block, and set up insertion point correctly since
17104 block may now end in a jump.
17105 (verify_hot_cold_block_grouping): Guard against checking when not in
17106 linearized RTL mode.
17107 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17109 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17110 rtl_verify_flow_info, so not called in cfglayout mode.
17111 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17112 (fixup_reorder_chain): Remove old code that attempted to fixup region
17113 crossing note as this is now handled in force_nonfallthru_and_redirect.
17114 (duplicate_insn_chain): Don't duplicate switch section notes.
17115 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17117 * basic-block.h (emit_barrier_after_bb): Declare.
17119 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17121 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17122 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17123 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17125 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17127 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17129 * config/arm/ldmstm.md: Regenerate.
17131 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17133 * config/arm/sync.md (atomic_loaddi_1):
17134 Disable predication for arm_restrict_it.
17135 (arm_load_exclusive<mode>): Likewise.
17136 (arm_load_exclusivesi): Likewise.
17137 (arm_load_exclusivedi): Likewise.
17138 (arm_load_acquire_exclusive<mode>): Likewise.
17139 (arm_load_acquire_exclusivesi): Likewise.
17140 (arm_load_acquire_exclusivedi): Likewise.
17141 (arm_store_exclusive<mode>): Likewise.
17142 (arm_store_exclusive<mode>): Likewise.
17143 (arm_store_release_exclusivedi): Likewise.
17144 (arm_store_release_exclusive<mode>): Likewise.
17146 2013-06-06 Richard Biener <rguenther@suse.de>
17148 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17150 (lto_tag_is_tree_code_p): Adjust.
17151 (lto_tag_is_gimple_code_p): Likewise.
17152 (lto_gimple_code_to_tag): Likewise.
17153 (lto_tag_to_gimple_code): Likewise.
17154 (lto_tree_code_to_tag): Likewise.
17155 (lto_tag_to_tree_code): Likewise.
17156 * data-streamer.h (streamer_write_hwi_in_range): Use
17157 uhwi streaming to stream the normalized range.
17158 (streamer_read_hwi_in_range): Likewise.
17160 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17162 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17163 (predicable_short_it): Likewise.
17164 (predicated): Likewise.
17165 (enabled): Handle above.
17166 (define_cond_exec): Set predicated attribute to yes.
17168 2013-06-05 Mike Stump <mikestump@comcast.net>
17170 * gdbinit.in (__FUNCTION__): Add.
17172 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17174 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17175 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17177 2013-06-05 Jan Hubicka <jh@suse.cz>
17179 * varasm.c (mark_decl_referenced): Revert the removal until targets
17182 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
17184 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
17185 instead of mark_decl_referenced.
17187 2013-06-05 Jan Hubicka <jh@suse.cz>
17189 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
17190 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
17191 and symtab_used_from_object_file_p.
17192 (cgraph_make_node_local_1): Clear forced_by_abi.
17193 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
17194 * cgraph.h (symtab_node_base): Add forced_by_abi.
17195 (decide_is_variable_needed): Remove.
17196 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
17197 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
17198 (decide_is_symbol_needed): ... this one; handle symbols in general;
17199 always analyze virtuals; honnor forced_by_abi.
17200 (cgraph_finalize_function): Update.
17201 (varpool_finalize_decl): Update.
17202 (symbol_defined_and_needed): Remove.
17203 (analyze_functions): Update.
17204 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17205 output_refs, input_overwrite_node): Handle forced_by_abi.
17206 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
17207 (address_taken_from_non_vtable_p): ... this one.
17208 (comdat_can_be_unshared_p_1): New function.
17209 (cgraph_comdat_can_be_unshared_p): Rename to ...
17210 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
17211 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
17212 (function_and_variable_visibility): Clear forced_by_abi as needed.
17213 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
17214 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
17215 * symtab.c (dump_symtab_base): Dump forced_by_abi.
17216 * varpool.c (decide_is_variable_needed): Remove.
17218 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17220 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
17221 (arm_option_override): Override arm_restrict_it where appropriate.
17222 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
17223 * config/arm/arm.opt (mrestrict-it): New command-line option.
17224 * doc/invoke.texi: Document -mrestrict-it.
17226 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17228 * tsan.c (tsan_atomic_table): Make const.
17230 2013-06-05 Richard Biener <rguenther@suse.de>
17232 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
17233 index associated with the tree we are supposed to replace.
17234 * tree-streamer-out.c (pack_ts_base_value_fields): Output
17235 TREE_ASM_WRITTEN as zero for everything but SSA names.
17237 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17239 * tree-ssa-structalias.c (call_stmt_vars): Make static.
17241 2013-06-04 Jan Hubicka <jh@suse.cz>
17243 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
17244 (input_node, input_varpool_node): Handle correctly external same
17246 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
17247 nodes at ltrans stage.
17249 2013-06-04 Jan Hubicka <jh@suse.cz>
17251 * ipa-inline.c (update_caller_keys): Fix availability test.
17252 (update_callee_keys): Likewise.
17253 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
17254 to follow ELF standard.
17256 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
17258 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
17259 (mips64r5900el-*-elf*): New configurations.
17260 * config/mips/mips-cpus.def (r5900): New processor.
17261 * config/mips/mips-tables.opt: Regenerate.
17262 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
17263 (mips_issue_rate): Handle PROCESSOR_R5900.
17264 (mips_reorg_process_insns): Force reorder mode for the R5900.
17265 * config/mips/mips.h (TARGET_MIPS5900): Define.
17266 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
17268 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
17270 * config/mips/mips.md (processor): Add r5900.
17271 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
17273 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17275 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
17276 into function to generate MOVI instruction.
17277 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
17278 (aarch64_preferred_simd_mode): Turn into wrapper.
17279 (aarch64_output_scalar_simd_mov_immediate): New function.
17280 * config/aarch64/aarch64-protos.h: Add prototype for above.
17282 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17284 * config/aarch64/aarch64.c (simd_immediate_info): Remove
17285 element_char member.
17286 (sizetochar): Return signed char.
17287 (aarch64_simd_valid_immediate): Remove elchar and other
17288 unnecessary variables.
17289 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
17290 Calculate element_char as required.
17291 * config/aarch64/aarch64-protos.h: Update and move prototype
17292 for aarch64_output_simd_mov_immediate.
17293 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
17296 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17298 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
17299 information completed by aarch64_simd_valid_immediate.
17300 (aarch64_legitimate_constant_p): Update arguments.
17301 (aarch64_simd_valid_immediate): Work with struct rather than many
17303 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
17304 (aarch64_simd_make_constant): Update arguments.
17305 (aarch64_output_simd_mov_immediate): Work with struct rather than
17306 many pointers. Output immediate directly rather than as operand.
17307 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17309 * config/aarch64/constraints.md (Dn): Update arguments.
17311 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17313 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
17315 (aarch64_simd_immediate_valid_for_move): Remove.
17316 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
17317 (aarch64_simd_make_constant): Update call.
17318 (aarch64_output_simd_mov_immediate): Update call.
17319 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17321 * config/aarch64/constraints.md (Dn): Update call.
17323 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17325 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
17326 return type to bool for prototype.
17327 (aarch64_legitimate_constant_p): Check for true instead of not -1.
17328 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
17329 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
17331 2013-06-04 Catherine Moore <clm@codesourcery.com>
17333 * config/mips/mips.opt (meva): New.
17334 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
17335 (ASM_SPEC): Handle -meva.
17336 * doc/invoke.texi (meva): Document.
17338 2013-06-04 Alan Modra <amodra@gmail.com>
17340 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
17343 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17345 * rtl.def: Add extra fourth optional field to define_cond_exec.
17346 * gensupport.c (process_one_cond_exec): Process attributes from
17348 * doc/md.texi: Document fourth field in define_cond_exec.
17350 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
17352 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
17353 out the processing order as in store_bit_field_1.
17355 2013-06-04 Jan Hubicka <jh@suse.cz>
17357 PR middle-end/57500
17358 * cgraphunit.c (cgraph_process_same_body_aliases): Create
17359 non-VAR_DECL node if it does not exist yet.
17361 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
17363 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
17364 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
17365 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
17366 target_cpu_default setting.
17368 2013-06-03 Teresa Johnson <tejohnson@google.com>
17370 * dumpfile.c (opt_info_switch_p): Change -fopt-info
17371 default to -fopt-info=optimized instead of all.
17372 * doc/invoke.texi: Ditto.
17373 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
17374 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
17375 (execute_vect_slp): Emit BB vectorization success under
17376 MSG_OPTIMIZED_LOCATIONS.
17377 * tree-vect-slp.c (vect_slp_transform_bb): Change
17378 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
17379 * tree-vect-loop.c (vect_transform_loop): Ditto.
17381 2013-06-03 Jason Merrill <jason@redhat.com>
17384 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
17385 Use TARGET_EXPR for C++.
17387 2013-06-03 Jakub Jelinek <jakub@redhat.com>
17389 PR rtl-optimization/57268
17390 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
17391 if DEBUG_INSN_P (insn).
17394 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
17396 PR rtl-optimization/57268
17397 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17398 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17400 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
17402 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
17403 (ix86_avoid_lea_for_addr): Likewise.
17404 (exact_dependency_1): Likewise.
17405 (ix86_adjust_cost): Likewise.
17406 (swap_top_of_ready_list): Fix formatting and !reload_completed check
17408 (do_reorder_for_imul): Fix typo, formatting and
17409 !reload_completed check removed.
17410 (ix86_sched_reorder): Fix typo and formatting.
17411 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
17414 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
17416 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
17418 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
17420 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
17421 <STRING_CST>: Likewise.
17422 <VECTOR_CST>: Likewise.
17424 2013-06-01 Janus Weil <janus@gcc.gnu.org>
17425 Mikael Morin <mikael@gcc.gnu.org>
17427 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
17428 * config.in: Regenerated.
17429 * configure: Regenerated.
17431 2013-06-01 Jan Hubicka <jh@suse.cz>
17433 PR middle-end/57366
17434 * cgraphunit.c (compile): When weakref is not supported,
17435 set up transparent aliases before final output pass.
17436 * varasm.c (assemble_alias): Do not try to do it here.
17438 2013-06-01 Jan Hubicka <jh@suse.cz>
17440 PR middle-end/57467
17441 * passes.c (for_per_function): Skip unanalyzed functions.
17443 2013-06-01 Jan Hubicka <jh@suse.cz>
17445 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
17446 (lto_symtab_merge_symbols_1): ... this one.
17447 (lto_symtab_merge_cgraph_nodes): Rename to ...
17448 (lto_symtab_merge_symbols): ... this one; simplify.
17449 * cgraph.c (same_body_aliases_done): Rename to ...
17450 (cpp_implicit_aliases_done): ... this one.
17451 (cgraph_create_function_alias): Update.
17452 (cgraph_same_body_alias): Update.
17453 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
17454 (verify_edge_corresponds_to_fndecl): Simplify.
17455 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
17456 (cgraph_node): Remove same_body_alias.
17457 (varpool_node): Remove alias_of and extra_name_alias.
17458 (same_body_aliases_done): Rename to ..
17459 (cpp_implicit_aliases_done): ... this one.
17460 (symtab_alias_ultimate_target): Add default parameter.
17461 (symtab_resolve_alias): New function.
17462 (fixup_same_cpp_alias_visibility): Declare.
17463 (cgraph_function_node): Add default parameter.
17464 (cgraph_node_asm_name): Likewise.
17465 (cgraph_function_or_thunk_node): Add default parameter; do
17466 not ICE when it is NULL.
17467 (varpool_variable_node): Likewise.
17468 * tree-emutls.c (create_emultls_var): Update.
17469 (ipa_lower_emutls): Update.
17470 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
17471 (cgraph_reset_node): Reset alias info.
17472 (cgraph_finalize_function): Update.
17473 (fixup_same_cpp_alias_visibility): Move to symtab.c.
17474 (analyze_function): Simplify.
17475 (cgraph_process_same_body_aliases): Simplify.
17476 (analyze_functions): Fixup same body aliases.
17477 (handle_alias_pairs): Simplify.
17478 (assemble_thunk): Update.
17479 (assemble_thunks_and_aliases): Update.
17480 (output_weakrefs): Rewrite.
17481 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
17482 (lto_output_varpool_node): Likewise.
17483 (compute_ltrans_boundary): Remve assert.
17484 (get_alias_symbol): New functoin.
17485 (input_node): Rewrite alias handling.
17486 (input_varpool_node): Likewise.
17487 * ipa-pure-const.c (propagate_pure_const): Fix formating.
17488 * ipa.c (process_references): Handle weakrefs correctly.
17489 (symtab_remove_unreachable_nodes): Likewise.
17490 * trans-mem.c (get_cg_data): Update.
17491 (ipa_tm_create_version_alias): Update.
17492 (ipa_tm_execute): Update.
17493 * symtab.c (dump_symtab_base): Dump aliases.
17494 (verify_symtab_base): Verify aliases.
17495 (symtab_node_availability): New function.
17496 (symtab_alias_ultimate_target): Simplify.
17497 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
17498 handle all the fixup cases.
17499 (symtab_resolve_alias): New function.
17500 * passes.c (ipa_write_summaries): Handle weakrefs.
17501 * varpool.c (varpool_analyze_node): Simplify.
17502 (assemble_aliases): Update.
17503 (varpool_create_variable_alias): Simplify.
17504 (varpool_extra_name_alias): Simplify.
17505 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
17506 (lto_symtab_merge_symbols): ... this one.
17508 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
17511 PR rtl-optimization/57268
17512 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17513 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17515 2013-06-01 Tobias Burnus <burnus@net-b.de>
17517 Partially reverted:
17518 2013-05-31 Tobias Burnus <burnus@net-b.de>
17520 PR middle-end/57073
17521 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
17524 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
17526 PR rtl-optimization/57268
17527 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17528 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17530 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
17532 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
17533 unordered comparison operators when -fno-trapping-math is in effect
17535 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
17536 and implement unordered comparison operators properly on the e500.
17538 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
17540 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
17541 for constant scalar integers.
17542 (simplify_relational_operation_1): Likewise.
17544 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
17546 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
17547 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
17550 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
17551 Igor Zamyatin <igor.zamyatin@intel.com>
17553 Silvermont (SLM) architecture performance tuning.
17554 * config/i386/i386.h (enum ix86_tune_indices): Add
17555 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
17556 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
17558 * config/i386/i386.c (initial_ix86_tune_features)
17559 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
17560 (ix86_lea_outperforms): Handle Silvermont tuning.
17561 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
17563 (ix86_use_lea_for_mov): Likewise.
17564 (ix86_avoid_lea_for_addr): Likewise.
17565 (ix86_lea_for_add_ok): Likewise.
17566 (exact_dependency_1): New function.
17567 (exact_store_load_dependency): Likewise.
17568 (ix86_adjust_cost): Handle Silvermont tuning.
17569 (do_reoder_for_imul): Likewise.
17570 (swap_top_of_ready_list): New function.
17571 (ix86_sched_reorder): Changed to handle Silvermont tuning.
17573 * config/i386/i386.md (peepholes that split memory operand in fp
17576 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
17578 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
17579 Remove un-necessary braces.
17581 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
17583 * config/aarch64/aarch64.c (aarch64_classify_symbol):
17584 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
17586 2013-05-31 Tobias Burnus <burnus@net-b.de>
17588 PR middle-end/57073
17589 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
17591 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17594 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
17595 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
17596 * config/arm/neon.md (iordi3_neon): Remove.
17597 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
17598 * config/arm/predicates.md (imm_for_neon_logic_operand):
17599 Move to earlier in the file.
17600 (neon_logic_op2): Likewise.
17601 (arm_iordi_operand_neon): New predicate.
17603 2013-05-31 Richard Biener <rguenther@suse.de>
17605 PR tree-optimization/57478
17606 PR tree-optimization/57453
17607 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
17610 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17612 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
17613 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
17615 2013-05-30 Tobias Burnus <burnus@net-b.de>
17616 Thomas Koenig <tkoenig@gcc.gnu.org>
17618 PR middle-end/57073
17619 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
17620 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
17622 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
17624 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
17626 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
17628 * target.def (register_usage_leveling_p): New hook.
17629 * targhooks.c (default_register_usage_leveling_p): New.
17630 * targhooks.h (default_register_usage_leveling_p): New prototype.
17631 * lra-assigns.c (register_usage_leveling_p): Use the hook.
17632 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
17633 * doc/tm.texi: Update.
17634 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
17636 2013-05-30 Ian Bolton <ian.bolton@arm.com>
17638 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
17639 (*insv_reg<mode>): New define_insn.
17641 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
17643 PR rtl-optimization/57439
17644 * postreload.c (move2add_valid_value_p): Check that we have
17645 a zero subreg_regno_offset when accessing the register in
17646 the requested mode.
17648 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
17649 Igor Zamyatin <igor.zamyatin@intel.com>
17651 Silvermont (SLM) architecture pipeline model, tuning and
17653 * config.gcc: Add slm config options and target.
17655 * config/i386/slm.md: New.
17657 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
17659 * config/i386/i386-c.c (ix86_target_macros_internal): New case
17661 (ix86_target_macros_internal): Likewise.
17663 * config/i386/i386.c (slm_cost): New cost.
17664 (m_SLM): New macro flag.
17665 (initial_ix86_tune_features): Set m_SLM.
17666 (x86_accumulate_outgoing_args): Likewise.
17667 (x86_arch_always_fancy_math_387): Likewise.
17668 (processor_target_table): Add slm cost.
17669 (cpu_names): Add slm cpu name.
17670 (x86_option_override_internal): Set SLM ISA.
17671 (ix86_issue_rate): New case PROCESSOR_SLM.
17672 (ia32_multipass_dfa_lookahead): Likewise.
17673 (fold_builtin_cpu): Add slm.
17675 * config/i386/i386.h (TARGET_SLM): New target macro.
17676 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
17677 (processor_type): Add PROCESSOR_SLM.
17679 * config/i386/i386.md (cpu): Add new value "slm".
17680 (slm.md): Include slm.md.
17682 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
17683 Zhenqiang Chen <zhenqiang.chen@linaro.org>
17685 * config/arm/arm-protos.h: Add and update function protos.
17686 * config/arm/arm.c (use_simple_return_p): New added.
17687 (thumb2_expand_return): Check simple_return flag.
17688 * config/arm/arm.md: Add simple_return and conditional simple_return.
17689 * config/arm/iterators.md: Add iterator for return and simple_return.
17691 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
17693 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
17694 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
17695 (arm_emit_vfp_multi_reg_pop): Likewise.
17696 (thumb2_emit_ldrd_pop): Likewise.
17697 (arm_expand_epilogue): Add misc REG_CFA notes.
17698 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
17700 2013-05-29 Lawrence Crowl <crowl@google.com>
17702 * config/arm/t-arm: Update for below.
17704 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
17705 Change type to hash_table. Update dependent calls and types.
17707 * config/i386/t-cygming: Update for below.
17709 * config/i386/t-interix: Update for below.
17711 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
17712 Change type to hash_table. Update dependent calls and types.
17713 (i386_find_on_wrapper_list::wrappers): Likewise.
17715 * config/ia64/t-ia64: Update for below.
17717 * config/ia64/ia64.c (bundle_state_table):
17718 Change type to hash_table. Update dependent calls and types.
17720 * config/mips/mips.c (mips_reorg_process_insns::htab):
17721 Change type to hash_table. Update dependent calls and types.
17723 * config/sol2.c (solaris_comdat_htab):
17724 Change type to hash_table. Update dependent calls and types.
17726 * config/t-sol2: Update for above.
17728 2013-05-29 Teresa Johnson <tejohnson@google.com>
17730 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
17731 functions are not yet marked as defined.
17733 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
17734 Pat Haugen <pthaugen@us.ibm.com>
17735 Peter Bergner <bergner@vnet.ibm.com>
17737 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
17741 (vrotl<mode>3): Likewise.
17742 (vashl<mode>3): Likewise.
17743 (vlshr<mode>3): Likewise.
17744 (vashr<mode>3): Likewise.
17746 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17747 support for power8 V2DI builtins.
17749 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
17750 power8 V2DI builtins.
17751 (vupkhsw): Likewise.
17752 (vupklsw): Likewise.
17753 (vaddudm): Likewise.
17754 (vminsd): Likewise.
17755 (vmaxsd): Likewise.
17756 (vminud): Likewise.
17757 (vmaxud): Likewise.
17758 (vpkudum): Likewise.
17759 (vpksdss): Likewise.
17760 (vpkudus): Likewise.
17761 (vpksdus): Likewise.
17766 (vsubudm): Likewise.
17767 (vcmpequd): Likewise.
17768 (vcmpgtsd): Likewise.
17769 (vcmpgtud): Likewise.
17770 (vcmpequd_p): Likewise.
17771 (vcmpgtsd_p): Likewise.
17772 (vcmpgtud_p): Likewise.
17773 (vupkhsw): Likewise.
17774 (vupklsw): Likewise.
17775 (vaddudm): Likewise.
17776 (vmaxsd): Likewise.
17777 (vmaxud): Likewise.
17778 (vminsd): Likewise.
17779 (vminud): Likewise.
17780 (vpksdss): Likewise.
17781 (vpksdus): Likewise.
17782 (vpkudum): Likewise.
17783 (vpkudus): Likewise.
17788 (vsubudm): Likewise.
17790 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
17791 support for power8 V2DI instructions.
17793 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
17794 power8 V2DI instructions. Combine pack and unpack insns to use an
17795 iterator for each mode. Check whether a particular mode supports
17796 Altivec instructions instead of just checking TARGET_ALTIVEC.
17797 (UNSPEC_VPKUWUM): Likewise.
17798 (UNSPEC_VPKSHSS): Likewise.
17799 (UNSPEC_VPKSWSS): Likewise.
17800 (UNSPEC_VPKUHUS): Likewise.
17801 (UNSPEC_VPKSHUS): Likewise.
17802 (UNSPEC_VPKUWUS): Likewise.
17803 (UNSPEC_VPKSWUS): Likewise.
17804 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
17805 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
17806 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
17807 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
17808 (UNSPEC_VUPKHSB): Likewise.
17809 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
17810 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
17811 (UNSPEC_VUPKHSH): Likewise.
17812 (UNSPEC_VUPKLSB): Likewise.
17813 (UNSPEC_VUPKLSH): Likewise.
17815 (VI_char): Likewise.
17816 (VI_scalar): Likewise.
17817 (VI_unit): Likewise.
17819 (VP_small): Likewise.
17820 (VP_small_lc): Likewise.
17821 (VU_char): Likewise.
17822 (add<mode>3): Likewise.
17823 (altivec_vaddcuw): Likewise.
17824 (altivec_vaddu<VI_char>s): Likewise.
17825 (altivec_vadds<VI_char>s): Likewise.
17826 (sub<mode>3): Likewise.
17827 (altivec_vsubcuw): Likewise.
17828 (altivec_vsubu<VI_char>s): Likewise.
17829 (altivec_vsubs<VI_char>s): Likewise.
17830 (altivec_vavgs<VI_char>): Likewise.
17831 (altivec_vcmpbfp): Likewise.
17832 (altivec_eq<mode>): Likewise.
17833 (altivec_gt<mode>): Likewise.
17834 (altivec_gtu<mode>): Likewise.
17835 (umax<mode>3): Likewise.
17836 (smax<mode>3): Likewise.
17837 (umin<mode>3): Likewise.
17838 (smin<mode>3): Likewise.
17839 (altivec_vpkuhum): Likewise.
17840 (altivec_vpkuwum): Likewise.
17841 (altivec_vpkshss): Likewise.
17842 (altivec_vpkswss): Likewise.
17843 (altivec_vpkuhus): Likewise.
17844 (altivec_vpkshus): Likewise.
17845 (altivec_vpkuwus): Likewise.
17846 (altivec_vpkswus): Likewise.
17847 (altivec_vpks<VI_char>ss): Likewise.
17848 (altivec_vpks<VI_char>us): Likewise.
17849 (altivec_vpku<VI_char>us): Likewise.
17850 (altivec_vpku<VI_char>um): Likewise.
17851 (altivec_vrl<VI_char>): Likewise.
17852 (altivec_vsl<VI_char>): Likewise.
17853 (altivec_vsr<VI_char>): Likewise.
17854 (altivec_vsra<VI_char>): Likewise.
17855 (altivec_vsldoi_<mode>): Likewise.
17856 (altivec_vupkhsb): Likewise.
17857 (altivec_vupkhs<VU_char>): Likewise.
17858 (altivec_vupkls<VU_char>): Likewise.
17859 (altivec_vupkhsh): Likewise.
17860 (altivec_vupklsb): Likewise.
17861 (altivec_vupklsh): Likewise.
17862 (altivec_vcmpequ<VI_char>_p): Likewise.
17863 (altivec_vcmpgts<VI_char>_p): Likewise.
17864 (altivec_vcmpgtu<VI_char>_p): Likewise.
17865 (abs<mode>2): Likewise.
17866 (vec_unpacks_hi_v16qi): Likewise.
17867 (vec_unpacks_hi_v8hi): Likewise.
17868 (vec_unpacks_lo_v16qi): Likewise.
17869 (vec_unpacks_hi_<VP_small_lc>): Likewise.
17870 (vec_unpacks_lo_v8hi): Likewise.
17871 (vec_unpacks_lo_<VP_small_lc>): Likewise.
17872 (vec_pack_trunc_v8h): Likewise.
17873 (vec_pack_trunc_v4si): Likewise.
17874 (vec_pack_trunc_<mode>): Likewise.
17876 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
17878 (vec_vmaxsd): Likewise.
17879 (vec_vmaxud): Likewise.
17880 (vec_vminsd): Likewise.
17881 (vec_vminud): Likewise.
17882 (vec_vpksdss): Likewise.
17883 (vec_vpksdus): Likewise.
17884 (vec_vpkudum): Likewise.
17885 (vec_vpkudus): Likewise.
17886 (vec_vrld): Likewise.
17887 (vec_vsld): Likewise.
17888 (vec_vsrad): Likewise.
17889 (vec_vsrd): Likewise.
17890 (vec_vsubudm): Likewise.
17891 (vec_vupkhsw): Likewise.
17892 (vec_vupklsw): Likewise.
17894 2013-05-29 Jan Hubicka <jh@suse.cz>
17896 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
17897 flags; reorder rest of fields in more consistent way.
17898 (varpool_node): Remove analyzed, finalized and alias.
17899 (cgraph_ndoe): Likewise.
17900 (symtab_alias_ultimate_target): New function.
17901 (cgraph_function_node): Move offline.
17902 (cgraph_reset_node): Declare.
17903 (cgraph_comdat_can_be_unshared_p): Remove.
17904 (varpool_remove_initializer): Declare.
17905 (varpool_first_defined_variable, varpool_next_defined_variable
17906 cgraph_first_defined_function, cgraph_next_defined_function): Update.
17907 (cgraph_function_with_gimple_body_p): Update.
17908 (varpool_all_refs_explicit_p): Update.
17909 (symtab_alias_target): New function.
17910 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
17911 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
17912 (cgraph_function_or_thunk_node): Simplify using
17913 symtab_alias_ultimate_target.
17914 (varpool_variable_node): Likewise.
17915 * cgraph.c (cgraph_create_function_alias): Update.
17916 (cgraph_add_thunk): Update.
17917 (cgraph_remove_node): Update.
17918 (dump_cgraph_node): Do not dump removed flags.
17919 (cgraph_function_body_availability): Update.
17920 (cgraph_propagate_frequency): Update.
17921 (verify_cgraph_node): Check sanity of local flag.
17922 (cgraph_function_node): Move here from cgraph.h; revamp for
17923 cgraph_function_or_thunk_node.
17924 * lto-symtab.c (lto_varpool_replace_node): Update.
17925 (lto_symtab_resolve_can_prevail_p): Update.
17926 (lto_symtab_merge_cgraph_nodes): Update.
17927 * ipa-cp.c (determine_versionability, initialize_node_lattices,
17928 propagate_constants_accross_call, devirtualization_time_bonus,
17929 ipcp_propagate_stage): Update.
17930 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
17931 * ipa-inline-transform.c (clone_inlined_nodes,
17932 preserve_function_body_p): Update.
17933 * ipa-reference.c (propagate): Update.
17934 (write_node_summary_p): Update.
17935 * toplev.c (wrapup_global_declaration_2): Update.
17936 * cgraphunit.c (cgraph_analyze_function): Rename to ...
17937 (analyze_function) ... this one.
17938 (cgraph_process_new_functions): Update.
17939 (cgraph_reset_node): Export.
17940 (cgraph_finalize_function): Update.
17941 (cgraph_add_new_function): Update.
17942 (process_function_and_variable_attributes): Update.
17943 (varpool_finalize_decl): Update.
17944 (symbol_finalized): Remove.
17945 (symbol_finalized_and_needed): Rename to ...
17946 (symbol_defined_and_needed): ... update.
17947 (cgraph_analyze_functions): Update.
17948 (handle_alias_pairs): Update.
17949 (mark_functions_to_output): Update.
17950 (assemble_thunk): Update.
17951 (output_in_order): Update.
17952 (output_weakrefs): Update.
17953 (finalize_compilation_unit): Update.
17954 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
17955 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
17956 input_node, input_varpool_node): Update.
17957 * dbxout.c (dbxout_expand_expr): Update.
17958 * cgraphclones.c (cgraph_clone_node): Update.
17959 (cgraph_copy_node_for_versioning): Update.
17960 (cgraph_materialize_clone): Update.
17961 (cgraph_materialize_all_clones): Update.
17962 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
17963 propagate_pure_const, propagate_nothrow): Update.
17964 * lto-streamer-out.c (lto_output, write_symbol): Update.
17965 * ipa-utils.c (ipa_reverse_postorder): Update.
17966 * ipa-inline.c (can_inline_edge_p): Update.
17967 (update_caller_keys, ipa_inline): Update.
17968 * dwarf2out.c (reference_to_unused,
17969 premark_types_used_by_global_vars_helper): Update.
17970 * tree-eh.c (tree_could_trap_p): Update.
17971 * ipa-split.c (consider_split, execute_split_functions): Update.
17972 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
17973 has_addr_references_p): Update; move ahead in file for better
17975 (process_references): Simplify.
17976 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
17977 bodies are removed.
17978 (cgraph_comdat_can_be_unshared_p): Make static.
17979 (cgraph_externally_visible_p): Update.
17980 (varpool_externally_visible_p): Update.
17981 (function_and_variable_visibility): Update.
17982 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
17983 ipa_tm_mark_force_output_node): Update.
17984 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
17985 estimate_edge_devirt_benefit, inline_generate_summary,
17986 inline_write_summary): Update.
17987 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
17988 * ipa-prop.c (ipa_compute_jump_functions): Update.
17989 (ipa_print_node_params, ipa_prop_read_section,
17990 ipa_update_after_lto_read, read_replacements_section): Update.
17991 * varasm.c (mark_decl_referenced): Update.
17992 (assemble_alias, dump_tm_clone_pairs): Update.
17993 * tree-inline.c (copy_bb): Update.
17994 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
17996 * symtab.c (dump_symtab_base): Print new flags.
17997 (verify_symtab_base): Verify new flags.
17998 (symtab_alias_ultimate_target): New function.
17999 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18000 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18002 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18004 * i386.c (ix86_get_function_versions_dispatcher,
18005 ix86_generate_version_dispatcher_body): Update.
18006 (fold_builtin_cpu): Use varpool_add_new_variable.
18007 * varpool.c (varpool_remove_initializer): Break out from ...
18008 (varpool_remove_node): ... this one.
18009 (dump_varpool_node, varpool_node_for_asm,
18010 cgraph_variable_initializer_availability, varpool_analyze_node,
18011 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18012 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18015 2013-05-29 Jan Hubicka <jh@suse.cz>
18017 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18019 2013-05-29 Easwaran Raman <eraman@google.com>
18021 PR tree-optimization/57442
18022 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18023 when control exits the main loop.
18025 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18027 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18029 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18030 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18031 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18034 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18036 PR tree-optimization/57441
18037 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18038 Don't limit size of incr_vec to number of candidates.
18040 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18042 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18043 and mips16 directories.
18044 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18045 (MULTILIB_DIRNAMES): Ditto.
18046 (MULTILIB_EXCEPTIONS): Add new exceptions.
18047 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18048 (MULTILIB_DIRNAMES): Ditto.
18049 (MULTILIB_EXCEPTIONS): Add new exceptions.
18051 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18052 Marcus Shawcroft <marcus.shawcroft@arm.com>
18054 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18055 SYMBOL_TINY_ABSOLUTE.
18056 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18057 SYMBOL_TINY_ABSOLUTE.
18058 (aarch64_expand_mov_immediate): Likewise.
18059 (aarch64_classify_symbol): Likewise.
18060 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18061 Permit SYMBOL_TINY_ABSOLUTE.
18062 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18064 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18065 Marcus Shawcroft <marcus.shawcroft@arm.com>
18067 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18068 Refactor if/switch. Replace gcc_assert with if.
18070 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18072 * config/i386/i386.c (initial_ix86_tune_features): Enable
18073 FP Reassociation for AMD bdver1 and bdver2.
18075 2013-05-29 Martin Jambor <mjambor@suse.cz>
18077 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18078 and IMAGPART_EXPR do not occur within other handled_components.
18080 2013-05-29 Richard Biener <rguenther@suse.de>
18082 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18083 access on whether the use is in the BB we currently try to
18085 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18086 vectorize to vect_bb_slp_scalar_cost.
18088 2013-05-29 Richard Biener <rguenther@suse.de>
18090 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18091 computing scalar cost offsetted by stmts that are kept live
18093 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18094 for computation of scalar cost.
18096 2013-05-28 Steve Ellcey <sellcey@mips.com>
18098 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18100 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18102 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18103 array notation built-in reduction functions.
18104 * doc/passes.texi (Passes): Added documentation about changes done
18106 * doc/invoke.texi (C Dialect Options): Added documentation about
18107 the -fcilkplus flag.
18108 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18109 (BUILTINS_DEF): Depend on cilkplus.def.
18110 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18111 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18112 * cilkplus.def: New file.
18114 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18116 PR rtl-optimization/57439
18117 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18119 2013-05-28 Easwaran Raman <eraman@google.com>
18121 PR tree-optimization/57337
18122 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18123 (find_insert_point): Correctly identify the insertion point
18124 when two statements with the same UID is compared.
18126 2013-05-28 Richard Biener <rguenther@suse.de>
18128 PR tree-optimization/56787
18129 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18130 from the list of data references.
18131 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18133 (vect_analyze_loop_operations): Likewise.
18134 (vect_transform_loop): Remove clobbers.
18136 2013-05-28 Martin Jambor <mjambor@suse.cz>
18138 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18139 and REALPART_EXPRs have scalar type.
18141 2013-05-28 Richard Biener <rguenther@suse.de>
18143 PR tree-optimization/57411
18144 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18146 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18147 virtual operand propagation.
18149 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18151 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18152 destination register for bmasksi_vis.
18153 (vector_init_bshuffle): Likewise.
18154 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18156 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18158 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18159 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18160 mode if the instruction isn't available in the original mode.
18161 * config/sparc/sparc.opt (mfix-ut699): New option.
18162 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18163 (divdf3): Turn into expander.
18164 (divdf3_nofix): New insn.
18165 (divdf3_fix): Likewise.
18166 (divsf3): Disable if -mfix-ut699.
18167 (sqrtdf2): Turn into expander.
18168 (sqrtdf2_nofix): New insn.
18169 (sqrtdf2_fix): Likewise.
18170 (sqrtsf2): Disable if -mfix-ut699.
18172 2013-05-27 Richard Biener <rguenther@suse.de>
18174 PR middle-end/57412
18175 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18176 block for the new loop.
18178 2013-05-27 Richard Biener <rguenther@suse.de>
18180 PR tree-optimization/57343
18181 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
18182 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
18183 (number_of_iterations_cond): Do not build the folded tree.
18185 2013-05-27 Richard Biener <rguenther@suse.de>
18188 PR middle-end/57381
18189 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18190 OEP_CONSTANT_ADDRESS_OF retained.
18192 PR tree-optimization/57417
18193 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
18194 for unchanged base.
18195 (set_ssa_val_to): Compare addresses using
18196 get_addr_base_and_unit_offset.
18198 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
18200 PR rtl-optimization/56833
18201 * postreload.c (move2add_record_mode): New function.
18202 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
18203 (move2add_use_add2_insn): Use move2add_record_sym_value.
18204 (move2add_use_add3_insn): Likewise.
18205 (reload_cse_move2add): Use move2add_valid_value_p and
18206 move2add_record_mode. Invalidate call-clobbered and REG_INC
18207 affected regs by setting reg_mode to VOIDmode.
18208 (move2add_note_store): Don't pretend the inside of a SUBREG is
18209 the actual destination. Invalidate single/leading registers by
18210 setting reg_mode to VOIDmode.
18211 Use move2add_record_sym_value, move2add_valid_value_p and
18212 move2add_record_mode.
18214 2013-05-27 Richard Biener <rguenther@suse.de>
18216 PR tree-optimization/57396
18217 * tree-affine.c (double_int_constant_multiple_p): Properly
18218 return false for val == 0 and div != 0.
18220 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18222 * config/mips/mips.h: Use #elif in preprocessor conditions.
18224 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18227 * config/mips/constraints.md (kl): New constraint.
18228 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
18229 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
18230 constraint for operand 0. Split after CSE for MIPS16. Emit a move
18231 from LO for MIPS16.
18232 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
18234 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18237 * config/mips/mips.c (mips_can_inline_p): New function.
18238 (TARGET_CAN_INLINE_P): Define.
18240 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
18242 * sched-int.h (ds_t, dw_t): Make unsigned int.
18243 Fix documentation that describes how all the ds_t bits are used.
18244 Reserve the last bit for delayed-branch scheduling.
18245 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
18246 (BITS_PER_DEP_WEAK): Fix definition and documentation.
18247 (gen_dep_weak_1): Remove prototype.
18248 * sched-deps.c (get_dep_weak_1): Make static.
18249 * target.def (speculate_insn, needs_block_p, gen_spec_check,
18250 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
18251 * doc/tm.texi: Regenerate.
18252 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
18254 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
18257 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
18259 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
18260 Sandra Loosemore <sandra@codesourcery.com>
18262 * config.gcc (powerpc-*): Allow native for with-cpu.
18264 2013-05-24 Jeff Law <law@redhat.com>
18266 PR tree-optimization/57124
18267 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
18268 conversion feeding a condition if the range has an overflow
18269 if -fstrict-overflow. Add warnings for when we do make the
18272 2013-05-24 Dehao Chen <dehao@google.com>
18274 * tree-cfg.c (locus_discrim_map): Fix the typo.
18275 (locus_discrim_hasher): Likewise.
18276 (locus_discrim_hasher::hash): Likewise.
18277 (locus_discrim_hasher::equal): Likewise.
18279 2013-05-24 Martin Jambor <mjambor@suse.cz>
18281 PR tree-optimization/57294
18282 * cgraph.h (ipa_record_stmt_references): Declare.
18283 * cgraphbuild.c (ipa_record_stmt_references): New function.
18284 (build_cgraph_edges): Use ipa_record_stmt_references.
18285 (rebuild_cgraph_edges): Likewise.
18286 (cgraph_rebuild_references): Likewise.
18287 * ipa-prop.c (ipa_modify_call_arguments): Discard references
18288 associated with the old statement and build references from the
18289 newly built statements.
18290 * ipa-ref.c (ipa_remove_stmt_references): New function.
18291 * ipa-ref.h (ipa_remove_stmt_references): Declare.
18293 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
18295 * lra-constraints.c (emit_spill_move): Use smaller mode for
18297 (check_and_process_move): Consider mem-reg moves for secondary
18299 (curr_insn_transform): Don't lose insns emitted before for
18300 secondary memory moves.
18301 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
18302 reg set up in the current insn.
18304 2013-05-24 Dehao Chen <dehao@google.com>
18306 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
18308 (locus_descrim_hasher::equal): Likewise.
18309 (build_gimple_cfg): New discriminator assignment algorithm.
18310 (make_edges): Likewise.
18311 (next_discriminator_for_locus): Likewise.
18312 (same_line_p): Likewise.
18313 (assign_discriminators): Likewise.
18314 (make_cond_expr_edges): Likewise.
18315 (make_gimple_switch_edges): Likewise.
18316 (make_goto_expr_edges): Likewise.
18317 (make_gimple_asm_edges): Likewise.
18319 2013-05-24 Ian Bolton <ian.bolton@arm.com>
18321 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
18322 X format specifier to only display bottom 16 bits.
18323 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
18324 immediate to match for operand 2, since it will be masked.
18326 2013-05-24 Richard Biener <rguenther@suse.de>
18328 PR tree-optimization/57287
18329 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
18330 all SSA names that occur in abnormal PHIs.
18332 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
18335 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
18336 that index is not negative.
18338 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
18340 PR rtl-optimization/55177
18341 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
18342 (simplify_byte_swapping_operation): New.
18343 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
18344 (simplify_relational_operation_1): Deal with BSWAP.
18346 2013-05-23 Richard Henderson <rth@redhat.com>
18349 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
18350 (ix86_reorg): Call it.
18352 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
18355 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
18356 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
18357 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
18359 2013-05-23 Christian Bruel <christian.bruel@st.com>
18362 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
18364 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18365 Marcus Shawcroft <marcus.shawcroft@arm.com>
18367 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
18368 * config/aarch64/constraints.md (Usa): Remove.
18369 * doc/md.texi (AArch64 Usa): Remove.
18371 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18372 Marcus Shawcroft <marcus.shawcroft@arm.com>
18374 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
18375 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
18376 * config/aarch64/predicates.md (aarch64_const_address): Remove.
18377 (aarch64_mov_operand): Use aarch64_mov_operand_p.
18379 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
18381 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
18382 instruction (AdvSIMD).
18383 * config/aarch64/aarch64-builtins.c
18384 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
18385 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
18387 2013-05-23 Martin Jambor <mjambor@suse.cz>
18389 PR middle-end/57347
18390 * tree.h (contains_bitfld_component_ref_p): Declare.
18391 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
18392 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
18394 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
18395 not access a bit-field. Assert all final offsets are byte-aligned.
18397 2013-05-23 Richard Biener <rguenther@suse.de>
18399 PR tree-optimization/57380
18400 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
18401 least one invariant or re-used load.
18402 * passes.c (init_optimization_passes): Move pass_phiprop before
18405 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
18407 * config/aarch64/aarch64-simd.md
18408 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
18410 2013-05-23 Richard Biener <rguenther@suse.de>
18412 PR middle-end/57381
18413 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18414 OEP_CONSTANT_ADDRESS_OF retained.
18416 2013-05-23 Jakub Jelinek <jakub@redhat.com>
18418 PR middle-end/57344
18419 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
18420 don't lower unit. Handle unit not being always BITS_PER_WORD.
18422 2013-05-23 Richard Biener <rguenther@suse.de>
18424 PR rtl-optimization/57341
18425 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
18426 instead of true_dependence.
18428 2013-05-22 David Malcolm <dmalcolm@redhat.com>
18430 * bb-reorder.c (branch_threshold): Make const.
18431 (exec_threshold): Ditto.
18433 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
18434 Pat Haugen <pthaugen@us.ibm.com>
18435 Peter Bergner <bergner@vnet.ibm.com>
18437 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
18438 documentation for the power8 crypto builtins.
18440 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
18442 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
18443 macros for defining power8 builtin functions.
18444 (BU_P8V_AV_2): Likewise.
18445 (BU_P8V_AV_P): Likewise.
18446 (BU_P8V_VSX_1): Likewise.
18447 (BU_P8V_OVERLOAD_1): Likewise.
18448 (BU_P8V_OVERLOAD_2): Likewise.
18449 (BU_CRYPTO_1): Likewise.
18450 (BU_CRYPTO_2): Likewise.
18451 (BU_CRYPTO_3): Likewise.
18452 (BU_CRYPTO_OVERLOAD_1): Likewise.
18453 (BU_CRYPTO_OVERLOAD_2): Likewise.
18454 (XSCVSPDP): Fix typo, point to the correct instruction.
18455 (VCIPHER): Add power8 crypto builtins.
18456 (VCIPHERLAST): Likewise.
18457 (VNCIPHER): Likewise.
18458 (VNCIPHERLAST): Likewise.
18459 (VPMSUMB): Likewise.
18460 (VPMSUMH): Likewise.
18461 (VPMSUMW): Likewise.
18462 (VPERMXOR_V2DI): Likewise.
18463 (VPERMXOR_V4SI: Likewise.
18464 (VPERMXOR_V8HI: Likewise.
18465 (VPERMXOR_V16QI: Likewise.
18466 (VSHASIGMAW): Likewise.
18467 (VSHASIGMAD): Likewise.
18468 (VPMSUM): Likewise.
18469 (VPERMXOR): Likewise.
18470 (VSHASIGMA): Likewise.
18472 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
18473 __CRYPTO__ if the crypto instructions are available.
18474 (altivec_overloaded_builtins): Add support for overloaded power8
18477 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
18478 support for power8 crypto builtins.
18479 (builtin_function_type): Likewise.
18480 (altivec_init_builtins): Add support for builtins that take vector
18481 long long (V2DI) arguments.
18483 * config/rs6000/crypto.md: New file, define power8 crypto
18486 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
18487 Pat Haugen <pthaugen@us.ibm.com>
18488 Peter Bergner <bergner@vnet.ibm.com>
18490 * doc/invoke.texi (Option Summary): Add power8 options.
18491 (RS/6000 and PowerPC Options): Likewise.
18493 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
18494 constraints.md instead of rs6000.h. Reorder w* constraints. Add
18495 wm, wn, wr documentation.
18497 * config/rs6000/constraints.md (wm): New constraint for VSX
18498 registers if direct move instructions are enabled.
18499 (wn): New constraint for no registers.
18500 (wq): New constraint for quad word even GPR registers.
18501 (wr): New constraint if 64-bit instructions are enabled.
18502 (wv): New constraint if power8 vector instructions are enabled.
18503 (wQ): New constraint for quad word memory locations.
18505 * config/rs6000/predicates.md (const_0_to_15_operand): New
18506 constraint for 0..15 for crypto instructions.
18507 (gpc_reg_operand): If VSX allow registers in VSX registers as well
18508 as GPR and floating point registers.
18509 (int_reg_operand): New predicate to match only GPR registers.
18510 (base_reg_operand): New predicate to match base registers.
18511 (quad_int_reg_operand): New predicate to match even GPR registers
18512 for quad memory operations.
18513 (vsx_reg_or_cint_operand): New predicate to allow vector logical
18514 operations in both GPR and VSX registers.
18515 (quad_memory_operand): New predicate for quad memory operations.
18516 (reg_or_indexed_operand): New predicate for direct move support.
18518 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
18519 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
18520 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
18521 (POWERPC_MASKS): Add power8 options.
18522 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
18525 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
18526 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
18528 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
18529 (-mpower8-fusion): New power8 options.
18530 (-mpower8-fusion-sign): Likewise.
18531 (-mpower8-vector): Likewise.
18532 (-mcrypto): Likewise.
18533 (-mdirect-move): Likewise.
18534 (-mquad-memory): Likewise.
18536 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
18537 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
18539 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
18540 (rs6000_debug_vector_unit): Add p8_vector.
18541 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
18542 definitions. Also print fusion state.
18543 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
18544 (rs6000_builtin_mask_calculate): Add power8 builtin support.
18545 (rs6000_option_override_internal): Add support for power8.
18546 (rs6000_common_init_builtins): Add debugging for skipped builtins
18547 if -mdebug=builtin.
18548 (rs6000_adjust_cost): Add power8 support.
18549 (rs6000_issue_rate): Likewise.
18550 (insn_must_be_first_in_group): Likewise.
18551 (insn_must_be_last_in_group): Likewise.
18552 (force_new_group): Likewise.
18553 (rs6000_register_move_cost): Likewise.
18554 (rs6000_opt_masks): Likewise.
18556 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
18557 power8 capable assembler, default to power7 options.
18558 (TARGET_DIRECT_MOVE): Likewise.
18559 (TARGET_CRYPTO): Likewise.
18560 (TARGET_P8_VECTOR): Likewise.
18561 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
18562 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
18563 (VECTOR_MEM_P8_VECTOR_P): Likewise.
18564 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
18565 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
18566 (TARGET_XSCVDPSPN): Likewise.
18567 (TARGET_XSCVSPDPN): Likewsie.
18568 (TARGET_SYNC_HI_QI): Likewise.
18569 (TARGET_SYNC_TI): Likewise.
18570 (MASK_CRYPTO): Likewise.
18571 (MASK_DIRECT_MOVE): Likewise.
18572 (MASK_P8_FUSION): Likewise.
18573 (MASK_P8_VECTOR): Likewise.
18574 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
18575 temporary used by some of the direct move instructions to get two FP
18576 temporary registers does not force creation of a stack frame.
18577 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
18578 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
18579 that any VSX registers are tieable, even if they are also an
18580 Altivec vector mode.
18581 (r6000_reg_class_enum): Add wm, wr, wv constraints.
18582 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
18583 (RS6000_BTM_CRYPTO): Likewise.
18584 (RS6000_BTM_COMMON): Likewise.
18586 * config/rs6000/rs6000.md (cpu attribute): Add power8.
18587 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
18588 (enum rs6000_vector): Add power8 vector support.
18590 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18594 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
18595 (any_sibcall_could_use_r3): this and handle indirect calls.
18596 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
18598 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18600 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
18602 2013-05-22 Richard Biener <rguenther@suse.de>
18604 PR middle-end/57349
18605 * profile.c (branch_prob): Do not split blocks that are
18606 abnormally receiving from ECF_RETURNS_TWICE functions.
18608 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18610 * recog.c (offsettable_address_addr_space_p): Fix calculation of
18611 address mode. Move pointer mode initialization to the same place.
18613 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
18615 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
18616 while it has any effect.
18618 2013-05-21 Easwaran Raman <eraman@google.com>
18620 PR tree-optimization/57322
18621 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
18622 UID of the statement added to the BB to be 1.
18624 2013-05-21 Jakub Jelinek <jakub@redhat.com>
18626 PR tree-optimization/57331
18627 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
18628 of conversion from pointer type to integral type with integer.
18630 2013-05-21 Martin Jambor <mjambor@suse.cz>
18633 * ipa-prop.c (ipa_read_node_info): Process param_used and
18634 controlled_uses in the same order as when writing.
18636 2013-05-21 Magnus Granberg <baldrick@free.fr>
18639 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
18641 2013-05-21 Richard Biener <rguenther@suse.de>
18643 PR tree-optimization/57318
18644 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
18645 estimate stmts with side-effects as likely eliminated.
18647 2013-05-21 Richard Biener <rguenther@suse.de>
18649 PR tree-optimization/57330
18650 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
18651 preserve the call stmts fntype.
18653 2013-05-21 Richard Biener <rguenther@suse.de>
18655 PR tree-optimization/57303
18656 * tree-ssa-sink.c (statement_sink_location): Improve killing
18657 stmt detection and properly handle self-assignments.
18659 2013-05-21 Christian Bruel <christian.bruel@st.com>
18661 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
18662 spanning registers. LEAF_REG_REMAP is supported only for contiguous
18663 registers. Set register size out of the PARALLEL loop.
18665 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
18668 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
18669 (*fmasf4, *fmasf4_media): New insns.
18671 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
18673 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
18674 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
18675 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
18676 (mips_idiv_insns): Update the comments to say that the returned
18677 instruction counts are in units of BASE_INSN_LENGTH.
18678 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
18679 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
18680 using 2 rather than 4 as the length of indirect MIPS16 and
18681 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
18682 length of a NOP. Don't divide MIPS16 lengths by 2.
18683 (mips16_split_long_branches): Assume a branch is long if the
18684 length is greater than 4 rather than 8.
18685 * config/mips/mips.md (length): Give MIPS16 lengths directly,
18686 rather than multiplying them by 2. Multiply instruction counts
18687 by BASE_INSN_LENGTH rather than 4.
18688 (*jump_mips16, tls_get_tp_mips16_<mode>)
18689 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
18691 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
18693 * config/mips/mips.md (extended_mips16): Remove branch case.
18694 (length): Remove duplicated extended_mips16 test.
18696 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
18698 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
18700 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
18702 * recog.h (Recog_data): Rename to...
18703 (recog_data_d): ...this.
18704 (recog_data): Update accordingly.
18705 * recog.c (recog_data): Likewise.
18706 * reload.c (save_recog_data): Likewise.
18707 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
18708 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
18710 2013-05-17 Julian Brown <julian@codesourcery.com>
18712 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
18713 found in a REG_EQUAL note, invalidate it.
18715 2013-05-17 Easwaran Raman <eraman@google.com>
18717 * tree-ssa-reassoc.c (find_insert_point): New function.
18718 (insert_stmt_after): Likewise.
18719 (get_def_stmt): Likewise.
18720 (ensure_ops_are_available): Likewise.
18721 (not_dominated_by): Likewise.
18722 (rewrite_expr_tree): Do not move statements beyond what is
18723 necessary. Remove call to swap_ops_for_binary_stmt...
18724 (reassociate_bb): ... and move it here.
18725 (build_and_add_sum): Assign UIDs for new statements.
18726 (linearize_expr): Likewise.
18727 (do_reassoc): Renumber gimple statement UIDs.
18729 2013-05-17 Jan Hubicka <jh@suse.cz>
18731 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
18733 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
18734 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
18736 (output_weakrefs): Update.
18738 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
18739 Martin Jambor <mjambor@suse.cz>
18741 PR middle-end/57276
18742 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
18743 value that corresponds to the given aggval is found in values vector.
18745 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
18747 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
18748 sse, sse2, sse3, ssse3 and sse4a flags to options.
18750 2013-05-17 David Malcolm <dmalcolm@redhat.com>
18752 * gengtype-state.c: (s_expr_writer): New class, to handle
18753 prettifying of output layout of s-expressions.
18754 (state_writer): New class, to write out gtype.state.
18755 (state_written_type_count): Move this variable into member data of
18757 (s_expr_writer::s_expr_writer): New code: constructor for new class
18758 (state_writer::state_writer(): ditto
18759 (s_expr_writer::write_new_line): New function
18760 (s_expr_writer::write_any_indent): ditto
18761 (s_expr_writer::begin_s_expr): ditto
18762 (s_expr_writer::end_s_expr): ditto
18763 (write_state_fileloc): convert to method of state_writer...
18764 (state_writer:: write_state_fileloc): ...and use methods of
18765 s_expr_writer to write indentation into the gtype.state output file
18766 to visually represent the hierarchical structure of the list
18768 (write_state_fields): ditto, renaming to...
18769 (state_writer::write_state_fields)
18770 (write_state_a_string): ditto, renaming to...
18771 (state_writer::write_state_a_string)
18772 (write_state_string_option): ditto, renaming to...
18773 (state_writer::write_state_string_option)
18774 (write_state_type_option): ditto, renaming to...
18775 (state_writer::write_state_type_option)
18776 (write_state_nested_option): ditto, renaming to...
18777 (state_writer::write_state_nested_option)
18778 (write_state_option): ditto, renaming to...
18779 (state_writer::write_state_option)
18780 (write_state_options): ditto, renaming to...
18781 (state_writer::write_state_options)
18782 (write_state_lang_bitmap): ditto, renaming to...
18783 (state_writer::write_state_lang_bitmap)
18784 (write_state_version): ditto, renaming to...
18785 (state_writer::write_state_version)
18786 (write_state_scalar_type): ditto, renaming to...
18787 (state_writer::write_state_scalar_type)
18788 (write_state_string_type): ditto, renaming to...
18789 (state_writer::write_state_string_type)
18790 (write_state_undefined_type): ditto, renaming to...
18791 (state_writer::write_state_undefined_type)
18792 (write_state_struct_union_type): ditto, renaming to...
18793 (state_writer::write_state_struct_union_type)
18794 (write_state_struct_type): ditto, renaming to...
18795 (state_writer::write_state_struct_type)
18796 (write_state_user_struct_type): ditto, renaming to...
18797 (state_writer::write_state_user_struct_type)
18798 (write_state_lang_struct_type): ditto, renaming to...
18799 (state_writer::write_state_lang_struct_type)
18800 (write_state_param_struct_type): ditto, renaming to...
18801 (state_writer::write_state_param_struct_type)
18802 (write_state_pointer_type): ditto, renaming to...
18803 (state_writer::write_state_pointer_type)
18804 (write_state_array_type): ditto, renaming to...
18805 (state_writer::write_state_array_type)
18806 (write_state_gc_used): ditto, renaming to...
18807 (state_writer::write_state_gc_used)
18808 (write_state_common_type_content): ditto, renaming to...
18809 (state_writer::write_state_common_type_content)
18810 (write_state_type): ditto, renaming to...
18811 (state_writer::write_state_type)
18812 (write_state_pair_list): ditto, renaming to...
18813 (state_writer::write_state_pair_list)
18814 (write_state_pair): ditto, renaming to...
18815 (state_writer::write_state_pair)
18816 (write_state_typedefs): ditto, renaming to...
18817 (state_writer::write_state_typedefs)
18818 (write_state_structures): ditto, renaming to...
18819 (state_writer::write_state_structures)
18820 (write_state_param_structs): ditto, renaming to...
18821 (state_writer::write_state_param_structs)
18822 (write_state_variables): ditto, renaming to...
18823 (state_writer::write_state_variables)
18824 (write_state_srcdir): ditto, renaming to...
18825 (state_writer::write_state_srcdir)
18826 (write_state_files_list): ditto, renaming to...
18827 (state_writer::write_state_files_list)
18828 (write_state_languages): ditto, renaming to...
18829 (state_writer::write_state_languages)
18830 (write_state): create a state_writer instance and use it when
18831 writing out the state file
18833 2013-05-17 Mike Stump <mikestump@comcast.net>
18835 PR rtl-optimization/57304
18836 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
18837 accessing DF_REF_REAL_LOC.
18839 2013-05-17 Jakub Jelinek <jakub@redhat.com>
18841 PR rtl-optimization/57281
18842 PR rtl-optimization/57300
18843 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
18844 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
18845 what the other splitter did if the registers are dead.
18847 2013-05-17 Richard Biener <rguenther@suse.de>
18849 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
18852 2013-05-17 Jakub Jelinek <jakub@redhat.com>
18854 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
18857 2013-05-17 Marek Polacek <polacek@redhat.com>
18859 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
18860 length when doing non-zero store of storing '\0' to '\0'.
18862 2013-05-17 Jakub Jelinek <jakub@redhat.com>
18864 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
18865 vect_external_def oprnd1 with loop_vinfo, try to emit
18866 optional cast, negation and and stmts on the loop preheader
18867 edge instead of into the pattern def seq.
18869 PR tree-optimization/57051
18870 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
18871 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
18873 2013-05-16 Nick Clifton <nickc@redhat.com>
18875 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
18876 (rl78_is_naked_func): New function.
18877 (rl78_expand_prologue): Skip prologue generation for naked functions.
18878 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
18879 * doc/extend.texi (naked): Add RL78 to the list of processors
18880 that supports this attribute.
18882 2013-05-16 Jeff Law <law@redhat.com>
18884 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
18886 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
18888 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
18889 cache parameters using detect_caches_amd also for CYRIX,
18890 NSC and TM2 signatures.
18892 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
18893 Dzianis Kahanovich <mahatma@eu.by>
18897 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
18898 VIA/Centaur processors and determine their cache parameters
18899 using detect_caches_amd.
18901 2013-05-16 Teresa Johnson <tejohnson@google.com>
18903 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
18904 (rtl_verify_edges): New function.
18905 (rtl_verify_bb_insns): Ditto.
18906 (rtl_verify_bb_pointers): Ditto.
18907 (rtl_verify_bb_insn_chain): Ditto.
18908 (rtl_verify_fallthru): Ditto.
18909 (rtl_verify_bb_layout): Ditto.
18910 (rtl_verify_flow_info_1): Outline checks into new functions.
18911 (rtl_verify_flow_info): Ditto.
18913 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
18915 * cfghooks.c (copy_bbs): Add update_dominance argument.
18916 * cfghooks.h (copy_bbs): Update prototype.
18917 * tree-cfg.c (gimple_duplicate_sese_region):
18918 Add update_dominance argument.
18919 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
18920 * tree-ssa-loop-ch.c (copy_loop_headers): Update
18921 gimple_duplicate_sese_region call.
18922 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
18923 Update copy_bbs call.
18924 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
18925 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
18927 2013-05-16 Jakub Jelinek <jakub@redhat.com>
18929 * tree-vectorizer.h (NUM_PATTERNS): Increment.
18930 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
18931 vect_recog_rotate_pattern.
18932 (vect_recog_rotate_pattern): New function.
18934 2013-05-16 Jason Merrill <jason@redhat.com>
18936 * Makefile.in (LLINKER): New variable.
18937 (mostlyclean): Remove link mutex.
18938 * configure.ac: Handle --enable-link-mutex.
18939 * lock-and-run.sh: New script.
18941 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18944 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
18947 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18949 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
18951 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
18953 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
18954 * config/arm/arm.c (next_consecutive_mem): New function.
18955 (gen_movmem_ldrd_strd): Likewise.
18956 * config/arm/arm.md (movmemqi): Update condition and code.
18957 (unaligned_loaddi, unaligned_storedi): New patterns.
18959 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18961 * config.gcc: Obsolete *-*-solaris2.9*.
18962 * doc/install.texi (Specific, *-*-solaris2*): Document it.
18964 2013-05-16 Richard Biener <rguenther@suse.de>
18966 * passes.c (init_optimization_passes): Move pass_parallelize_loops
18967 earlier, after GRAPHITE transforms and IV canonicalization.
18969 2013-05-16 Jakub Jelinek <jakub@redhat.com>
18971 * omp-low.c (extract_omp_for_data): For collapsed loops,
18972 if at least one of the loops is known at compile time to
18973 iterate zero times, set count to 0.
18974 (expand_omp_regimplify_p): New function.
18975 (expand_omp_for_generic): For collapsed loops, if at least
18976 one of the loops isn't known to iterate at least once,
18977 add runtime check with setting count to 0.
18978 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
18979 For unsigned types if it isn't known at compile time that
18980 the loop will iterate at least once, add runtime check to bypass
18981 the whole loop if initial condition isn't true.
18983 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
18985 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
18987 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
18989 PR middle-end/57286
18990 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
18991 transformations to avoid an infinite loop.
18993 2013-05-16 Marek Polacek <polacek@redhat.com>
18995 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
18997 2013-05-15 Leif Ekblad <leif@rdos.net>
18999 * config/i386/i386.c (ix86_decompose_address): Use
19000 DEFAULT_TLS_SEG_REG to access TLS segment register.
19001 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19002 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19003 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19005 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19008 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19009 sibling calls to functions that would normally be lazily bound,
19010 unless $gp is call-clobbered.
19012 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19014 * config/i386/i386.c (ix86_option_override_internal): Update
19015 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19016 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19017 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19018 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19020 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19022 2013-05-15 Andreas Schwab <schwab@suse.de>
19024 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19025 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19028 2013-05-15 Teresa Johnson <tejohnson@google.com>
19030 * loop-unroll.c (report_unroll_peel): Check decision before
19031 emitting unroll/peel message.
19033 2013-05-15 Teresa Johnson <tejohnson@google.com>
19035 * function.h (has_bb_partition): New rtl_data flag.
19036 (bb_reorder_complete): Ditto.
19037 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19038 instead of flag_reorder_blocks_and_partition.
19039 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19040 with some enhancements.
19041 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19042 * bb-reorder.c (connect_traces): Check for has_bb_partition
19043 instead of flag_reorder_blocks_and_partition.
19044 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19045 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19046 verify_hot_cold_block_grouping.
19047 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19049 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19052 * config/arm/predicates.md (call_insn_operand): New predicate.
19053 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19054 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19055 if insn is not a tail call.
19056 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19058 * config/arm/arm.h (enum reg_class): New caller save register class.
19059 (REG_CLASS_NAMES): Likewise.
19060 (REG_CLASS_CONTENTS): Likewise.
19061 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19064 2013-05-15 Richard Biener <rguenther@suse.de>
19066 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19067 of MSG_OPTIMIZED_LOCATIONS.
19068 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19069 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19071 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19072 of MSG_OPTIMIZED_LOCATIONS.
19073 (execute_vect_slp): Likewise.
19074 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19075 (vect_create_cond_for_alias_checks): Likewise.
19076 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19077 (vect_recog_widen_mult_pattern): Likewise.
19078 (vect_recog_widen_sum_pattern): Likewise.
19079 (vect_recog_over_widening_pattern): Likewise.
19080 (vect_recog_widen_shift_pattern): Likewise.
19081 (vect_recog_vector_vector_shift_pattern): Likewise.
19082 (vect_recog_divmod_pattern): Likewise.
19083 (vect_recog_mixed_size_cond_pattern): Likewise.
19084 (vect_recog_bool_pattern): Likewise.
19085 (vect_pattern_recog_1): Likewise.
19087 2013-05-15 Martin Jambor <mjambor@suse.cz>
19089 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19090 non-functions to builtin_unreachable.
19091 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19092 correct when new direct edges were discovered.
19094 2013-05-15 Martin Jambor <mjambor@suse.cz>
19096 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19097 header, print symbol order instead of node uid, print more information
19098 about indirect edge targets.
19099 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19101 (ipa_make_edge_direct_to_target): Likewise.
19102 (remove_described_reference): Likewise.
19103 (propagate_controlled_uses): Likewise.
19104 (ipa_print_node_params): Also print symbol order.
19105 (ipcp_transform_function): Print symbol order instead of node uids.
19106 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19107 (cgraph_get_create_real_symbol_node): Likewise.
19108 * ipa-cp.c (print_lattice): Likewise.
19109 (print_all_lattices): Likewise.
19110 (determine_versionability): Likewise.
19111 (initialize_node_lattices): Likewise.
19112 (estimate_local_effects): Likewise.
19113 (update_profiling_info): Likewise.
19114 (create_specialized_node): Likewise.
19115 (perhaps_add_new_callers): Likewise.
19116 (decide_about_value): Likewise.
19117 (decide_whether_version_node): Likewise.
19118 (identify_dead_nodes): Likewise.
19119 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19120 (dump_inline_summary): Likewise.
19121 (estimate_node_size_and_time): Likewise.
19122 (inline_analyze_function): Likewise.
19123 * ipa-inline.c (report_inline_failed_reason): Likewise.
19124 (want_early_inline_function_p): Likewise.
19125 (edge_badness): Likewise.
19126 (update_edge_key): Likewise.
19127 (inline_small_functions): Likewise. Add dumping of order to two other
19129 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19130 instead of node uids.
19131 (propagate_pure_const): Likewise.
19132 (propagate_pure_const): Likewise.
19133 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19134 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19135 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19137 * tree-pretty-print.c (dump_function_header): Likewise.
19138 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19139 Print symbol order instead of node uids.
19141 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19143 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19144 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19146 2013-05-15 Richard Biener <rguenther@suse.de>
19148 PR tree-optimization/57275
19149 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19150 return value for fail to do runtime alias checks for gather loads.
19152 2013-05-15 Jan Hubicka <jh@suse.cz>
19156 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19157 weakrefs are not external.
19158 (lto_symtab_merge_decls): Fix thinko when dealing with
19159 non-lto_symtab decls.
19160 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19161 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19162 * varpool.c (dump_varpool_node): Dump more flags.
19164 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19166 * config/i386/i386.c (processor_alias_table): Add instruction
19167 FSGSBASE for AMD bdver3 architecture.
19169 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19171 * tree.c (warn_deprecated_use): Print file:line using locus color.
19172 * diagnostic.c (diagnostic_report_current_module): Print file:line
19173 and file:line:column using locus color.
19175 2013-05-14 Mike Stump <mikestump@comcast.net>
19177 * gdbinit.in: Add __null.
19179 2013-05-14 Mike Stump <mikestump@comcast.net>
19181 * recog.h: Rename struct recog_data to Recog_data.
19182 * recog.c: Likewise.
19183 * reload.c (can_reload_into): Likewise.
19184 * config/picochip/picochip.c: Likewise.
19186 2013-05-14 Mike Stump <mikestump@comcast.net>
19188 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
19190 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
19192 * resource.h (struct resources): Remove unch_memory member.
19193 (CLEAR_RESOURCE): Don't clear unch_memory.
19194 * resource.c (mark_referenced_resources): Don't set it.
19195 (mark_set_resources): Likewise.
19196 (mark_target_live_regs): Don't clear it.
19197 (init_resource_info): Likewise.
19198 * reorg.c (resource_conflicts_p): Don't compare it.
19199 (redundant_insn): Don't set it.
19201 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
19203 * emit-rtl.c (next_label): Remove unused function.
19204 (skip_consecutive_labels, link_cc0_insns): Move to ...
19205 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
19206 only place where these functions are used, and make them static.
19208 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19210 * fold-const.c (fold_negate_expr): Handle vectors.
19211 (fold_truth_not_expr): Make it static.
19212 (fold_invert_truthvalue): New static function.
19213 (invert_truthvalue_loc): Handle vectors. Do not call
19214 fold_truth_not_expr directly.
19215 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
19216 <TRUTH_NOT_EXPR>: Do not cast to boolean.
19217 (fold_comparison): Handle vector constants.
19218 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
19219 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
19220 * tree.h (fold_truth_not_expr): Remove declaration.
19222 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
19224 * config/aarch64/aarch64-simd.md
19225 (aarch64_vcond_internal<mode>): Rename to...
19226 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
19227 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
19228 float modes. Clarify all iterator modes.
19229 (vcond<mode><mode>): Use new name for vcond expanders.
19230 (vcond<v_cmp_result><mode>): Likewise.
19231 (vcondu<mode><mode>: Likewise.
19232 * config/aarch64/iterators.md (VDQF_COND): New.
19234 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19237 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
19238 variable for the shift amount. Check that we shift by non-negative
19241 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
19244 * config/arm/arm.h (EPILOGUE_USES): Only return true
19245 for LR_REGNUM after epilogue_completed.
19247 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
19249 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
19250 is error_mark_node.
19252 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19255 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
19256 and Solaris 11+/x86 with gld.
19257 * configure: Regenerate.
19259 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19261 * expmed.c (expand_shift_1): Canonicalize rotates by
19262 constant bitsize / 2 to bitsize - 1.
19263 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
19264 case ROTATERT>: Likewise.
19267 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19269 * config/i386/i386.md (rotateinv): New code attr.
19270 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19271 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19272 roll $31, %eax, etc.
19274 2013-05-14 Richard Biener <rguenther@suse.de>
19276 PR middle-end/57235
19277 * tree-eh.c (sink_clobbers): Give up for successors with
19278 multiple predecessors and no virtual uses.
19280 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
19282 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
19283 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
19285 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19287 PR middle-end/57251
19288 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
19289 the case when both op0 and op1 have VOIDmode.
19291 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19293 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
19294 in multiply-accumulate mode.
19296 2013-05-13 Guozhi Wei <carrot@google.com>
19298 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
19300 2013-05-13 Kai Tietz <ktietz@redhat.com>
19303 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
19304 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
19305 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
19306 * config/i386/i386.c (ix86_option_override_internal): Likewise.
19307 (ix86_expand_prologue): Likewise.
19308 (ix86_expand_split_stack_prologue): Likewise.
19309 (legitimate_pic_address_disp_p): Likewise.
19310 (legitimize_pic_address): Likewise.
19311 (legitimize_tls_address): Likewise.
19312 (legitimize_pe_coff_symbol): Likewise.
19313 (output_pic_addr_const): Likewise.
19314 (construct_plt_address): Likewise.
19315 (ix86_expand_call): Likewise.
19316 (x86_output_mi_thunk): Likewise.
19317 (x86_function_profiler): Likewise.
19319 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
19321 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
19322 similar switch cases.
19323 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
19324 (aarch64_simd_mov_to_<mode>low): Delete.
19325 (aarch64_simd_mov_to_<mode>high): Delete.
19326 (move_lo_quad_<mode>): Add w<-r alternative.
19327 (aarch64_simd_move_hi_quad_<mode>): Likewise.
19328 (aarch64_simd_mov_from_*): Update type attribute.
19329 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
19332 2013-05-13 Jan Hubicka <jh@suse.cz>
19334 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
19335 * config/i386/i386.c (ix86_compute_frame_layout,
19336 ix86_expand_epilogue, emit_i387_cw_initialization,
19337 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
19338 ix86_local_alignment): Fix use of size/speed predicates.
19340 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19342 PR tree-optimization/45216
19343 PR tree-optimization/57157
19344 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
19345 the (-Y) & (B - 1) variant if OP is |.
19346 * expmed.c (expand_shift_1): For rotations by const0_rtx just
19347 return shifted. Use (-op1) & (prec - 1) as other_amount
19348 instead of prec - op1.
19350 2013-05-13 Martin Jambor <mjambor@suse.cz>
19352 PR middle-end/42371
19353 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
19354 (ipa_constant_data): New type.
19355 (ipa_jump_func): Use ipa_constant_data to hold information about
19356 constant jump functions.
19357 (ipa_get_jf_constant): Adjust to jump function type changes.
19358 (ipa_get_jf_constant_rdesc): New function.
19359 (ipa_param_descriptor): New field controlled_uses.
19360 (ipa_get_controlled_uses): New function.
19361 (ipa_set_controlled_uses): Likewise.
19362 * ipa-ref.h (ipa_find_reference): Declare.
19363 * ipa-prop.c (ipa_cst_ref_desc): New type.
19364 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
19366 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
19367 New parameter cs. Adjust all callers.
19368 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
19369 (remove_described_reference): New function.
19370 (jfunc_rdesc_usable): Likewise.
19371 (try_make_edge_direct_simple_call): Decrement controlled use count,
19372 attempt to remove reference if it hits zero.
19373 (combine_controlled_uses_counters): New function.
19374 (propagate_controlled_uses): Likewise.
19375 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
19376 (ipa_edge_duplication_hook): Duplicate reference descriptions.
19377 (ipa_print_node_params): Print described use counter.
19378 (ipa_write_jump_function): Adjust to jump function type changes.
19379 (ipa_read_jump_function): New parameter CS, pass it to
19380 ipa_set_jf_constant. Adjust caller.
19381 (ipa_write_node_info): Stream controlled use count
19382 (ipa_read_node_info): Likewise.
19383 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
19385 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
19386 count. Remove cloning-added reference if it reaches zero.
19387 * ipa-ref.c (ipa_find_reference): New function.
19389 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19391 * config/i386/i386.c (processor_target_table): Modified default
19392 alignment values for AMD BD and BT architectures.
19394 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
19396 * tree-vect-generic.c (uniform_vector_p): Move ...
19397 * tree.c (uniform_vector_p): ... here.
19398 * tree.h (uniform_vector_p): Declare it.
19399 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
19402 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19404 PR tree-optimization/57230
19405 * tree-ssa-strlen.c (handle_char_store): Record length for
19406 array store from STRING_CST.
19408 PR tree-optimization/57230
19409 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
19412 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
19414 * config/epiphany/epiphany.c (epiphany_init): Check size of
19415 NUM_MODES_FOR_MODE_SWITCHING.
19416 (epiphany_expand_prologue):
19417 Remove CONFIG_REGNUM initial value handling code.
19418 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
19419 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
19420 (emit_set_fp_mode, epiphany_mode_after): Likewise.
19421 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
19422 Don't return 1 for FP_MODE_NONE.
19423 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
19424 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
19425 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
19426 * config/epiphany/epiphany.md (save_config): New pattern.
19428 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
19430 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
19432 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
19434 * config/i386/i386.md (memory): Handle sseishft1.
19435 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
19436 (*vec_extractv2di_1): Ditto.
19438 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
19440 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
19443 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
19445 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
19446 Add mthumb/march=armv7-a multilib.
19447 Add mthumb/march=armv7-r multilib.
19448 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
19450 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
19452 * config/v850/t-rtems: Add more multilibs.
19454 2013-05-10 Richard Biener <rguenther@suse.de>
19456 PR tree-optimization/57214
19457 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
19458 not propagate from SSA names that occur in abnormal PHI nodes.
19460 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
19462 * stor-layout.c (element_precision): New function.
19463 * machmode.h (element_precision): Declare it.
19464 * tree.c (build_minus_one_cst): New function.
19465 (element_precision): Likewise.
19466 * tree.h (build_minus_one_cst): Declare new function.
19467 (element_precision): Likewise.
19468 * fold-const.c (operand_equal_p): Use element_precision.
19469 (fold_binary_loc): Handle vector types.
19470 * convert.c (convert_to_integer): Use element_precision.
19471 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
19474 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
19476 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
19477 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
19478 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
19479 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
19480 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
19481 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
19482 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
19483 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
19484 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
19485 (Uuw8): New constraints.
19486 (Usb4): Move into alphabetical order.
19487 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
19488 (sd8_operand, ub8_operand, uw8_operand): New predicates.
19489 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
19490 previously unnamed patterns.
19491 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
19492 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
19493 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
19494 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
19495 of set_attr_alternative/if_then_else. Use extended_mips16 instead
19496 of specific lengths.
19498 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19500 * config/i386/i386.md (rotateinv): New code attr.
19501 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19502 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19503 roll $31, %eax, etc.
19505 PR tree-optimization/45216
19506 PR tree-optimization/57157
19507 * tree-ssa-forwprop.c (simplify_rotate): New function.
19508 (ssa_forward_propagate_and_combine): Call it.
19510 2013-05-10 Richard Biener <rguenther@suse.de>
19512 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
19513 disable peeling when we version for aliasing.
19514 (vector_alignment_reachable_p): Honor explicit user alignment.
19515 (vect_supportable_dr_alignment): Likewise.
19516 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
19517 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
19518 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
19519 then peeling to arrange for the cost-model check to come first.
19521 2013-05-10 Alan Modra <amodra@gmail.com>
19523 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
19524 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
19525 * configure: Regenerate.
19527 2013-05-10 Alan Modra <amodra@gmail.com>
19530 * varasm.c (default_elf_select_section): Move !DECL_P check..
19531 (get_named_section): ..to here before calling get_section_name.
19533 (default_section_type_flags): Add DECL_P check.
19534 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
19535 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
19537 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
19539 * config/epiphany/epiphany.c (epiphany_expand_prologue):
19540 When using gen_stack_adjust_str with a register offset, add a
19541 REG_FRAME_RELATED_EXPR note.
19543 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
19545 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
19546 (*vec_extractv4si_zext_mem): Ditto.
19547 (*vec_extractv2di): Add 0->x and x->x alternatives.
19548 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
19549 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
19551 2013-05-09 Jason Merrill <jason@redhat.com>
19553 N3639 C++1y VLA support
19554 * gimplify.c (gimplify_vla_decl): Don't touch an existing
19557 * tree.c (build_constructor_va): New.
19558 * tree.h: Declare it.
19560 2013-05-09 Martin Jambor <mjambor@suse.cz>
19563 * gimple-fold.c (canonicalize_constructor_val): Call
19564 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
19566 2013-05-09 Jan Hubicka <jh@suse.cz>
19567 Richard Biener <rguenther@suse.de>
19570 * symtab.c (symtab_make_decl_local): Do not add private names.
19572 2013-05-09 Jan Hubicka <jh@suse.cz>
19575 * symtab.c (insert_to_assembler_name_hash): Handle clones.
19576 (unlink_from_assembler_name_hash): Likewise.
19577 (symtab_prevail_in_asm_name_hash, symtab_register_node,
19578 symtab_unregister_node, symtab_initialize_asm_name_hash,
19579 change_decl_assembler_name): Update.
19581 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
19583 * config/aarch64/aarch64.md: New movtf split.
19584 (*movtf_aarch64): Update.
19585 (aarch64_movdi_tilow): Handle TF modes and rename to
19586 aarch64_movdi_<mode>low.
19587 (aarch64_movdi_tihigh): Handle TF modes and rename to
19588 aarch64_movdi_<mode>high
19589 (aarch64_movtihigh_di): Handle TF modes and rename to
19590 aarch64_mov<mode>high_di
19591 (aarch64_movtilow_di): Handle TF modes and rename to
19592 aarch64_mov<mode>low_di
19593 (aarch64_movtilow_tilow): Remove spurious whitespace.
19594 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
19596 (aarch64_print_operand): Update.
19598 2013-05-09 Alan Modra <amodra@gmail.com>
19600 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
19602 * configure: Regenerate.
19604 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
19606 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
19607 splitter preparation statements.
19608 * config/i386/sse.md (*vec_extract* splitters): Ditto.
19609 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
19612 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19614 * gimple-ssa-strength-reduction.c (count_candidates): Change
19615 return value to int.
19616 (analyze_candidates_and_replace): Change type of length to int.
19618 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
19620 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
19621 (*vec_extract<mode>): Use VI12_128 mode iterator.
19622 (*vec_extract<mode>_mem): Ditto.
19623 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
19626 2013-05-08 Diego Novillo <dnovillo@google.com>
19631 2012-08-17 Diego Novillo <dnovillo@google.com>
19634 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
19635 * config.in: Regenerate.
19636 * configure: Regenerate.
19637 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
19639 2013-05-08 Jan Hubicka <jh@suse.cz>
19642 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
19643 * cgraph.h (symtab_node_base): Add unique_name.
19644 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
19645 input_overwrite_node, input_varpool_node): Stream unique_name.
19646 * cgraphclones.c (cgraph_create_virtual_clone,
19647 cgraph_function_versioning): Set unique_name.
19648 * ipa.c (function_and_variable_visibility): Set unique_name.
19650 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19652 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
19653 (alloc_cand_and_find_basis): Restrict conditional candidate
19654 processing to CAND_MULTs.
19656 2013-05-08 Jan Hubicka <jh@suse.cz>
19659 lto-symtab.c (lto_symtab_symbol_p): New function.
19660 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
19661 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
19662 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
19663 Skip static symbols.
19665 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
19667 PR tree-optimization/57200
19668 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
19669 Only call inform if the preceding warning_at returns true.
19671 2013-05-07 Han Shen <shenhan@google.com>
19673 * cfgexpand.c (record_or_union_type_has_array_p): New function.
19674 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
19675 * common.opt (fstack-protector-strong): New option.
19676 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
19677 * doc/invoke.texi (Optimization Options): Document
19678 "-fstack-protector-strong".
19679 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
19681 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
19683 * config/mips/mips.c (mips_machine_reorg2): Return 0.
19685 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
19687 * ira.c (update_equiv_regs): Add insn having equiv memory even if
19688 it is not lhs of the insn.
19689 (setup_reg_equiv): Remove insn having equiv memory which it is not
19691 * lra-constraints.c (process_address): Try to improve generation
19692 code for address base + disp.
19693 (lra_constraints): Make correct the code for checking insn setting
19694 up backward equivalence. Remove insn only if it is in the init
19696 * lra-eliminations.c (update_reg_eliminate): Change return value.
19697 (lra_eliminate): Use the result.
19699 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
19701 * config/i386/sse.md (ssescalarnummask): New mode attribute.
19702 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
19703 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
19704 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
19705 register target operands.
19706 (*vec_extractv8hi_sse2): New pattern.
19707 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
19708 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
19709 (*vec_extract<mode>_mem): New insn and split pattern.
19711 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
19713 * config/arm/arm.c (arm_asan_shadow_offset): New function.
19714 (TARGET_ASAN_SHADOW_OFFSET): Define.
19715 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
19716 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
19718 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19720 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
19721 (incr_vec_index): Return -1 if increment not found.
19722 (create_add_on_incoming_edge): Assert if increment not found.
19723 (record_increment): Limit number of increments recorded.
19724 (all_phi_incrs_profitable): Return false if an increment not found.
19725 (replace_profitable_candidates): Don't process increments that were
19727 (analyze_candidates_and_replace): Limit size of incr_vec.
19729 2013-05-07 Richard Biener <rguenther@suse.de>
19731 * calls.c (special_function_p): setjmp-like functions are leaf.
19732 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
19733 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
19735 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
19737 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
19738 (aarch64_simd_mov<mode>): New expander.
19739 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
19740 (aarch64_simd_mov_to_<mode>high): Likewise.
19741 (aarch64_simd_mov_from_<mode>low): Likewise.
19742 (aarch64_simd_mov_from_<mode>high): Likewise.
19743 (aarch64_dup_lane<mode>): Update.
19744 (aarch64_dup_lanedi): New instruction pattern.
19745 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
19746 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
19748 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19750 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
19751 (replace_mult_candidate): Remove unnecessary argument; remove
19752 unnecessary parameter from call to introduce_cast_before_cand.
19753 (replace_unconditional_candidate): Remove unnecessary parameter
19754 from call to replace_mult_candidate.
19755 (replace_conditional_candidate): Likewise.
19756 (insert_initializers): Use make_temp_ssa_name.
19757 (introduce_cast_before_cand): Remove unnecessary argument; use
19758 make_temp_ssa_name.
19759 (replace_one_candidate): Remove unnecessary argument; remove
19760 unnecessary parameter from calls to introduce_cast_before_cand.
19761 (replace_profitable_candidates): Remove unnecessary parameters
19762 from calls to replace_one_candidate.
19764 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19766 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
19767 phi def as possibly hiding a basis for a CAND_ADD whose operands
19768 have been commuted in the analysis.
19769 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
19771 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19773 * config/aarch64/aarch64.md
19774 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
19775 shift value between 0-4.
19777 2013-05-07 Richard Biener <rguenther@suse.de>
19779 * double-int.h (rshift): New overload.
19780 * double-int.c (rshift): New function.
19781 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
19782 (create_reference_ops_from_ref): Remove.
19783 (vn_reference_insert): Use shared ops for constructing the
19784 reference and copy it.
19786 2013-05-07 Richard Biener <rguenther@suse.de>
19788 PR middle-end/57190
19789 * tree-eh.c (sink_clobbers): Properly propagate
19790 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
19792 2013-05-07 Jakub Jelinek <jakub@redhat.com>
19794 PR tree-optimization/57149
19795 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
19796 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
19797 collect_phi_def_edges, execute_late_warn_uninitialized): Use
19798 uninit_undefined_value_p instead of ssa_undefined_value_p.
19801 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
19802 for modifier == EXPAND_INITIALIZER.
19804 2013-05-07 Anton Blanchard <anton@samba.org>
19806 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
19807 for powerpc64 little endian.
19808 * configure: Regenerate.
19810 2013-05-06 Graham Stott <grahams@btinternet.com>
19812 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
19813 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
19814 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
19815 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
19817 2013-05-06 Graham Stott <grahams@btinternet.com>
19819 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
19820 codes which allow non-lvalues.
19822 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
19824 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
19825 components are all 1s.
19826 (integer_minus_onep): New function.
19827 * tree.h (integer_minus_onep): Declare it.
19828 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
19829 integer_minus_onep instead of integer_all_onesp.
19831 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
19834 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
19835 variations of these patterns.
19837 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
19839 * config/i386/i386.md (isa): Add x64_sse4 member.
19840 (enabled): Handle x64_sse4.
19841 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
19842 instruction for 64bit SSE4_1 targets. Update insn attributes.
19843 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
19844 instruction for SSE4_1 targets. Update insn attributes.
19845 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
19846 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
19847 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
19849 (*vec_extractv4si): Rename from *sse4_1_pextrd.
19850 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
19851 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
19853 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
19856 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
19858 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
19860 * final.c (do_assembler_dialects): Don't handle curly braces and
19861 vertical bar escaped by % as dialect delimiters.
19862 (output_asm_insn): Print curly braces and vertical bar if escaped
19863 by % and ASSEMBLER_DIALECT defined.
19864 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
19865 * doc/tm.texi: Regenerated.
19867 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
19869 * config/mips/mips.c: Include tree-pass.h.
19870 (mips_reorg): Split in pre- and post-dbr_schedule parts.
19871 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
19872 (pass_mips_machine_reorg2): New machine specific pass.
19873 (insert_pass_mips_machine_reorg2): New pass plugin definition.
19874 (mips_option_override): Register the new pass.
19875 * rtl.h (cleanup_barriers): Remove prototype.
19876 (dbr_schedule): Likewise.
19877 * jump.c (cleanup_barriers): Make static.
19878 * reorg.c (dbr_schedule): Likewise.
19880 2013-05-06 Richard Biener <rguenther@suse.de>
19882 PR tree-optimization/57185
19883 * tree-parloops.c (add_field_for_reduction): Handle anonymous
19884 SSA names properly.
19886 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
19889 * config/i386/i386.c (add_parameter_dependencies): Add dependence
19890 between "first_arg" and "insn", not "last" and "insn".
19892 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19894 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
19895 (find_candidates_in_block): Re-enable slsr_process_phi.
19896 (create_phi_basis): Fix double counting of candidate adjustment.
19898 2013-05-06 Richard Biener <rguenther@suse.de>
19900 PR middle-end/57147
19901 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
19902 the edge is also fallthru, preserve it and just clear the
19904 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
19905 also complex, preserve that and just clear the fallthru flag.
19906 * tree-inline.c (update_ssa_across_abnormal_edges): Also
19907 update virtual operands.
19909 2013-05-06 Alan Modra <amodra@gmail.com>
19911 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
19912 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
19913 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
19914 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
19915 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
19916 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
19918 2013-05-06 Alan Modra <amodra@gmail.com>
19920 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
19921 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
19922 (DEFAULT_ASM_ENDIAN): Define.
19923 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
19924 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
19925 Update -K PIC clause from sysv4.h.
19926 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
19927 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
19929 2013-05-06 Alan Modra <amodra@gmail.com>
19931 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
19932 twice for little-endian.
19933 (ashrdi3_no_power, ashrdi3): Support little-endian.
19935 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
19938 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
19939 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
19941 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
19942 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
19943 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
19946 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
19947 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19949 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
19950 * config/pa/pa.opt: Make mbig-switch a no-op.
19951 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
19952 (CASE_VECTOR_MODE): Always return SImode.
19953 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
19954 for the !TARGET_BIG_SWITCH case.
19955 * config/pa/pa-linux.h: Likewise.
19956 * config/pa/pa-openbsd.h: Likewise.
19957 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
19958 * config/pa/pa.md (short_jump): Remove define_insn.
19959 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
19960 (casesi0): Remove define_insn.
19961 (type): Remove btable_branch.
19962 (pa_combine_type): Likewise.
19963 (in_nullified_branch_delay): Likewise.
19964 (in_call_delay): Likewise.
19965 (define_delay): Likewise.
19966 (define_insn_reservation "Z3"): Likewise.
19967 (define_insn_reservation "Z4"): Likewise.
19968 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
19969 (pa_adjust_insn_length): Remove adjustment for btable branches.
19970 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
19973 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
19975 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
19976 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
19977 Add m->r,x alternatives.
19978 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
19979 splitters using SWI48x mode iterator.
19980 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
19981 TARGET_64BIT. Add m->x alternative.
19982 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
19983 Add o->x alternative. Enable for TARGET_SSE.
19984 (sse_storeq): Remove expander.
19985 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
19986 with memory input operand.
19987 (*vec_extractv2di_1 splitter): New.
19988 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
19989 * config/i386/i386.md (ssevecmodelower): New mode attribute.
19991 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
19993 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
19994 (INT_LOWPART): Delete.
19995 (extract_MB): Adjust.
19996 (extract_ME): Adjust.
19997 (print_operand): Adjust.
19999 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20001 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20002 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20003 (reg_or_logical_cint_operand, easy_fp_constant,
20004 logical_const_operand): Delete "CONST_DOUBLE" case.
20005 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20006 "HOST_BITS_PER_WIDE_INT == 64" test.
20007 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20008 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20009 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20010 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20012 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20013 CONST_DOUBLE DImode/VOIDmode case.
20014 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20015 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20016 CONST_DOUBLE VOIDmode case.
20017 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20018 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20019 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20020 Delete CONST_DOUBLE case.
20021 (splitters for mov FMOVE64 const_double): Delete
20022 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20023 "HOST_BITS_PER_WIDE_INT >= 64" test.
20024 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20026 (mov DI const_double): Delete.
20028 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20030 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20031 on op shows all bits zero in mode of a lowpart subreg, return zero.
20033 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20036 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20037 to save TFmode registers and DImode to save TImode registers for
20038 caller save operations.
20039 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20040 mark being partially clobbered since they only use the first
20043 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20044 and TDmode only use the upper 64-bits of each VSX register.
20046 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20048 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20049 (find_candidates_in_block): Disable slsr_process_phi.
20051 2013-05-03 Guozhi Wei <carrot@google.com>
20053 * coverage.c (coverage_obj_init): Move the construction of gcov
20055 (build_init_ctor): ... here.
20057 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20059 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20060 (slsr_cand_d): Redefine def_phi.
20061 (stride_status, phi_adjust_status, count_phis_status): New enums.
20062 (find_phi_def): New.
20063 (find_basis_for_base_expr): New.
20064 (find_basis_for_candidate): Handle hidden bases.
20065 (alloc_cand_and_find_basis): Handle phi candidates.
20066 (slsr_process_phi): New.
20067 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20068 (create_mul_imm_cand): Likewise.
20069 (create_add_ssa_cand): Exclude phi base candidates.
20070 (create_add_imm_cand): Likewise.
20071 (slsr_process_cast): Likewise.
20072 (slsr_process_copy): Likewise.
20073 (find_candidates_in_block): Handle phi candidates.
20074 (dump_candidate): Likewise.
20075 (unconditional_cands): Delete.
20076 (unconditional_cands_with_known_stride_p): Delete.
20077 (phi_dependent_cand_p): New.
20078 (cand_increment): Handle phi-dependent candidates.
20079 (replace_dependent): Delete.
20080 (replace_mult_candidate): New.
20081 (replace_unconditional_candidate): New.
20082 (incr_vec_index): Move to avoid forward reference.
20083 (create_add_on_incoming_edge): New.
20084 (create_phi_basis): New.
20085 (replace_dependents): Delete.
20086 (replace_conditional_candidate): New.
20087 (phi_add_costs): New.
20088 (replace_uncond_cands_and_profitable_phis): New.
20089 (record_increment): Handle phi adjustments.
20090 (record_phi_increments): New.
20091 (record_increments): Handle phi adjustments.
20092 (phi_incr_cost): New.
20093 (lowest_cost_path): Handle phis.
20094 (total_savings): Likewise.
20095 (analyze_increments): Likewise.
20096 (ncd_with_phi): New.
20097 (ncd_of_cand_and_phis): New.
20098 (nearest_common_dominator_for_cands): Handle phi increments.
20099 (all_phi_incrs_profitable): New.
20100 (replace_profitable_candidates): Handle phi-dependent candidates.
20101 (analyze_candidates_and_replace): Likewise.
20103 2013-05-03 Teresa Johnson <tejohnson@google.com>
20106 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20107 do not exceed REG_BR_PROB_BASE.
20109 2013-05-03 Jeff Law <law@redhat.com>
20111 PR tree-optimization/57144
20112 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20113 operand of the condition will bit into the new type when eliminating
20114 a cast feeding a condition.
20116 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20118 PR rtl-optimization/57130
20119 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20120 of COMPARE as in_code to the recursive call if needed.
20122 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20124 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20125 (enabled): Handle new members.
20126 * config/i386/sse.md (*vec_concatv2si): Merge from
20127 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20128 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20130 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20132 PR tree-optimization/57027
20133 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20134 for fnms opportunity, check we got the prerequisite kind
20135 of tree / gimple before using accessor functions.
20137 2013-05-03 Richard Biener <rguenther@suse.de>
20139 * double-int.h (lshift): New overload without precision
20140 and arith argument.
20141 (operator *=, operator +=, operator -=): Move ...
20142 * double-int.c (operator *=, operator +=, operator -=): ... here
20143 and implement more efficiently.
20144 (mul_double_with_sign): Remove.
20145 (lshift_double): Adjust to take unsinged shift argument, push
20146 dispatching code to callers.
20147 (mul_double_wide_with_sign): Add early out for callers that
20148 are not interested in high parts or overflow.
20149 (lshift): New function.
20150 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20151 dispatch code here.
20152 (lrotate, rrotate): Use logical shifts.
20153 * expr.c (get_inner_reference): Use lshift.
20154 * fixed-value.c (do_fixed_divide): Likewise.
20155 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20156 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20157 (indirect_refs_may_alias_p): Likewise.
20158 (stmt_kills_ref_p_1): Likewise.
20160 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20162 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20164 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20166 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20167 scalar form of FABD instruction.
20169 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20171 * lra-constraints.c (process_alt_operands): Add checking alt
20172 number to choose the best alternative.
20174 2013-05-02 Richard Biener <rguenther@suse.de>
20176 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20177 bitmap and its handling.
20178 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20180 2013-05-02 Richard Biener <rguenther@suse.de>
20182 PR middle-end/57140
20183 * tree-inline.c (copy_loops): Properly handle removed loops.
20184 (copy_cfg_body): Mark destination loops for fixup if source
20185 loops needed fixup.
20187 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
20190 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
20191 generating simple_return for naked functions.
20193 2013-05-02 Martin Jambor <mjambor@suse.cz>
20195 PR middle-end/56988
20196 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
20197 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
20199 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
20200 ipa_agg_replacement_value structures.
20201 (known_aggs_to_agg_replacement_list): Likewise.
20202 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
20203 (read_agg_replacement_chain): Likewise.
20204 (ipcp_transform_function): Also check that by_ref flags match.
20206 2013-05-02 Richard Biener <rguenther@suse.de>
20208 * graphds.h (struct graph): Add obstack member.
20209 * graphds.c (new_graph): Initialize obstack and allocate
20211 (add_edge): Allocate edge from the obstack.
20212 (free_graph): Free the obstack instead of all edges and vertices.
20214 2013-05-02 Teresa Johnson <tejohnson@google.com>
20216 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
20218 * cfg.c (update_bb_profile_for_threading): Ditto.
20219 * tree-inline.c (copy_bb): Ditto.
20220 (copy_edges_for_bb): Ditto.
20221 (initialize_cfun): Ditto.
20222 (copy_cfg_body): Ditto.
20223 (expand_call_inline): Ditto.
20224 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
20225 (estimate_node_size_and_time): Ditto.
20226 (inline_merge_summary): Ditto.
20227 * cgraphclones.c (cgraph_clone_edge): Ditto.
20228 (cgraph_clone_node): Ditto.
20229 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20230 (compute_trg_info): Ditto.
20232 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20234 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
20235 S reg when fp attribute set.
20236 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
20238 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20240 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
20242 (*and_one_cmplsi3_compare0_uxtw): Likewise.
20243 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
20244 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
20246 2013-05-02 Richard Biener <rguenther@suse.de>
20248 * tree-scalar-evolution.c (scev_info_hasher): Remove.
20249 (struct instantiate_cache_entry): New type.
20250 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
20251 (struct instantiate_cache_type): New type.
20252 (set_instantiated_value, get_instantiated_value): Remove.
20253 (get_instantiated_value_entry): New function.
20254 (instantiate_scev_name): Use the new cache and adjust.
20255 (instantiate_scev_poly): Adjust.
20256 (instantiate_scev_binary): Likewise.
20257 (instantiate_array_ref): Likewise.
20258 (instantiate_scev_convert): Likewise.
20259 (instantiate_scev_not): Likewise.
20260 (instantiate_scev_3): Likewise.
20261 (instantiate_scev_2): Likewise.
20262 (instantiate_scev_r): Likewise.
20263 (instantiate_scev): Likewise.
20264 (resolve_mixers): Likewise.
20266 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
20269 * lra-constraints.c (best_small_class_operands_num): Remove.
20270 (process_alt_operands): Remove small_class_operands_num. Take
20271 small classes operands into losers and only if the operand is not
20272 matched. Modify debugging output.
20273 (curr_insn_transform): Remove best_small_class_operands_num.
20276 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20278 * config/aarch64/aarch64-builtins.c
20279 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
20280 * config/aarch64/aarch64-simd-builtins.def
20281 (reduc_splus_): Add new modes.
20282 (reduc_uplus_): New.
20283 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
20284 (reduc_uplus_v4sf): Likewise.
20285 (reduc_splus_v4sf): Likewise.
20286 (aarch64_addv<mode>): Likewise.
20287 (reduc_uplus_<mode>): Likewise.
20288 (reduc_splus_<mode>): Likewise.
20289 (aarch64_addvv2di): Likewise.
20290 (reduc_uplus_v2di): Likewise.
20291 (reduc_splus_v2di): Likewise.
20292 (aarch64_addvv2si): Likewise.
20293 (reduc_uplus_v2si): Likewise.
20294 (reduc_splus_v2si): Likewise.
20295 (reduc_<sur>plus_<mode>): New.
20296 (reduc_<sur>plus_v2di): Likewise.
20297 (reduc_<sur>plus_v2si): Likewise.
20298 (reduc_<sur>plus_v4sf): Likewise.
20299 (aarch64_addpv4sf): Likewise.
20300 * config/aarch64/arm_neon.h
20301 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
20302 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
20303 add UNSPEC_SADDV, UNSPEC_UADDV.
20305 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
20307 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20309 * config/aarch64/arm_neon.h
20310 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
20312 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20314 * config/aarch64/aarch64-builtins
20315 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
20317 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20319 * config/aarch64/aarch64-simd-builtins.def
20320 (reduc_smax_): New.
20321 (reduc_smin_): Likewise.
20322 (reduc_umax_): Likewise.
20323 (reduc_umin_): Likewise.
20324 (reduc_smax_nan_): Likewise.
20325 (reduc_smin_nan_): Likewise.
20328 (smax): Update for V2SF, V4SF and V2DF modes.
20331 (smin_nan): Likewise.
20332 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
20333 (<su><maxmin><mode>3): ...This, refactor.
20334 (s<maxmin><mode>3): New.
20335 (<maxmin_uns><mode>3): Likewise.
20336 (reduc_<maxmin_uns>_<mode>): Refactor.
20337 (reduc_<maxmin_uns>_v4sf): Likewise.
20338 (reduc_<maxmin_uns>_v2si): Likewise.
20339 (aarch64_<fmaxmin><mode>: Remove.
20340 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
20342 (vmin<q>_f<32,64>): Likewise.
20343 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20345 (su): Add mappings for smax, smin, umax, umin.
20347 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20348 (FMAXMIN): Rename as...
20349 (FMAXMIN_UNS): ...This.
20351 (fmaxminv): Likewise.
20352 (fmaxmin): Likewise.
20354 (maxmin_uns_op): Likewise.
20356 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20358 * config/aarch64/arm_neon.h
20359 (vac<ge, gt><sd>_f<32, 64>): Rename to...
20360 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
20361 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
20363 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20365 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
20366 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
20368 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20370 * config/aarch64/aarch64-simd.md
20371 (vcond<mode>_internal): Handle special cases for constant masks.
20372 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
20373 (vcondu<mode><mode>): Likewise.
20374 (vcond<v_cmp_result><mode>): New.
20376 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20378 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
20379 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
20380 * config/aarch64/aarch64-simd-builtins.def
20381 (cmeq): Update to BUILTIN_VALLDI.
20386 * config/aarch64/arm_neon.h
20387 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
20388 to builtins or C as appropriate.
20390 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20392 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
20394 (cmhi): Rename to...
20396 * config/aarch64/aarch64-simd.md
20397 (simd_mode): Add SF.
20398 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
20399 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
20400 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
20401 (cstore<mode>_neg): ...This.
20402 * config/aarch64/iterators.md
20404 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
20405 (COMPARISONS): New.
20406 (UCOMPARISONS): Likewise.
20407 (optab): Add missing comparisons.
20413 (VCMP_S): Likewise.
20414 (VCMP_U): Likewise.
20415 (V_cmp_result): Add DF, SF modes.
20416 (v_cmp_result): Likewise.
20418 (vmtype): Likewise.
20419 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
20421 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
20423 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
20424 define_insn to define_insn_and_split.
20425 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
20426 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
20427 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
20428 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
20429 (thumb2_negscc): Likewise.
20431 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
20433 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
20435 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
20437 * config/arm/thumb2.md: Remove trailing whitespaces.
20439 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
20441 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
20442 Use gen_int_mode rather than GEN_INT.
20444 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
20446 * value-prof.c (stream_in_histogram_value): Remove the strayed
20449 2013-04-30 Richard Biener <rguenther@suse.de>
20451 PR middle-end/57122
20452 * cfghooks.c (split_edge): Properly check for the loop latch edge.
20454 2013-04-30 Richard Biener <rguenther@suse.de>
20456 PR middle-end/57107
20457 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
20459 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
20461 PR rtl-optimization/56957
20462 PR rtl-optimization/57105
20463 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
20464 variable. Use just INSN_UID for determining whether an insn
20465 should be only disconnected from the insn stream.
20466 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
20468 2013-04-30 Jakub Jelinek <jakub@redhat.com>
20470 PR tree-optimization/57104
20471 * tsan.c (instrument_expr): Don't instrument accesses to
20472 DECL_HARD_REGISTER VAR_DECLs.
20474 2013-04-30 Richard Biener <rguenther@suse.de>
20476 * function.h (loops_for_fn): New inline function.
20477 (set_loops_for_fn): Likewise.
20478 * cfgloop.h (place_new_loop): Add struct function parameter.
20479 (get_loop): Likewise.
20480 (get_loops): Likewise.
20481 (number_of_loops): Likewise.
20482 (fel_next): Adjust.
20483 (fel_init): Likewise.
20484 * cfg.c (get_loop_copy): Adjust.
20485 * cfgloop.c (flow_loops_dump): Likewise.
20486 (record_loop_exits): Likewise.
20487 (verify_loop_structure): Likewise.
20488 * cfgloopanal.c (mark_irreducible_loops): Likewise.
20489 (estimate_reg_pressure_cost): Likewise.
20490 (mark_loop_exit_edges): Likewise.
20491 * cfgloopmanip.c (place_new_loop): Likewise.
20492 (add_loop): Likewise.
20493 (duplicate_loop): Likewise.
20494 * graph.c (draw_cfg_nodes): Likewise.
20495 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
20496 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
20497 (extract_affine_chrec): Likewise.
20498 (build_scop_iteration_domain): Likewise.
20499 * graphite.c (graphite_initialize): Likewise.
20500 * ira-build.c (create_loop_tree_nodes): Likewise.
20501 (more_one_region_p): Likewise.
20502 (rebuild_regno_allocno_maps): Likewise.
20503 (mark_loops_for_removal): Likewise.
20504 (mark_all_loops_for_removal): Likewise.
20505 (remove_unnecessary_regions): Likewise.
20506 (ira_build): Likewise.
20507 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
20508 * loop-init.c (fix_loop_structure): Likewise.
20509 (gate_rtl_move_loop_invariants): Likewise.
20510 (gate_rtl_unswitch): Likewise.
20511 (gate_rtl_unroll_and_peel_loops): Likewise.
20512 (rtl_doloop): Likewise.
20513 * lto-streamer-in.c (input_cfg): Likewise.
20514 * lto-streamer-out.c (output_cfg): Likewise.
20515 * modulo-sched.c (sms_schedule): Likewise.
20516 * predict.c (tree_estimate_probability): Likewise.
20517 (tree_estimate_probability_driver): Likewise.
20518 (estimate_loops): Likewise.
20519 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
20520 (move_sese_region_to_fn): Likewise.
20521 (debug_loop_num): Likewise.
20522 * tree-chrec.c (chrec_evaluate): Likewise.
20523 (hide_evolution_in_other_loops_than_loop): Likewise.
20524 (chrec_component_in_loop_num): Likewise.
20525 (reset_evolution_in_loop): Likewise.
20526 (evolution_function_is_invariant_rec_p): Likewise.
20527 * tree-if-conv.c (main_tree_if_conversion): Likewise.
20528 * tree-inline.c (copy_loops): Likewise.
20529 (copy_cfg_body): Likewise.
20530 (tree_function_versioning): Likewise.
20531 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
20532 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
20534 (add_to_evolution_1): Likewise.
20535 (scev_const_prop): Likewise.
20536 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
20537 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
20538 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
20539 (tree_ssa_lim_initialize): Likewise.
20540 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
20541 (verify_loop_closed_ssa): Likewise.
20542 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
20543 (tree_ssa_loop_im): Likewise.
20544 (tree_ssa_loop_unswitch): Likewise.
20545 (tree_vectorize): Likewise.
20546 (check_data_deps): Likewise.
20547 (tree_ssa_loop_ivcanon): Likewise.
20548 (tree_ssa_loop_bounds): Likewise.
20549 (tree_complete_unroll): Likewise.
20550 (tree_complete_unroll_inner): Likewise.
20551 (tree_parallelize_loops): Likewise.
20552 (tree_ssa_loop_prefetch): Likewise.
20553 (tree_ssa_loop_ivopts): Likewise.
20554 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
20555 * tree-vectorizer.c (vectorize_loops): Likewise.
20557 2013-04-29 Mike Frysinger <vapier@gentoo.org>
20559 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
20560 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
20561 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
20562 with EABI_LINK_SPEC.
20564 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
20567 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
20570 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
20573 * lra-constraints.c (process_alt_operands): Discourage a bit more
20574 using memory for pseudos. Print cost dump for alternatives.
20575 Modify cost values for conflicts with early clobbers.
20576 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
20578 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
20581 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
20583 2013-04-29 Ian Bolton <ian.bolton@arm.com>
20585 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
20586 from/to S register.
20587 (movdi_aarch64): Support LDR/STR from/to D register.
20589 2013-04-29 Ian Bolton <ian.bolton@arm.com>
20591 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
20592 or higher by default.
20594 2013-04-29 Richard Biener <rguenther@suse.de>
20596 PR middle-end/57075
20597 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
20598 even if not adding abnormal edges for calls that can make
20601 2013-04-29 Richard Biener <rguenther@suse.de>
20603 PR middle-end/57103
20604 * tree-cfg.c (move_stmt_op): Fix condition under which to update
20606 (move_stmt_r): Remove redundant checking.
20608 2013-04-29 Teresa Johnson <tejohnson@google.com>
20611 * basic-block.h (apply_scale): New function.
20612 (apply_probability): Use apply_scale.
20613 * gimple-streamer-in.c (input_bb): Ditto.
20614 * lto-streamer-in.c (input_cfg): Ditto.
20615 * lto-cgraph.c (merge_profile_summaries): Ditto.
20616 * tree-optimize.c (execute_fixup_cfg): Ditto.
20617 * tree-inline.c (copy_bb): Update comment to use apply_scale.
20618 (copy_edges_for_bb): Ditto.
20619 (copy_cfg_body): Ditto.
20621 2013-04-29 Tom de Vries <tom@codesourcery.com>
20623 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
20624 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
20625 (tail_merge_optimize): Handle current_loops == NULL.
20627 2013-04-26 Jeff Law <law@redhat.com>
20629 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
20630 (simplify_cond_using_ranges): Generalize code to simplify
20631 COND_EXPRs where one argument is a constant and the other
20632 is an SSA_NAME created by an integral type conversion.
20634 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20636 * config/arm/arm.md (store_minmaxsi): Use only when
20637 optimize_insn_for_size_p.
20639 2013-04-29 Christian Bruel <christian.bruel@st.com>
20642 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
20644 2013-04-29 Richard Biener <rguenther@suse.de>
20646 PR middle-end/57089
20647 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
20648 loop tree make sure to schedule a fixup for the child as well.
20649 (expand_omp_for_generic): Properly add loops.
20650 (expand_omp_for_static_nochunk): Likewise.
20651 (expand_omp_for_static_chunk): Likewise.
20652 (expand_omp_for): For the degenerate case fixup loops.
20653 (expand_omp_sections): Fix default bb placement in loops.
20654 (expand_omp_atomic_pipeline): Properly add loops.
20656 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20658 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
20660 2013-04-29 Tom de Vries <tom@codesourcery.com>
20662 * tree-ssa-tail-merge.c: Update header comment.
20664 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20666 * config/aarch64/arm_neon.h
20667 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
20668 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
20669 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
20670 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
20671 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
20672 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
20673 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
20674 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
20676 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20678 * config/aarch64/aarch64-simd.md
20679 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
20680 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
20681 fix_trunc, fixuns_trunc.
20682 (ftrunc<VDQF:mode>2): New.
20683 * config/aarch64/iterators.md (optab): Add fix, fixuns.
20684 (fix_trunc_optab): New.
20686 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20688 * config/aarch64/aarch64-builtins.c
20689 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
20690 iceilf, lround, iroundf.
20692 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
20695 * config/i386/i386.h (enum ix86_tune_indices)
20696 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
20697 New, split from X86_TUNE_INTER_UNIT_MOVES.
20698 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
20699 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
20700 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
20701 (TARGET_INTER_UNIT_MOVES): Remove.
20702 * config/i386/i386.c (initial_ix86_tune_features): Update.
20703 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
20704 (ix86_expand_convert_uns_didf_sse): Use
20705 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
20706 (ix86_expand_vector_init_one_nonzero): Ditto.
20707 (ix86_expand_vector_init_interleave): Ditto.
20708 (inline_secondary_memory_needed): Return true for moves from SSE class
20709 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
20710 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
20711 * config/i386/constraints.md (Yi, Ym): Depend on
20712 TARGET_INTER_UNIT_MOVES_TO_VEC.
20713 (Yj, Yn): New constraints.
20714 * config/i386/i386.md (*movdi_internal): Change constraints of
20715 operand 1 from Yi to Yj and from Ym to Yn.
20716 (*movsi_internal): Ditto.
20717 (*movdf_internal): Ditto.
20718 (*movsf_internal): Ditto.
20719 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
20720 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
20721 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
20722 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
20723 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
20724 * config/i386/sse.md (movdi_to_sse): Ditto.
20725 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
20726 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
20727 TARGET_INTER_UNIT_MOVES.
20728 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
20729 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
20730 instead of TARGET_INTER_UNIT_MOVES.
20731 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
20732 operand 1 from Yi to Yj and from Ym to Yn.
20734 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20736 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
20737 (float_truncate_hi_): Likewise.
20738 (float_extend_lo_): Likewise.
20739 (float_truncate_lo_): Likewise.
20740 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
20741 (aarch64_float_extend_lo_v2df): Likewise.
20742 (vec_unpacks_hi_v4sf): Likewise.
20743 (aarch64_float_truncate_lo_v2sf): Likewise.
20744 (aarch64_float_truncate_hi_v4sf): Likewise.
20745 (vec_pack_trunc_v2df): Likewise.
20746 (vec_pack_trunc_df): Likewise.
20748 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20750 * config/aarch64/aarch64-builtins.c
20751 (aarch64_fold_builtin): Fold float conversions.
20752 * config/aarch64/aarch64-simd-builtins.def
20753 (floatv2si, floatv4si, floatv2di): New.
20754 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
20755 * config/aarch64/aarch64-simd.md
20756 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
20757 * config/aarch64/iterators.md (FLOATUORS): New.
20758 (optab): Add float, floatuns.
20759 (su_optab): Likewise.
20761 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20763 * config/aarch64/aarch64-builtins.c
20764 (aarch64_builtin_vectorized_function): Use new names for
20766 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
20767 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
20768 (fcvtzu): Split as...
20769 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
20770 (fcvtas): Split as...
20771 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
20772 (fcvtau): Split as...
20773 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
20774 (fcvtps): Split as...
20775 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
20776 (fcvtpu): Split as...
20777 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
20778 (fcvtms): Split as...
20779 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
20780 (fcvtmu): Split as...
20781 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
20782 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
20783 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
20784 (lfrintnusf, lfrintnudf): Likewise.
20785 * config/aarch64/aarch64-simd.md
20786 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
20788 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
20789 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
20790 (fcvt_pattern): Likewise.
20792 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20794 * config/aarch64/aarch64-simd.md
20795 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
20796 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
20798 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20800 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
20801 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
20802 (vrnd<a,m,n,p>_f32): Implement using builtins.
20803 (vrnd<i,x><q>_f<32, 64>): New.
20805 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20807 * config/aarch64/aarch64-builtins.c
20808 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
20809 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
20810 (frintz): Rename to...
20812 (frintp): Rename to...
20814 (frintm): Rename to...
20816 (frinti): Rename to...
20817 (nearbyint): ...this.
20818 (frintx): Rename to...
20820 (frinta): Rename to...
20822 * config/aarch64/aarch64-simd.md
20823 (aarch64_frint<frint_suffix><mode>): Delete.
20824 (<frint_pattern><mode>2): Convert to insn.
20825 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
20826 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
20827 (frint_pattern): Likewise.
20828 (frint_suffix): Likewise.
20830 2013-04-29 Richard Biener <rguenther@suse.de>
20832 PR tree-optimization/57081
20833 * loop-init.c: Include tree-flow.h.
20834 (loop_optimizer_finalize): Free number of iteration estimates.
20835 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
20837 2013-04-29 Jakub Jelinek <jakub@redhat.com>
20839 PR tree-optimization/57083
20840 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
20841 non-singleton shift count range, zero extend low_bound for uns case.
20843 * config/i386/predicates.md (general_vector_operand): New predicate.
20844 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
20845 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
20846 if they aren't nonimmediate operands. If their original values
20847 satisfy const_vector_equal_evenodd_p, don't shift them.
20848 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
20849 predicates. For the SSE4.1 case force operands[{1,2}] into registers
20850 if not nonimmediate_operand.
20851 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
20852 instead of register_operand.
20853 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
20855 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
20857 * stor-layout.c (finalize_size_functions): Allocate a structure and
20858 reset cfun before dumping the functions.
20860 2013-04-27 Jakub Jelinek <jakub@redhat.com>
20862 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
20865 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
20866 use xop_pmacsdqh if uns_p.
20867 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
20868 the immediate rotate count.
20870 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
20872 * rtl.h (struct rtx_def): Add comment for field jump.
20873 (LRA_SUBREG_P): New macro.
20874 * recog.c (register_operand): Check LRA_SUBREG_P.
20875 * lra.c (lra): Add note at the end of RTL code. Align non-empty
20877 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
20878 (lra_final_code_change): Skip subreg change for operators.
20879 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
20880 if there are no operand changes.
20881 * lra-constraints.c (curr_insn_set): New.
20882 (match_reload): Set LRA_SUBREG_P.
20883 (emit_spill_move): Ditto.
20884 (check_and_process_move): Use curr_insn_set. Process only single
20885 set insns. Don't initialize sec_mem_p and change_p.
20886 (simplify_operand_subreg): Use LRA_SUBREG_P.
20887 (reg_in_class_p): New function.
20888 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
20889 of #ifdef. Add code to remove cycling.
20890 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
20891 non-null disp. Reload inner instead of disp when base and index
20892 are null. Try to put lo_sum into register.
20893 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
20894 (check_and_process_move): Move code for move cost check to
20895 simple_move_p. Remove equiv_substitution.
20896 (simple_move_p): New function.
20897 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
20898 curr_insn_set. Call check_and_process_move only for single set
20899 insns. Use the new function. Move call of check_and_process_move
20900 after operand equiv substitution and address process.
20902 2013-04-26 Jakub Jelinek <jakub@redhat.com>
20905 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
20906 with nonlocal goto receivers or returns twice calls, ignore
20907 unininitialized values from abnormal edges to nl goto receiver
20908 or returns twice call.
20910 2013-04-26 Jakub Jelinek <jakub@redhat.com>
20912 PR tree-optimization/57051
20913 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
20914 and VEC_RSHIFT_EXPR if shift count is a multiple of element
20917 2013-04-26 Richard Biener <rguenther@suse.de>
20919 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
20920 (expand_omp_taskreg): Likewise. Mark loops for fixup.
20921 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
20922 (fixup_loop_arrays_after_move): New function.
20923 (move_sese_region_to_fn): Properly outline the loop tree parts
20924 of the SESE region.
20926 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
20928 * config/i386/i386.md (type, unit): Fix long lines.
20930 2013-04-26 Richard Biener <rguenther@suse.de>
20932 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
20933 (lto-streamer-out.o): Likewise.
20934 * cfgloop.c (init_loops_structure): Export, add struct function
20935 argument and adjust.
20936 (flow_loops_find): Adjust.
20937 * cfgloop.h (enum loop_estimation): Add EST_LAST.
20938 (init_loops_structure): Declare.
20939 * lto-streamer-in.c: Include cfgloop.h.
20940 (input_cfg): Input the loop tree.
20941 * lto-streamer-out.c: Include cfgloop.h.
20942 (output_cfg): Output the loop tree.
20943 (output_struct_function_base): Do not drop PROP_loops.
20945 2013-03-26 Richard Biener <rguenther@suse.de>
20947 * tree-cfg.c (execute_build_cfg): Build the loop tree.
20948 (pass_build_cfg): Provide PROP_loops.
20949 (move_sese_region_to_fn): Remove loops that are outlined into fn
20951 * tree-inline.c: Include cfgloop.h.
20952 (initialize_cfun): Do not drop PROP_loops.
20953 (copy_loops): New function.
20954 (copy_cfg_body): Copy loop structure.
20955 (tree_function_versioning): Initialize destination loop tree.
20956 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
20957 (pass_parallelize_loops): Do IL verification.
20958 * loop-init.c (loop_optimizer_init): Fixup loops if required.
20959 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
20960 the CFG make sure we fixup loops as well.
20961 * tree-ssa-tail-merge.c: Include cfgloop.h.
20962 (replace_block_by): When merging loop latches mark loops for fixup.
20963 * lto-streamer-out.c (output_struct_function_base): Drop
20964 PROP_loops for now.
20965 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
20966 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
20967 * ipa-split.c: Include cfgloop.h.
20968 (split_function): Add the new return block to the loop tree root.
20969 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
20970 whether we have removed the forwarder block.
20971 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
20972 * cfgloop.h (place_new_loop): Declare.
20973 * cfgloopmanip.c (place_new_loop): Export.
20974 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
20975 (tree-switch-conversion.o): Likewise.
20976 (tree-complex.o): Likewise.
20977 (tree-inline.o): Likewise.
20978 (tree-ssa-tailmerge.o): Likewise.
20979 (ipa-split.o): Likewise.
20980 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
20981 (tree-ssa-copy.o): Likewise.
20982 * tree-switch-conversion.c: Include cfgloop.h
20983 (process_switch): If we emit a bit-test cascade, schedule loops
20985 * tree-complex.c: Include cfgloop.h.
20986 (expand_complex_div_wide): Properly add new basic-blocks to loops.
20987 * asan.c: Include cfgloop.h.
20988 (create_cond_insert_point): Properly add new basic-blocks to
20989 loops, schedule loop fixup.
20990 * cfgloop.c (verify_loop_structure): Check that looks are not
20992 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
20994 (expand_omp_for_generic): Likewise.
20995 (expand_omp_sections): Likewise.
20996 (expand_omp_atomic_pipeline): Schedule loops for fixup.
20997 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
20998 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
20999 is initialized, not when loops are present.
21000 * tree-parloops.c (parallelize_loops): Remove checking here.
21001 * passes.c (init_optimization_passes): Schedule a copy-propagation
21002 pass before complete unrolling of inner loops.
21004 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21006 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21007 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21008 (colorize_init): Add argument to _WIN32 version.
21009 * toplev.c: Include diagnostic-color.h.
21010 (process_options): Default to -fdiagnostics-color=auto if
21011 GCC_COLORS env var is in the environment.
21012 * common.opt (fdiagnostics-color=): Add Var and Init.
21013 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21014 env var is in the environment, the default is auto rather than never.
21016 * diagnostic.h (file_name_as_prefix): Add context argument.
21017 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21018 the string as locus.
21019 * langhooks.c (lhd_print_error_function): Adjust caller.
21021 2013-04-25 Lawrence Crowl <crowl@google.com>
21023 * var-tracking.c (shared_hash_def::htab):
21024 Change type to hash_table. Update dependent calls and types.
21026 2013-04-25 Lawrence Crowl <crowl@google.com>
21028 * Makefile.in: Update as needed below.
21030 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21031 Move declaration to after the type's method definitons.
21033 * attribs.c (htab_t scoped_attributes::attribute_hash):
21034 Change type to hash_table. Update dependent calls and types.
21036 * bitmap.c (htab_t bitmap_desc_hash):
21037 Change type to hash_table. Update dependent calls and types.
21039 * cselib.c (htab_t cselib_hash_table):
21040 Change type to hash_table. Update dependent calls and types.
21042 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21043 (hash_string_slot_node): Move implementation into lto-streamer.h
21044 struct string_slot_hasher.
21045 (eq_string_slot_node): Likewise.
21047 * data-streamer-out.c: Update output_block::string_hash_table
21048 dependent calls and types.
21050 * dwarf2cfi.c (htab_t trace_index):
21051 Change type to hash_table. Update dependent calls and types.
21053 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21054 Change type to hash_table. Update dependent calls and types.
21055 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21056 (htab_t optimize_external_refs::map): Likewise.
21057 (htab_t output_comp_unit::extern_map): Likewise.
21058 (htab_t output_comdat_type_unit::extern_map): Likewise.
21059 (htab_t output_macinfo::macinfo_htab): Likewise.
21060 (htab_t optimize_location_lists::htab): Likewise.
21061 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21063 * except.c (htab_t ehspec_hash_type):
21064 Change type to hash_table. Update dependent calls and types.
21065 (assign_filter_values::ttypes): Likewise.
21066 (assign_filter_values::ehspec): Likewise.
21067 (sjlj_assign_call_site_values::ar_hash): Likewise.
21068 (convert_to_eh_region_ranges::ar_hash): Likewise.
21070 * gcse.c (htab_t pre_ldst_table):
21071 Change type to hash_table. Update dependent calls and types.
21073 * ggc-common.c (htab_t saving_htab):
21074 Change type to hash_table. Update dependent calls and types.
21075 (htab_t loc_hash): Likewise.
21076 (htab_t ptr_hash): Likewise.
21077 (call_count): Rename ggc_call_count.
21078 (call_alloc): Rename ggc_call_alloc.
21079 (loc_descriptor): Rename make_loc_descriptor.
21080 (add_statistics): Rename ggc_add_statistics.
21082 * ggc-common.c (saving_htab):
21083 Change type to hash_table. Update dependent calls and types.
21085 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21086 (push_gimplify_context): Likewise.
21087 (pop_gimplify_context): Likewise.
21088 (struct gimple_temp_hash_elt): Added.
21089 (struct gimplify_hasher): Likewise.
21090 (struct gimplify_ctx.temp_htab):
21091 Change type to hash_table. Update dependent calls and types.
21093 * gimple-fold.c: Include gimplify-ctx.h.
21095 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21096 Change type to hash_table. Update dependent calls and types.
21097 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21098 avoid potential global name collision.
21100 * gimplify.c: Include gimplify-ctx.h.
21101 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21102 (htab_t gimplify_ctx::temp_htab):
21103 Update dependent calls and types for new type hash_table.
21104 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21105 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21107 * gimplify-ctx.h: New.
21108 (struct gimple_temp_hash_elt): Move from gimplify.c.
21109 (class gimplify_hasher): New.
21110 (struct gimplify_ctx): Move from gimple.h.
21111 (htab_t gimplify_ctx::temp_htab):
21112 Change type to hash_table. Update dependent calls and types.
21114 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21115 (htab_t ivs_params::newivs_index):
21116 Change type to hash_table. Update dependent calls and types.
21117 (htab_t ivs_params::params_index): Likewise.
21118 (htab_t print_generated_program::params_index): Likewise.
21119 (htab_t gloog::newivs_index): Likewise.
21120 (htab_t gloog::params_index): Likewise.
21122 * graphite.c: Include graphite-htab.h.
21123 4htab_t graphite_transform_loops::bb_pbb_mapping):
21124 Change type to hash_table. Update dependent calls and types.
21126 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21127 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21128 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21130 * graphite-dependences.c: Include graphite-htab.h.
21131 (loop_is_parallel_p): Change hash table type of parameter.
21133 * graphite-htab.h: New.
21134 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21135 (extern find_pbb_via_hash): Move from graphite-poly.h.
21136 (extern loop_is_parallel_p): Move from graphite-poly.h.
21137 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21139 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21140 (extern loop_is_parallel_p): Move to graphite-htab.h.
21141 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21143 * haifa-sched.c (htab_t delay_htab):
21144 Change type to hash_table. Update dependent calls and types.
21145 (htab_t delay_htab_i2): Likewise.
21147 * ira-color.c (htab_t allocno_hard_regs_htab):
21148 Change type to hash_table. Update dependent calls and types.
21150 * ira-costs.c (htab_t cost_classes_htab):
21151 Change type to hash_table. Update dependent calls and types.
21153 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21154 Change type to hash_table. Update dependent calls and types.
21156 * loop-iv.c (htab_t bivs):
21157 Change type to hash_table. Update dependent calls and types.
21159 * loop-unroll.c (htab_t opt_info::insns_to_split):
21160 Change type to hash_table. Update dependent calls and types.
21161 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21163 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21164 (struct string_slot_hasher): New.
21165 (htab_t output_block::string_hash_table):
21166 Change type to hash_table. Update dependent calls and types.
21168 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21169 (htab_t file_name_hash_table):
21170 Change type to hash_table. Update dependent calls and types.
21172 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21175 * lto-streamer.c (htab_t tree_htab):
21176 Change type to hash_table. Update dependent calls and types.
21178 * omp-low.c: Include gimplify-ctx.h.
21180 * passes.c (htab_t name_to_pass_map):
21181 Change type to hash_table. Update dependent calls and types.
21182 (pass_traverse): Rename to passes_pass_traverse.
21184 * plugin.c (htab_t event_tab):
21185 Change type to hash_table. Update dependent calls and types.
21187 * postreload-gcse.c (htab_t expr_table):
21188 Change type to hash_table. Update dependent calls and types.
21189 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
21191 * sese.c (debug_rename_map_1): Make extern.
21192 (htab_t copy_bb_and_scalar_dependences::rename_map):
21193 Change type to hash_table. Update dependent calls and types.
21195 * sese.h (extern debug_rename_map): Move to .c file.
21197 * store-motion.c (htab_t store_motion_mems_table):
21198 Change type to hash_table. Update dependent calls and types.
21200 * trans-mem.c (htab_t tm_new_mem_hash):
21201 Change type to hash_table. Update dependent calls and types.
21203 * tree-browser.c (htab_t TB_up_ht):
21204 Change type to hash_table. Update dependent calls and types.
21206 * tree-cfg.c (htab_t discriminator_per_locus):
21207 Change type to hash_table. Update dependent calls and types.
21209 * tree-complex.c: Include tree-hasher.h
21210 (htab_t complex_variable_components):
21211 Change type to hash_table. Update dependent calls and types.
21213 * tree-eh.c (htab_t finally_tree):
21214 Change type to hash_table. Update dependent calls and types.
21216 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
21217 struct int_tree_hasher.
21218 (extern int_tree_map_eq): Likewise.
21219 (uid_decl_map_hash): Removed.
21220 (extern decl_tree_map_eq): Likewise.
21222 * tree-hasher.h: New.
21223 (struct int_tree_hasher): New.
21224 (typedef int_tree_htab_type): New.
21226 * tree-inline.c: Include gimplify-ctx.h.
21228 * tree-mudflap.c: Include gimplify-ctx.h.
21230 * tree-parloops.c: Include tree-hasher.h.
21231 (htab_t eliminate_local_variables_stmt::decl_address):
21232 Change type to hash_table. Update dependent calls and types.
21233 (htab_t separate_decls_in_region::decl_copies): Likewise.
21235 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
21236 Change type to hash_table. Update dependent calls and types.
21238 * tree-sra.c (candidates):
21239 Change type to hash_table. Update dependent calls and types.
21241 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
21243 (int_tree_map_hash): Likewise.
21245 * tree-ssa-dom.c (htab_t avail_exprs):
21246 Change type to hash_table. Update dependent calls and types.
21248 * tree-ssa-live.c (var_map_base_init::tree_to_index):
21249 Change type to hash_table. Update dependent calls and types.
21251 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
21252 Change type to hash_table. Update dependent calls and types.
21254 * tree-ssa-phiopt.c (seen_ssa_names):
21255 Change type to hash_table. Update dependent calls and types.
21257 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
21258 Change type to hash_table. Update dependent calls and types.
21260 * tree-ssa-uncprop.c (equiv):
21261 Change type to hash_table. Update dependent calls and types.
21263 2013-04-25 Jakub Jelinek <jakub@redhat.com>
21265 PR rtl-optimization/57003
21266 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
21267 call note_stores with kill_clobbered_value callback again after
21268 killing regs_invalidated_by_call.
21270 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21272 * config/aarch64/aarch64-simd.md
21273 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
21274 (aarch64_simd_bsl<mode>): Likewise.
21275 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
21277 2013-04-25 Marek Polacek <polacek@redhat.com>
21279 PR tree-optimization/57066
21280 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
21282 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21284 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
21286 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21288 * config/aarch64/aarch64-builtins.c
21289 (aarch64_fold_builtin): New.
21290 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
21291 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
21292 * config/aarch64/aarch64-simd-builtins.def (abs): New.
21293 * config/aarch64/arm_neon.h
21294 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
21296 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21297 Tejas Belagod <tejas.belagod@arm.com>
21299 * config/aarch64/aarch64-builtins.c
21300 (aarch64_gimple_fold_builtin): New.
21301 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
21302 * config/aarch64/aarch64-simd-builtins.def (addv): New.
21303 * config/aarch64/aarch64-simd.md (addpv4sf): New.
21304 (addvv4sf): Update.
21305 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
21307 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21309 * config/aarch64/aarch64.md
21310 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
21312 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21314 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
21315 (*ngcsi_uxtw): New pattern.
21317 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21318 Julian Brown <julian@codesourcery.com>
21320 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
21321 (TB_DREG): Add T_V4HF.
21322 (v4hf_UP): New macro.
21323 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21324 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21325 Handle initialisation of V4HF. Adjust initialisation of reinterpret
21327 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21328 (arm_vector_mode_supported_p): Handle V4HF.
21329 (arm_mangle_map): Handle V4HFmode.
21330 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
21331 * config/arm/arm_neon_builtins.def: Add entries for
21332 vcvtv4hfv4sf, vcvtv4sfv4hf.
21333 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
21334 (neon_vcvtv4hfv4sf): Likewise.
21335 * config/arm/neon-gen.ml: Handle half-precision floating point
21337 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
21338 * config/arm/arm_neon.h: Regenerate.
21339 * config/arm/neon.ml (type elts): Add F16.
21340 (type vectype): Add T_float16x4, T_floatHF.
21341 (type vecmode): Add V4HF.
21342 (type features): Add Requires_FP_bit feature.
21343 (elt_width): Handle F16.
21344 (elt_class): Likewise.
21345 (elt_of_class_width): Likewise.
21346 (mode_of_elt): Refactor.
21347 (type_for_elt): Handle F16, fix error messages.
21348 (vectype_size): Handle T_float16x4.
21349 (vcvt_sh): New function.
21350 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
21351 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
21352 (string_of_mode): Handle V4HF.
21353 * doc/arm-neon-intrinsics.texi: Regenerate.
21355 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21357 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
21358 format specifier in 'X' case.
21360 2013-04-25 Alan Modra <amodra@gmail.com>
21363 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
21364 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
21365 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
21366 Repeat for many other rotate/shift and mask patterns using subregs.
21367 Name lshiftrt insns.
21368 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
21369 on WORDS_BIG_ENDIAN.
21371 2013-04-25 Alan Modra <amodra@gmail.com>
21373 * config.gcc: Support little-endian powerpc-linux targets.
21374 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
21375 (LINK_OS_LINUX_SPEC): Define.
21376 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
21377 Preserve MASK_LITTLE_ENDIAN.
21378 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
21379 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
21380 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
21381 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
21382 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
21383 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
21384 Correct fp word order for little-endian. Don't shift toc entries
21385 smaller than a word for little-endian.
21386 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
21387 (bswapdi2 splits): Correct low-part subreg for little-endian.
21388 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
21389 low/high where such is correct only for be.
21390 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
21391 little-endian for -mcall-aixdesc.
21393 2013-04-25 Alan Modra <amodra@gmail.com>
21395 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
21396 replace_equiv_address_nv.
21398 2013-04-25 Alan Modra <amodra@gmail.com>
21400 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
21402 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21405 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21406 * rtl.h (struct rtx_def): ...
21408 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21410 PR rtl-optimizations/57046
21411 * lra-constraints (split_reg): Set up lra_risky_transformations_p
21412 for multi-reg splits.
21414 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
21416 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
21418 2013-04-24 Sterling Augustine <saugustine@google.com>
21420 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
21421 (comp_dir_string, debug_str_dwo_section): New.
21422 (DEBUG_STR_DWO_SECTION): Rename to ...
21423 (DEBUG_DWO_STR_SECTION): ... this.
21424 (DEBUG_NORM_STR_SECTION): Delete.
21425 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
21426 (DEBUG_STR_DWO_SECTION_FLAGS): New.
21427 (find_AT_string): Move most logic to ...
21428 (find_AT_string_in_table): ... here. New.
21429 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
21430 add_skeleton_AT_string. Delete logic.
21431 (output_skeleton_debug_sections): Remove call to
21432 add_top_level_skeleton_die_attrs.
21433 (add_comp_dir_attribute): Move logic to comp_dir_string.
21434 (dwarf2out_init): Initialize debug_str_dwo_section.
21435 (output_indirect_string): Call find_string_form.
21436 (output_indirect_strings): Rewrite.
21437 (prune_unused_types): Empty skeleton_debug_str_hash.
21438 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
21439 (dwarf2out_finish): Call output_indirect_strings.
21441 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
21443 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
21445 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21447 * rtl.h (struct rtx_def): Add comment for field jump.
21448 (LRA_SUBREG_P): New macro.
21449 * recog.c (register_operand): Check LRA_SUBREG_P.
21450 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21452 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21453 (lra_final_code_change): Skip subreg change for operators.
21454 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21455 if there are no operand changes.
21456 * lra-constraints.c (curr_insn_set): New.
21457 (match_reload): Set LRA_SUBREG_P.
21458 (emit_spill_move): Ditto.
21459 (check_and_process_move): Use curr_insn_set. Process only single
21460 set insns. Don't initialize sec_mem_p and change_p.
21461 (simplify_operand_subreg): Use LRA_SUBREG_P.
21462 (reg_in_class_p): New function.
21463 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21464 of #ifdef. Add code to remove cycling.
21465 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21466 non-null disp. Reload inner instead of disp when base and index
21467 are null. Try to put lo_sum into register.
21468 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21469 (check_and_process_move): Move code for move cost check to
21470 simple_move_p. Remove equiv_substitution.
21471 (simple_move_p): New function.
21472 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21473 curr_insn_set. Call check_and_process_move only for single set
21474 insns. Use the new function. Move call of check_and_process_move
21475 after operand equiv substitution and address process.
21477 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
21479 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
21480 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
21481 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
21483 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
21485 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
21487 2013-04-24 Marek Polacek <polacek@redhat.com>
21489 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
21490 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
21491 (select_loops_exit_conditions): Likewise.
21492 (number_of_iterations_for_all_loops): Likewise.
21493 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
21494 (scev_analysis): Likewise.
21496 2013-04-02 Catherine Moore <clm@codesourcery.com>
21497 Chao-ying Fu <fu@mips.com>
21499 * config/mips/micromips.md (jraddiusp): New pattern.
21500 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
21501 instruction if possible.
21503 2013-04-24 Alan Modra <amodra@gmail.com>
21505 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
21507 2013-04-24 Julian Brown <julian@codesourcery.com>
21508 Chung-Lin Tang <cltang@codesourcery.com>
21510 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
21511 dependency behavior in enumeration type DIE generation. Add TODO note
21512 to comments about future DW_FORM_sdata/udata re-work of related code.
21514 2013-04-23 Lawrence Crowl <crowl@google.com>
21516 * Makefile.in: Update as needed below.
21518 * hash-table.h (class hash_table):
21519 Correct many methods with parameter types compare_type to the correct
21520 value_type. (Correct code was unlikely to notice the change.)
21521 (hash_table::elements_with_deleted) New.
21522 (class hashtable::iterator): New.
21523 (hashtable::begin()): New.
21524 (hashtable::end()): New.
21525 (FOR_EACH_HASH_TABLE_ELEMENT): New.
21527 * statistics.c (statistics_hashes):
21528 Change type to hash_table. Update dependent calls and types.
21530 * tree-into-ssa.c (var_infos):
21531 Change type to hash_table. Update dependent calls and types.
21533 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
21534 Change type to hash_table. Update dependent calls and types.
21536 * tree-ssa-loop-im.c (struct mem_ref.refs):
21537 Change type to hash_table. Update dependent calls and types.
21539 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
21540 Change type to hash_table. Update dependent calls and types.
21542 * tree-ssa-sccvn.c (vn_tables_s::nary):
21543 Change type to hash_table. Update dependent calls and types.
21544 (vn_tables_s::phis): Likewise.
21545 (vn_tables_s::references): Likewise.
21547 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
21548 (vn_reference_eq): Update parameter and return types.
21550 * tree-ssa-structalias.c (pointer_equiv_class_table):
21551 Change type to hash_table. Update dependent calls and types.
21552 (location_equiv_class_table): Likewise.
21554 * tree-vect-data-refs.c: Consequential changes for making
21555 peeling a hash_table.
21557 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
21558 (destroy_loop_vec_info): Dependent hash_table update.
21560 * tree-vectorizer.h (peeling_htab):
21561 Change type to hash_table. Update dependent calls and types.
21563 2013-04-23 Shiva Chen <shiva0217@gmail.com>
21565 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
21566 to check the register content is equal or not.
21567 * lra-constraints.c (match_reload): Use lra_assign_reg_val
21568 to assign register content record.
21569 * lra-eliminations.c (update_reg_eliminate): Use
21570 lra_update_reg_val_offset to update register content offset.
21571 * lra-int.h (struct lra_reg): Add offset member.
21572 (lra_reg_val_equal_p): New static inline function.
21573 (lra_update_reg_val_offset): New static inline function.
21574 (lra_assign_reg_val): New static inline function.
21575 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
21576 to assign register content record.
21577 (initialize_lra_reg_info_element): Initial offset to zero.
21579 2013-04-23 Catherine Moore <clm@codesourcery.com>
21581 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
21582 operands. Record compression.
21584 2013-04-23 Xinliang David Li <davidxl@google.com>
21586 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
21588 2013-04-23 Richard Biener <rguenther@suse.de>
21590 PR middle-end/57036
21591 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
21592 parameter, only add abnormal goto edges from the copied body
21593 if the call could perform abnormal gotos.
21594 (copy_cfg_body): Adjust.
21596 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
21598 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
21600 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
21602 * coretypes.h (gimple_stmt_iterator): Add struct to make
21605 2013-04-23 Richard Biener <rguenther@suse.de>
21607 PR tree-optimization/57026
21608 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
21609 from SSA names occuring in abnormal PHI nodes.
21611 2013-04-22 Andi Kleen <ak@linux.intel.com>
21613 * lto/lto.c (print_lto_report_1): Fix LTO report names.
21615 2013-04-22 Andi Kleen <ak@linux.intel.com>
21617 * lto/lto.c (print_lto_report_1): Declare early.
21618 (read_cgraph_and_symbols): Call print_lto_report_1 early.
21620 2013-04-22 Andi Kleen <ak@linux.intel.com>
21622 * common.opt (-flto-report-wpa): Add.
21623 * doc/invoke.texi (-flto-report-wpa): Add.
21624 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
21627 2013-04-22 Xinliang David Li <davidxl@google.com>
21629 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
21630 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
21631 * Makefile.in: New dependency
21633 David Daney <ddaney.cavm@gmail.com>
21635 * configure.ac (gcc_cv_as_micromips_support): Use the
21636 --fatal-warnings option.
21637 * configure: Regenerate.
21639 2013-04-22 Marek Polacek <polacek@redhat.com>
21642 * tsan.c (instrument_expr): Don't instrument expression
21643 in case its size is zero.
21645 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
21649 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
21651 * config/alpha/alpha.c (TARGET_LRA_P): New define.
21653 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
21655 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
21656 (gimple_stmt_iterator): New typedef.
21657 * gimple.h (gimple_stmt_iterator): Rename to...
21658 (gimple_stmt_iterator_d): ... This.
21659 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
21660 trees be valid for GIMPLE and GENERIC.
21661 (TARGET_GIMPLE_FOLD_BUILTIN): New.
21662 * gimple-fold.c (gimple_fold_call): Call target hook
21663 gimple_fold_builtin.
21664 * hooks.c (hook_bool_gsiptr_false): New.
21665 * hooks.h (hook_bool_gsiptr_false): New.
21666 * target.def (fold_stmt): New.
21667 * doc/tm.texi: Regenerate.
21669 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
21672 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
21673 a set sp if no stack realignment.
21675 2013-04-22 Nick Clifton <nickc@redhat.com>
21677 * config.gcc (tilegx-linux): Extend extra_objs rather than
21679 (tilepro-linux): Likewise.
21681 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
21683 * config/aarch64/aarch64-builtins.c
21685 (CF0, CF1, CF2, CF3, CF4, CF10): New.
21686 (VAR<1-12>): Add MAP parameter.
21687 (BUILTIN_*): Likewise.
21688 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
21689 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
21690 (aarch64_ushl_n<mode>): Likewise.
21691 (aarch64_sshr_n<mode>): Likewise.
21692 (aarch64_ushr_n<mode>): Likewise.
21693 (aarch64_<maxmin><mode>): Likewise.
21694 (aarch64_sqrt<mode>): Likewise.
21695 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
21696 (vshr<q>_n_*): Likewise.
21698 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
21700 * config/aarch64/aarch64-builtins.c
21701 (aarch64_simd_builtin_type_mode): Handle SF types.
21703 (BUILTIN_GPF): Define.
21704 (aarch64_init_simd_builtins): Handle SF types.
21705 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
21706 (frecps): Likewise.
21707 (frecpx): Likewise.
21708 * config/aarch64/aarch64-simd.md
21709 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
21710 (aarch64_frecpe<mode>): New.
21711 (aarch64_frecps<mode>): Likewise.
21712 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
21713 (v8type): Add frecp<esx>.
21714 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
21715 (aarch64_frecps<mode>): Likewise.
21716 * config/aarch64/iterators.md (FRECP): New.
21717 (frecp_suffix): Likewise.
21718 * config/aarch64/arm_neon.h
21719 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
21721 2013-04-22 Christian Bruel <christian.bruel@st.com>
21724 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
21725 (REG_CLASS_NAMES): Idem.
21726 (REG_CLASS_CONTENTS): Idem.
21727 (REGCLASS_HAS_FP_REG): Idem.
21728 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
21729 (sh_conditional_register_usage): Idem.
21731 2013-04-21 Jeff Law <law@redhat.com>
21733 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
21734 (ssa_forward_propagate_and_combine): Use it.
21736 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
21738 * lra.c: Update the flow chart diagram.
21740 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
21742 PR rtl-optimization/56847
21743 * lra-constraints.c (process_alt_operands): Discourage alternative
21744 with non-matche doffsettable memory constraint fro memory with
21747 2013-04-19 Richard Biener <rguenther@suse.de>
21749 PR tree-optimization/56982
21750 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
21752 * gimplify.c (gimplify_call_expr): Notice special calls.
21753 (gimplify_modify_expr): Likewise.
21754 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
21755 abnormal control flow receivers.
21756 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
21757 in the same way as cfun->has_nonlocal_labels.
21758 (gimple_purge_dead_abnormal_call_edges): Likewise.
21759 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
21760 receivers start a basic-block.
21762 2013-04-19 Richard Biener <rguenther@suse.de>
21764 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
21766 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
21767 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
21768 (SLP_TREE_LOAD_PERMUTATION): Add.
21769 (vect_transform_slp_perm_load): Adjust prototype.
21770 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
21771 (vect_free_slp_instance): Likewise.
21772 (vect_create_new_slp_node): Likewise.
21773 (vect_supported_slp_permutation_p): Remove.
21774 (vect_slp_rearrange_stmts): Adjust.
21775 (vect_supported_load_permutation_p): Likewise. Inline
21776 vect_supported_slp_permutation_p here.
21777 (vect_analyze_slp_instance): Compute load permutations per
21778 slp node instead of per instance.
21779 (vect_get_slp_defs): Adjust.
21780 (vect_transform_slp_perm_load): Likewise.
21781 (vect_schedule_slp_instance): Remove redundant code.
21782 (vect_schedule_slp): Remove hack for PR56270, add it ...
21783 * tree-vect-stmts.c (vectorizable_load): ... here, do not
21784 CSE loads for SLP. Adjust.
21786 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
21788 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
21789 spelling in two comments.
21791 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
21794 * config/arm/arm.c (load_multiple_sequence): Require SP
21795 as base register for loads if SP is in the register list.
21797 2013-04-19 Martin Jambor <mjambor@suse.cz>
21799 PR tree-optimization/56718
21800 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
21801 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
21802 and made public. Adjusted all callers.
21803 (ipa_intraprocedural_devirtualization): New function.
21804 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
21805 (ipa_intraprocedural_devirtualization): Likewise.
21806 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
21808 2013-04-19 Richard Biener <rguenther@suse.de>
21810 PR tree-optimization/57000
21811 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
21813 2013-04-19 Terry Guo <terry.guo@arm.com>
21815 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
21817 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
21818 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
21819 (cortex_m4_fmacs): Use new reservations.
21820 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
21822 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
21824 PR rtl-optimization/56999
21825 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
21827 (lra_coalesce): Remove split_origin_bitmap and related code.
21828 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
21829 ranges if necessary.
21831 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
21833 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
21835 (ix86_expand_call): Remove clobbered_registers array and use
21836 x86_64_ms_sysv_extra_clobbered_registers instead.
21837 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
21839 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
21841 * config/i386/i386.md (*call_rex64_ms_sysv): Use
21842 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
21843 (*call_value_rex64_ms_sysv): Ditto.
21845 2013-04-18 Cary Coutant <ccoutant@google.com>
21847 * dwarf2out.c (output_pubnames): Check die_perennial_p of
21848 parent instead of die_mark.
21850 2013-04-18 Diego Novillo <dnovillo@google.com>
21852 * gimple.c (create_gimple_tmp): New.
21853 (get_expr_type): New.
21854 (build_assign): New.
21855 (build_type_cast): New.
21856 * gimple.h (enum ssa_mode): Define.
21857 (gimple_seq_set_location): New.
21858 * asan.c (build_check_stmt): Change some gimple_build_* calls
21859 to use build_assign and build_type_cast.
21861 2013-04-18 Richard Biener <rguenther@suse.de>
21863 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
21864 handle negative step. Remove redundant checks.
21865 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
21866 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
21867 for negative step and grouped loads fail to vectorize.
21869 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
21871 * emit-rtl.c (reset_insn_used_flags): New function.
21872 (reset_all_used_flags): Use it.
21873 (verify_insn_sharing): New function.
21874 (verify_rtl_sharing): Fix verification for SEQUENCEs.
21876 2013-04-18 Jakub Jelinek <jakub@redhat.com>
21878 PR tree-optimization/56984
21879 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
21880 and (x >> M) >= N don't register any assertion if N << M is the
21883 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
21885 * lower-subreg.c (resolve_simple_move): If called self-recursive,
21886 do not delete_insn insns that have not yet been emitted, only
21887 unlink them with remove_insn.
21888 * df-scan.c (df_insn_delete): Revert r197492.
21890 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
21892 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
21893 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
21895 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
21897 * config/arm/arm.md (movsicc_insn): Convert define_insn into
21898 define_insn_and_split.
21899 (and_scc,ior_scc,negscc): Likewise.
21900 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
21902 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
21904 * config/arm/arm.c (use_return_insn): Return 0 for targets that
21905 can benefit from using a sequence of LDRD instructions in epilogue
21906 instead of a single LDM instruction.
21908 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
21911 * doc/extend.texi: Fix typo.
21913 2013-04-17 Richard Biener <rguenther@suse.de>
21915 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
21916 (vect_build_slp_tree): ... here.
21917 (vect_build_slp_tree_1): Compute which stmts of the SLP group
21918 match. Remove special-casing of mismatched complex loads.
21919 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
21920 re-try the match with swapped commutative operands.
21921 (vect_supported_load_permutation_p): Remove special-casing of
21922 mismatched complex loads.
21923 (vect_analyze_slp_instance): Adjust.
21925 2013-04-17 Richard Biener <rguenther@suse.de>
21927 PR rtl-optimization/56921
21928 * cfgloop.h (struct loop): Add simple_loop_desc member.
21929 (struct niter_desc): Mark with GTY(()).
21930 (simple_loop_desc): Do not use aux field but simple_loop_desc.
21931 * loop-iv.c (get_simple_loop_desc): Likewise.
21932 (free_simple_loop_desc): Likewise.
21935 2013-04-16 Richard Biener <rguenther@suse.de>
21937 PR rtl-optimization/56921
21938 * loop-init.c (pass_rtl_move_loop_invariants): Add
21939 TODO_do_not_ggc_collect to todo_flags_finish.
21940 (pass_rtl_unswitch): Same.
21941 (pass_rtl_unroll_and_peel_loops): Same.
21942 (pass_rtl_doloop): Same.
21944 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
21946 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
21947 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
21948 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
21950 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
21951 * tree-streamer.c (record_common_node): Adjust reference in comment.
21953 2013-04-17 Terry Guo <terry.guo@arm.com>
21955 * config/arm/cortex-m4.md: Add a new bypass.
21957 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21959 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
21961 (*subs_<optab><mode>_multp2): New pattern.
21962 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
21963 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
21965 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21967 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
21968 (*subs_mul_imm_<mode>): New pattern.
21970 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
21973 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
21974 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
21975 (vsx_movti_32bit): Change j->wa to O->wa.
21977 2013-04-16 Richard Biener <rguenther@suse.de>
21979 PR rtl-optimization/56921
21980 * loop-init.c (pass_rtl_move_loop_invariants): Add
21981 TODO_do_not_ggc_collect to todo_flags_finish.
21982 (pass_rtl_unswitch): Same.
21983 (pass_rtl_unroll_and_peel_loops): Same.
21984 (pass_rtl_doloop): Same.
21986 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
21988 * config/arm/arm.c (emit_multi_reg_push): New declaration
21989 for an existing function.
21990 (arm_emit_strd_push): New function.
21991 (arm_expand_prologue): Used here.
21992 (arm_emit_ldrd_pop): New function.
21993 (arm_expand_epilogue): Used here.
21994 (arm_get_frame_offsets): Update condition.
21995 (arm_emit_multi_reg_pop): Add a special case for load of a single
21996 register with writeback.
21998 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22000 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22003 2013-04-16 Richard Biener <rguenther@suse.de>
22005 PR tree-optimization/56756
22006 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22007 (first_mem_ref_loc): New.
22008 (execute_sm): Place the load temporarily before a previous
22009 access instead of in the latch edge to ensure its SSA dependencies
22010 are defined at points dominating the load.
22012 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22014 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22015 correct fix by moving header and footer insn to the footer of
22016 the merged basic block. Clear BB_END of the merged-away block.
22018 PR middle-end/43631
22019 * emit-rtl.c (make_note_raw): New function.
22020 (link_insn_into_chain): New static inline function.
22021 (add_insn): Use it.
22022 (add_insn_before, add_insn_after): Factor insn chain linking code...
22023 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22024 using link_insn_into_chain.
22025 (note_outside_basic_block_p): New helper function for emit_note_after
22026 and emit_note_before.
22027 (emit_note_after): Use nobb variant of add_insn_after if the note
22028 should not be contained in a basic block.
22029 (emit_note_before): Use nobb variant of add_insn_before if the note
22030 should not be contained in a basic block.
22031 (emit_note_copy): Use make_note_raw.
22032 (emit_note): Likewise.
22033 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22034 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22035 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22036 the moved barrier the tail of the basic block it follows.
22037 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22039 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22041 PR tree-optimization/56962
22042 * gimple-ssa-strength-reduction.c (record_increment): Only set
22043 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22044 either rhs1 or rhs2 is equal to c->base_expr.
22046 2013-04-15 Richard Biener <rguenther@suse.de>
22048 PR tree-optimization/56933
22049 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22051 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22052 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22053 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22054 dependence check ...
22055 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22057 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22058 GROUP_READ_WRITE_DEPENDENCE.
22060 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22062 * emit-rtl.c (reset_all_used_flags): New function.
22063 (verify_rtl_sharing): Call reset_all_used_flags before and after
22064 performing the checks.
22066 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22068 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22069 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22070 * config/arm/constraints.md (De): New constraint.
22071 * config/arm/neon.md (anddi3_neon): Delete.
22072 (neon_vand<mode>): Expand to standard anddi3 pattern.
22073 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22074 Move earlier in the file.
22075 (neon_inv_logic_op2): Likewise.
22076 (arm_anddi_operand_neon): New predicate.
22078 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22080 * configure.ac (gcc_cv_ld_as_needed): Set
22081 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22082 Use -z ignore, -z record on *-*-solaris2*.
22083 (HAVE_LD_AS_NEEDED): Update comment.
22084 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22085 * configure: Regenerate.
22086 * config.in: Regenerate.
22087 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22088 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22089 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22090 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22091 equivalents. Fix markup.
22092 * doc/tm.texi: Regenerate.
22094 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22096 * config/i386/i386.opt: New option mstack-protector-guard=.
22097 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22098 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22099 TARGET_SSP_TLS_GUARD.
22100 * config/i386/i386.c (ix86_option_override_internal): Set
22101 ix86_stack_protector_guard.
22102 * config/i386/i386.md (stack_protect_set): Enable for
22103 TARGET_SSP_TLS_GUARD only.
22104 (stack_protect_set_<mode>): Ditto.
22105 (stack_protect_test): Ditto.
22106 (stack_protect_test_<mode>): Ditto.
22107 * doc/invoke.texi (i386 Option): Document.
22109 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22112 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22113 (S_MODES): Set H_MODE bit.
22114 (SF_MODES): Set only S_MODE and SF_MODE bits.
22115 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22116 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22117 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22118 <MODE_FLOAT>: Likewise.
22120 2013-04-15 Joey Ye <joey.ye@arm.com>
22122 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22124 2013-04-15 Joey Ye <joey.ye@arm.com>
22126 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22128 (thumb_far_jump_used_p): Count instruction size and set
22131 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22133 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22134 * resource.c (mark_target_live_regs): Reformat conditional block.
22136 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22138 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22139 notes, they are emitted only just before final.
22140 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22142 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22144 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22145 * cfgrtl.c (delete_insn): Call it here instead.
22146 * lra-spills.c (lra_final_code_change): Use delete_insn.
22147 * haifa-sched.c (sched_remove_insn): Likewise.
22148 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22149 returning to the nop pool.
22150 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22151 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22153 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22155 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22156 * doc/tm.texi: Regenerated.
22158 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22160 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22163 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22165 * df-core.c (df_find_def): Compare register numbers.
22166 (df_find_use): Likewise.
22168 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22171 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22172 lra_in_progress for return.
22174 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22176 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22177 define_insn into define_insn_and_split and emit movsicc patterns.
22179 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22181 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
22183 2013-04-12 Richard Biener <rguenther@suse.de>
22185 * tree-pass.h (TODO_do_not_ggc_collect): New.
22186 * passes.c (execute_one_ipa_transform_pass): Honor
22187 TODO_do_not_ggc_collect.
22188 (execute_one_pass): Likewise.
22191 2013-04-10 Richard Biener <rguenther@suse.de>
22193 * passes.c (init_optimization_passes): Remove reload pass.
22194 * ira.c (do_reload): Merge into ...
22196 (rest_of_handle_reload): Remove.
22197 (pass_reload): Likewise.
22198 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22199 of reload for vzeroupper pass placement.
22201 2013-04-12 Jakub Jelinek <jakub@redhat.com>
22203 PR tree-optimization/56918
22204 PR tree-optimization/56920
22205 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
22206 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
22207 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
22208 use wide_mul_with_sign method.
22210 2013-04-12 Richard Biener <rguenther@suse.de>
22212 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
22213 not be considered a gimple constant.
22215 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
22217 * fold-const.c (const_binop): Handle vector shifts by a scalar.
22218 (fold_binary_loc): Call const_binop also for mixed vector-scalar
22221 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
22222 Jakub Jelinek <jakub@redhat.com>
22224 * opts.c: Include diagnostic-color.h.
22225 (common_handle_option): Handle OPT_fdiagnostics_color_.
22226 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
22227 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
22228 (diagnostic-color.o): New.
22229 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
22230 (diagnostic_color_rule): New enum.
22231 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
22232 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
22233 the location string.
22234 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
22235 either NULL, or color kind.
22236 * diagnostic-color.c: New file.
22237 * diagnostic-color.h: New file.
22238 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
22240 * doc/invoke.texi (-fdiagnostics-color): Document.
22241 * pretty-print.h (pp_show_color): Define.
22242 (struct pretty_print_info): Add show_color field.
22243 * diagnostic.c: Include diagnostic-color.h.
22244 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
22245 macros. Colorize error:, warning: etc. strings and also the location
22247 (diagnostic_show_locus): Colorize the caret line.
22248 * pretty-print.c: Include diagnostic-color.h.
22249 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
22250 inside of %< %> quotes or quoted through q format modifier.
22252 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22254 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
22256 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22258 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
22259 code in CC_NZ mode.
22260 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
22263 2013-04-11 Marek Polacek <polacek@redhat.com>
22265 PR tree-optimization/48184
22266 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
22268 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
22270 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
22271 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
22272 (skip_simple_arithmetic): Tidy up.
22273 * tree.h (skip_simple_constant_arithmetic): Declare.
22275 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22277 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
22279 2013-04-11 Richard Biener <rguenther@suse.de>
22281 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22282 generate vector constants.
22284 2013-04-11 Richard Biener <rguenther@suse.de>
22286 PR tree-optimization/56878
22287 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
22288 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
22290 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
22291 Prefer to align the DR with the most invariant base address.
22293 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22295 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
22298 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
22300 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
22301 floating-point vector comparisons against 0.
22303 2013-04-11 Jakub Jelinek <jakub@redhat.com>
22305 PR tree-optimization/56899
22306 * fold-const.c (extract_muldiv_1): Apply distributive law
22307 only if TYPE_OVERFLOW_WRAPS (ctype).
22309 2013-04-11 Bin Cheng <bin.cheng@arm.com>
22312 * ira-costs.c (scan_one_insn): Check whether the source rtx of
22313 loading has side effect.
22315 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
22317 * config/sparc/sparc.c: Include tree-pass.h.
22318 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
22319 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
22320 head of file. Change return type. Split off gate function.
22321 (sparc_gate_work_around_errata): New function.
22322 (pass_work_around_errata): New pass definition.
22323 (insert_pass_work_around_errata) New pass insert definition to
22324 insert pass_work_around_errata just after delayed-branch scheduling.
22325 (sparc_option_override): Insert the pass.
22326 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
22328 2013-04-10 David S. Miller <davem@davemloft.net>
22330 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
22331 or -mcpu=hypersparc.
22333 * target.def (cstore_mode): New hook.
22334 * target.h: Include insn-codes.h
22335 * targhooks.c: Likewise.
22336 (default_cstore_mode): New function.
22337 * targhooks.h: Declare it.
22338 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
22339 * doc/tm.texi: Rebuild.
22340 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
22341 target hook, rather than inspecting the insn_data.
22342 * config/sparc/sparc.c (sparc_cstore_mode): New function.
22343 (TARGET_CSTORE_MODE): Redefine.
22344 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
22346 * config/sparc/predicates.md (cstore_result_operand): New special
22348 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
22349 Use it for operand 0.
22350 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
22351 (*snesi_special): Likewise.
22352 (*snesi_zero): Likewise.
22353 (*seqsi_zero): Likewise.
22354 (*sltu_insn): Likewise.
22355 (*sgeu_insn): Likewise.
22356 (*seqdi_special): Make operand 0 and comparison operation be of
22358 (*snedi_special): Likewise.
22359 (*snedi_special_vis3): Likewise.
22360 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
22361 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
22362 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
22363 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
22364 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
22365 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
22366 (*sltu_extend_sp64): Likewise.
22367 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
22368 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
22369 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
22370 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
22371 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
22373 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
22375 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
22376 (aarch64_start_file): Use the new function.
22378 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22379 Jason Merrill <jason@redhat.com>
22381 * common.opt: Add -gdwarf.
22382 * opts.c (common_handle_option): Handle it.
22383 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
22385 2013-04-10 Richard Biener <rguenther@suse.de>
22387 * passes.c (execute_todo): Do not call ggc_collect conditional here.
22388 (execute_one_ipa_transform_pass): But unconditionally here.
22389 (execute_one_pass): And here.
22390 (init_optimization_passes): Remove reload pass.
22391 * tree-pass.h (TODO_ggc_collect): Remove.
22392 (pass_reload): Likewise.
22393 * ira.c (do_reload): Merge into ...
22395 (rest_of_handle_reload): Remove.
22396 (pass_reload): Likewise.
22397 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22398 of reload for vzeroupper pass placement.
22399 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
22400 and todo_flags_finish of all passes.
22402 2013-04-10 Richard Biener <rguenther@suse.de>
22404 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
22405 first_const_oprnd field, rename first_def_type to first_op_type.
22406 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
22407 (vect_get_and_check_slp_defs): Always use the type of the
22408 operand. Allow mixed vect_external_def, vect_constant_def types.
22409 (vect_get_constant_vectors): Handle mixed vect_external_def,
22410 vect_constant_def types.
22412 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
22414 PR tree-optimization/55524
22415 * tree-ssa-math-opts.c
22416 (convert_mult_to_fma): Don't use an fms construct
22417 when we don't have an fms operation, but fnma, and it looks
22418 likely that we'll be able to use the latter.
22420 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
22422 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
22424 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
22425 inline fail caused by overwritable functions.
22427 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
22429 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
22430 unnecessary bits in the constant power of two case.
22432 2013-04-10 Richard Biener <rguenther@suse.de>
22434 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
22435 broken code swapping operands.
22436 (vect_build_slp_tree): Do not compute load permutations here.
22437 (vect_analyze_slp_instance): Compute load permutations here,
22438 after building the SLP tree.
22440 2013-04-09 Christian Bruel <christian.bruel@st.com>
22442 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
22443 of next/prev_real_insn.
22445 2013-04-09 Jan Hubicka <jh@suse.cz>
22447 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
22448 Drop aliased parameter.
22449 (function_and_variable_visibility): Do not handle alias pairs.
22450 * cgraph.c (varpool_externally_visible_p): Update prototype.
22451 * varpool.c (varpool_add_new_variable): Update.
22453 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22455 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
22457 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
22459 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
22461 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
22463 2013-04-09 Marek Polacek <polacek@redhat.com>
22465 PR tree-optimization/48762
22466 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
22468 2013-04-09 Richard Biener <rguenther@suse.de>
22470 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
22472 (vect_build_slp_tree): Likewise.
22473 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
22474 calculating the cost of a SLP instance.
22475 (vect_analyze_slp_instance): Use it from here, after building
22478 2013-04-09 Jakub Jelinek <jakub@redhat.com>
22480 PR middle-end/56883
22481 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
22482 expand_omp_for_static_chunk): Use simple_p = true in
22483 force_gimple_operand_gsi calls when assigning to addressable decls.
22485 2013-04-09 Jeff Law <law@redhat.com>
22487 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
22488 when the boolean was created by converting a wider object which
22489 had a boolean range.
22491 2013-04-09 Richard Biener <rguenther@suse.de>
22493 * tree-vectorizer.h (slp_void_p): Remove.
22494 (slp_tree): Typedef before _slp_tree declaration.
22495 (struct _slp_tree): Use a vector of slp_tree as children.
22496 (vect_get_place_in_interleaving_chain): Remove.
22497 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
22499 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
22501 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
22502 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
22503 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
22504 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
22505 Use slp_node instead of slp_void_p and adjust.
22507 2013-04-09 Richard Biener <rguenther@suse.de>
22509 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
22510 work that is not necessary.
22512 2013-04-09 Jakub Jelinek <jakub@redhat.com>
22514 PR tree-optimization/56854
22515 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
22516 forward into clobber stmts if it would change MEM_REF lhs into
22519 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
22521 * tree.c (type_hash_lookup, type_hash_add): Make static.
22522 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
22524 2013-04-09 Richard Biener <rguenther@suse.de>
22526 * tree.h (unsave_expr_now): Remove.
22527 * tree-inline.c (mark_local_for_remap_r): Remove.
22528 (unsave_expr_1): Likewise.
22529 (unsave_r): Likewise.
22530 (unsave_expr_now): Likewise.
22531 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
22532 (propagate_tree_value): Likewise.
22534 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
22536 * doc/rtl.texi (sequence): Rewrite documentation to match the
22537 current use of SEQUENCE rtl objects.
22538 * rtl.def (SEQUENCE): Likewise.
22540 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
22541 Update documentation.
22542 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
22543 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
22545 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
22547 2013-04-08 Teresa Johnson <tejohnson@google.com>
22549 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
22550 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
22552 (estimate_edge_size_and_time): Add comment to suggest using rounding
22554 (estimate_node_size_and_time): Ditto.
22555 (remap_edge_change_prob): Use helper rounding divide methods.
22556 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
22557 (gimple_mod_pow2_value_transform): Ditto.
22558 (gimple_mod_subtract_transform): Ditto.
22559 (gimple_ic_transform): Ditto.
22560 (gimple_stringops_transform): Ditto.
22561 * stmt.c (conditional_probability): Ditto.
22562 (emit_case_dispatch_table): Ditto.
22563 * lto-cgraph.c (merge_profile_summaries): Ditto.
22564 * tree-optimize.c (execute_fixup_cfg): Ditto.
22565 * cfgcleanup.c (try_forward_edges): Ditto.
22566 * cfgloopmanip.c (scale_loop_profile): Ditto.
22568 (duplicate_loop_to_header_edge): Ditto.
22569 (lv_adjust_loop_entry_edge): Ditto.
22570 * tree-vect-loop.c (vect_transform_loop): Ditto.
22571 * profile.c (compute_branch_probabilities): Ditto.
22572 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
22573 * lto-streamer-in.c (input_cfg): Ditto.
22574 * gimple-streamer-in.c (input_bb): Ditto.
22575 * ipa-cp.c (update_profiling_info): Ditto.
22576 (update_specialized_profile): Ditto.
22577 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
22578 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
22580 * sched-rgn.c (compute_dom_prob_ps): Ditto.
22581 (compute_trg_info): Ditto.
22582 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
22583 (purge_dead_edges): Ditto.
22584 * loop-unswitch.c (unswitch_loop): Ditto.
22585 * cgraphclones.c (cgraph_clone_edge): Ditto.
22586 (cgraph_clone_node): Ditto.
22587 * tree-inline.c (copy_bb): Ditto.
22588 (copy_edges_for_bb): Ditto.
22589 (initialize_cfun): Ditto.
22590 (copy_cfg_body): Ditto.
22591 (expand_call_inline): Ditto.
22593 2013-04-08 Kai Tietz <ktietz@redhat.com>
22595 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
22596 TARGET_CYGWIN64 by TARGET_64BIT.
22598 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
22600 * config/epiphany/epiphany.md (GPR_1): New constant.
22601 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
22602 * config/epiphany/epiphany.c (gen_compare_reg):
22603 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
22604 is already in place.
22605 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
22606 Don't require being called during rtl expansion; If y operlaps r0,
22608 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
22609 (epiphany_expand_epilogue): Likewise.
22611 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
22612 Don't use CC_FPmode for ORDERED / UNORDERED.
22613 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
22615 * config/epiphany/constraints.md (CnL): New constraint.
22616 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
22617 * config/epiphany/predicates.md (add_operand): Allow 1024.
22619 * config/epiphany/epiphany.md (logical_op): New code iterator.
22620 (op_mnc): New code attribute.
22621 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
22622 (mov_f+1, mov_f+2): New peephole2 patterns.
22624 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
22625 (cstoresi4): Also allow re-use of zero result when doing a NE
22626 comparison to a non-zero operand.
22627 Use (clobber (scratch)) for first insn if the gpr output is not needed.
22629 * config/epiphany/epiphany.md (<insn_opname>v2si3):
22630 Use gen_addsi3_i / gen_subsi3_i.
22632 2013-04-08 Jakub Jelinek <jakub@redhat.com>
22636 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
22637 contain anything but clobbers, at most one __builtin_stack_restore,
22638 optionally debug stmts and final resx, and if it has at least one
22639 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
22640 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
22641 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
22642 which isn't defaut definition, remove them.
22643 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
22644 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
22645 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
22646 with MEM_REF LHS with SSA_NAME address.
22648 2013-04-08 Jeff Law <law@redhat.com>
22650 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
22652 2013-04-08 Richard Biener <rguenther@suse.de>
22654 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
22656 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
22657 determined vector type.
22658 (vect_analyze_data_refs): Likewise.
22659 (vect_get_new_vect_var): Adjust.
22660 (vect_create_destination_var): Preserve SSA name versions.
22661 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
22662 not dump anything here.
22664 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
22666 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
22667 Add member lr_slot_known.
22668 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
22670 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
22671 Remove code that sets lr_slot_offset according to what a previous
22672 version of epiphany_emit_save_restore used to do.
22673 (epiphany_emit_save_restore): When doing an lr save or restore,
22674 set/verify lr_slot_known and lr_slot_offset.
22676 2013-04-08 Xinyu Qi <xyqi@marvell.com>
22679 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
22682 2013-04-08 Richard Biener <rguenther@suse.de>
22684 * alias.c (find_base_term): Fix thinko in previous change.
22686 2013-04-08 Jakub Jelinek <jakub@redhat.com>
22688 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
22689 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
22690 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
22691 if possible to compute val.
22692 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
22693 For QImode integers don't require anything about precision. Use
22694 const_with_all_bytes_same to find out if the constant doesn't have
22695 repeated bytes in it.
22697 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22699 * config/s390/s390.c (s390_expand_insv): Only accept insertions
22702 2013-04-08 Marek Polacek <polacek@redhat.com>
22704 PR rtl-optimization/48182
22705 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
22708 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22711 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
22712 nuses, make sure we have a label.
22714 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22717 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
22718 (rs6000_emit_swdiv_low_precision): Remove.
22719 (rs6000_emit_swdiv): Rewrite to handle between one and four
22720 iterations of Newton-Raphson generally; modify required number of
22721 iterations for some cases.
22722 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
22724 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
22726 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
22727 set-but-unused variable.
22729 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
22730 basic blocks of released function bodies garbage-collectable.
22732 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
22733 (struct rtl_opt_pass): Add TODO_df_finish.
22735 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
22737 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22739 * config/arm/constraints.md (q): New constraint.
22740 * config/arm/ldrdstrd.md: New file.
22741 * config/arm/arm.md (ldrdstrd.md) New include.
22742 (arm_movdi): Use "q" instead of "r" constraint
22743 for double-word memory access.
22744 (movdf_soft_insn): Likewise.
22745 * config/arm/vfp.md (movdi_vfp): Likewise.
22746 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
22747 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
22748 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
22749 (mem_ok_for_ldrd_strd): Likewise.
22750 (output_move_double): Update assertion.
22752 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22754 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
22756 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22758 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
22759 define_insn_and_split.
22760 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
22762 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22764 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
22765 define_insn_and_split.
22766 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
22767 (shiftsi3_compare): New pattern.
22768 (rrx): New pattern.
22769 * config/arm/unspecs.md (UNSPEC_RRX): New.
22771 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22773 * config/arm/arm.md (negdi_extendsidi): New pattern.
22774 (negdi_zero_extendsidi): Likewise.
22776 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22778 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
22779 define_insn_and_split.
22780 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
22781 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
22783 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22785 * config/arm/arm.md (arm_subdi3): Convert define_insn into
22786 define_insn_and_split.
22787 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
22788 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
22790 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22792 * config/arm/arm.md (subsi3_carryin): New pattern.
22793 (subsi3_carryin_const): Likewise.
22794 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
22795 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
22797 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22799 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
22801 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22803 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
22804 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
22806 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22808 * config/arm/arm.c (arm_expand_builtin): Change fcode
22809 type to unsigned int.
22811 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22813 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
22815 2013-04-04 Ian Lance Taylor <iant@google.com>
22817 * doc/standards.texi (Standards): The Go frontend supports the Go 1
22820 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
22822 PR middle-end/56729
22823 * df-scan.c (df_insn_delete): Disable failing assert.
22825 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22827 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
22828 New function prototype.
22829 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
22830 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
22831 (arm_builtin_vectorized_function): New function.
22833 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22835 * config/arm/arm_neon_builtins.def: New file.
22836 * config/arm/arm.c (neon_builtin_data): Move contents to
22837 arm_neon_builtins.def.
22838 (enum arm_builtins): Include neon builtin definitions.
22839 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
22840 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
22842 2013-04-04 Marek Polacek <polacek@redhat.com>
22844 PR tree-optimization/48186
22845 * predict.c (maybe_hot_frequency_p): Return false if
22846 HOT_BB_FREQUENCY_FRACTION is 0.
22847 (cgraph_maybe_hot_edge_p): Likewise.
22849 2013-04-04 Richard Biener <rguenther@suse.de>
22851 PR tree-optimization/56826
22852 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
22855 2013-04-04 Richard Biener <rguenther@suse.de>
22857 PR tree-optimization/56213
22858 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
22859 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
22861 2013-04-04 Richard Biener <rguenther@suse.de>
22863 PR tree-optimization/56837
22864 * tree-loop-distribution.c (classify_partition): For non-zero
22865 values require that the value has the same precision as its
22866 mode to be useful as memset value.
22868 2013-04-03 Nick Clifton <nickc@redhat.com>
22870 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
22871 (fmssf4): Use fmsf.s on E3V5 architectures.
22872 (fnmasf4): Use fnmaf.s on E3V5 architectures.
22873 (fnmssf4): Use fnmsf.s on E3V5 architectures.
22875 2013-04-03 Jeff Law <law@redhat.com>
22877 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
22878 (lra-eliminations.o): Likewise.
22880 2013-04-03 Teresa Johnson <tejohnson@google.com>
22882 * gcov-io.c (compute_working_sets): Moved most of body of old
22883 compute_working_sets here from profile.c.
22884 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
22885 (gcov_working_set_t): Moved typedef here from basic-block.h
22886 (compute_working_set): Declare.
22887 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
22888 (get_working_sets): Renamed from compute_working_set,
22889 replace most of body with call to new compute_working_sets.
22890 (get_exec_counts): Replace call to compute_working_sets
22891 to get_working_sets.
22892 * profile.h (get_working_sets): Renamed from compute_working_set.
22893 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
22894 to get_working_sets.
22895 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
22896 * gcov-dump.c (dump_working_sets): New function.
22898 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
22900 * hwint.c (sext_hwi, zext_hwi): New functions.
22901 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
22902 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
22903 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
22904 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
22905 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
22906 (sext_hwi, zext_hwi): New functions.
22908 2013-04-03 Jeff Law <law@redhat.com>
22910 PR tree-optimization/56799
22911 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
22912 back test for widening conversion erroneously dropped in prior change.
22914 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22917 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
22918 instead of next_real_insn.
22920 2013-04-03 Marek Polacek <polacek@redhat.com>
22923 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
22925 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22928 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
22930 (thumb1_output_casesi): Likewise.
22931 (thumb2_output_casesi): Likewise.
22933 2013-04-03 Richard Biener <rguenther@suse.de>
22935 PR tree-optimization/56817
22936 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
22938 (tree_unroll_loops_completely_1): ... new function to manually
22939 walk the loop tree, properly defering outer loops of unrolled
22940 loops to later iterations.
22942 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22944 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
22945 (vectorizable_load): Likewise.
22946 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
22947 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
22949 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22951 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
22954 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
22956 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
22958 2013-04-03 Bin Cheng <bin.cheng@arm.com>
22960 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
22962 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22964 PR tree-optimization/56790
22965 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
22968 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22970 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
22972 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
22973 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
22976 2013-04-03 Jakub Jelinek <jakub@redhat.com>
22979 * tree.h (force_folding_builtin_constant_p): New decl.
22980 * builtins.c (force_folding_builtin_constant_p): New variable.
22981 (fold_builtin_constant_p): Fold immediately also if
22982 force_folding_builtin_constant_p.
22984 2013-04-03 Richard Biener <rguenther@suse.de>
22986 PR tree-optimization/56812
22987 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
22988 DRs of the same interleaving chain are independent.
22990 2013-04-02 Jason Merrill <jason@redhat.com>
22992 * gdbinit.in (pbb): Use debug fn.
22994 2013-04-02 Lawrence Crowl <crowl@google.com>
22996 * sese.h (struct ivtype_map_elt_s): Remove unused.
22997 (extern debug_ivtype_map): Remove unused.
22998 (extern eq_ivtype_map_elts): Remove unused.
22999 * sese.c (debug_ivtype_map): Removed unused.
23000 (debug_ivtype_map_1): Removed unused.
23001 (debug_ivtype_elt): Remove unused.
23002 (eq_ivtype_map_elts): Remove unused.
23005 2013-04-02 Kai Tietz <ktietz@redhat.com>
23008 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23009 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23010 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23012 (legitimize_pe_coff_symbol): Likewise.
23013 (is_imported_p): New helper-function.
23014 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23015 for Windows x64 targets.
23016 (ix86_expand_prologue): Optimize for pe-coff targets.
23017 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23018 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23019 medium/large code-model.
23020 (legitimize_pic_address): Likewise.
23021 (legitimize_tls_address): Likewise.
23022 (ix86_expand_call): Likewise.
23023 (x86_output_mi_thunk): Likewise.
23024 (get_dllimport_decl): Add new beimport argument.
23025 (construct_plt_address): Don't assert for x64 pe-coff targets.
23026 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23028 (SYMBOL_FLAG_STUBVAR): New macro.
23029 (SYMBOL_REF_STUBVAR_P): Likewise.
23030 * config/i386/winnt.c (stub_list): New structure.
23031 (stub_head): New local variable.
23032 (i386_pe_record_stub): New function.
23033 (i386_pe_file_end): Emit refptr-stubs.
23035 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23037 PR rtl-optimization/56745
23038 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23039 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23042 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23043 and both of them are MEM_REFs, just compare first argument for
23044 equality and attempt to deal even with differing offsets.
23047 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23048 of gimple_clobber_p to be MEM_REF.
23049 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23050 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23051 after gimplification.
23052 * asan.c (get_mem_ref_of_assignment): Don't instrument
23053 gimple_clobber_p stmts.
23054 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23055 gimple_clobber_p stmt if they have MEM_REF lhs and
23056 are dead because of another gimple_clobber_p stmt.
23057 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23058 gimple_clobber_p stmts like debug stmts.
23059 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23060 that refer to unused VAR_DECLs or uninitialized values.
23061 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23062 gimple_clobber_p stmts if they refer to removed parameters.
23063 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23066 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23068 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23069 using SWI48 mode attribute.
23071 2013-04-02 Wei Mi <wmi@google.com>
23073 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23074 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23075 *<rotate_insn><mode>3_mask in i386.md.
23077 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23079 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23081 2013-04-02 Richard Biener <rguenther@suse.de>
23083 PR tree-optimization/56778
23084 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23085 Runtime alias tests are not supported for gather loads.
23086 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23087 stmts referenced from SSA operands before updating SSA form.
23089 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23090 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23092 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23093 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23094 * config/arm/cortex-a53.md: New file.
23095 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23096 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23097 * config/arm/arm.c (arm_issue_rate): Likewise.
23098 * config/arm/arm-tune.md: Regenerate
23099 * config/arm/arm-tables.opt: Regenerate.
23100 * config/arm/arm-cores.def: Add cortex-a53.
23102 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23104 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23107 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23109 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23110 scalar load/store operations using B/H registers.
23111 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23113 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23115 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23117 * config/aarch64/aarch64.c
23118 (aarch64_simd_scalar_immediate_valid_for_move): New.
23119 * config/aarch64/aarch64-protos.h
23120 (aarch64_simd_scalar_immediate_valid_for_move): New.
23121 * config/aarch64/constraints.md (Dh, Dq): New.
23122 * config/aarch64/iterators.md (hq): New.
23124 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23126 * reorg.c (get_branch_condition): Deal with conditional returns.
23127 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23129 2013-04-01 Wei Mi <wmi@google.com>
23131 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23132 Truncate operand 2 using %b asm operand modifier.
23133 (*<shift_insn><mode>3_mask): Ditto.
23134 (*<rotate_insn><mode>3_mask): Ditto.
23136 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23138 PR middle-end/56798
23139 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23141 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23143 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23145 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23147 2013-03-30 Lawrence Crowl <crowl@google.com>
23149 * dse.c (clear_alias_sets): Remove never set.
23150 (disqualified_clear_alias_sets): Remove never set.
23151 (clear_alias_mode_pool): Remove never set.
23152 (dse_step0): Remove condition that is never true.
23153 (canon_address): Remove condition that is never true.
23154 (dse_step7): Remove condition that is never true.
23155 (rest_of_handle_dse): Remove condition that is never true.
23156 (rest_of_handle_dse::did_global): Remove never read from above.
23157 (dse_step2_spill): Remove never called from above.
23158 (dse_step5_spill): Remove never called from above.
23160 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23162 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23163 JUMP_TABLE_DATA changes.
23164 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23165 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23166 (Insns) <jump_table_data>: New entry.
23167 * doc/tm.texi: Regenerate.
23169 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23171 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23172 for table jump at the end of a basic block using tablejump_p.
23173 * targhooks.c (default_invalid_within_doloop): Likewise.
23174 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23175 target hook implementation that is identical to the default hook.
23176 (rs6000_invalid_within_doloop): Remove.
23178 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23179 unused variable from tablejump_p call.
23181 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
23182 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
23183 (INSN_DELETED_P): Likewise.
23184 (emit_jump_table_data): New prototype.
23185 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
23186 after 4th as unused.
23187 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
23188 * sched-vis.c (print_insn): Likewise.
23189 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
23190 insn for compatibility with back ends that use next_active_insn to
23191 identify jump table data.
23192 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
23193 (remove_insn): Likewise.
23194 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
23196 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
23197 (emit_jump_table_data): New function.
23199 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
23200 basic block, a JUMP_TABLE_DATA never is.
23201 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
23202 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
23203 off from code handling real insns.
23204 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
23205 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
23206 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
23207 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
23208 is not a NONDEBUG_INSN_P.
23209 * ira-costs.c (scan_one_insn): Likewise.
23210 * jump.c (mark_all_labels): Likewise.
23211 (mark_jump_label_1): Likewise.
23212 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23213 * lra.c (get_insn_freq): Expect all insns reaching here to be in
23215 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
23216 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
23217 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
23218 JUMP_TABLE_DATA_P insns.
23219 (calculate_elim_costs_all_insns): Likewise.
23220 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
23221 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
23222 (delete_output_reload): Code style fixups.
23223 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
23224 insn flags on this non-insn.
23225 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
23226 as scheduling barriers, for pre-change compatibility.
23227 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
23228 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
23230 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
23231 redundant JUMP_TABLE_DATA_P test.
23232 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
23233 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23234 (frv_for_each_packet): Likewise.
23235 * config/i386/i386.c (min_insn_size): Likewise.
23236 (ix86_avoid_jump_mispredicts): Likewise.
23237 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23238 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
23239 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23240 (mips16_insn_length): Robustify.
23241 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
23242 (mips16_split_long_branches): Likewise.
23243 * config/pa/pa.c (pa_combine_instructions): Likewise.
23244 * config/rs6000/rs6000.c (get_next_active_insn): Treat
23245 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
23246 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
23247 as contributing to pool range lengths.
23248 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
23249 Remove redundant JUMP_TABLE_DATA_P test.
23250 (sh_loop_align): Likewise.
23251 (split_branches): Likewise.
23252 (sh_insn_length_adjustment): Likewise.
23253 * config/spu/spu.c (get_branch_target): Likewise.
23255 2013-03-29 Jan Hubicka <jh@suse.cz>
23257 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
23258 gcov streaming; stream hot bb threshold to ltrans.
23259 * predict.c (get_hot_bb_threshold): Break out from ....
23260 (maybe_hot_count_p): ... here.
23261 (set_hot_bb_threshold): New function.
23262 * lto-section-in.c (lto_section_name): Add profile.
23263 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
23264 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
23265 and data-streamer.h
23266 (histogram_entry): New structure.
23267 (histogram, histogram_pool): New global vars.
23268 (histogram_hash): New structure.
23269 (histogram_hash::hash): New method.
23270 (histogram_hash::equal): Likewise.
23271 (account_time_size): New function.
23272 (cmp_counts): New function.
23273 (dump_histogram): New function.
23274 (ipa_profile_generate_summary): New function.
23275 (ipa_profile_write_summary): New function.
23276 (ipa_profile_read_summary): New function.
23277 (ipa_profile): Decide on threshold.
23278 (pass_ipa_profile): Add ipa_profile_write_summary and
23279 ipa_profile_read_summary.
23280 * Makefile.in (ipa.o): Update dependencies.
23281 * lto-streamer.h (LTO_section_ipa_profile): New section.
23283 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
23285 * tree.h (VAR_P): New.
23287 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
23290 * doc/invoke.texi ([-fwhole-program]): Fix typo.
23292 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
23294 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
23295 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23296 (control_flow_insn_p): Likewise.
23297 * cfgrtl.c (duplicate_insn_chain): Likewise.
23298 * final.c (get_attr_length_1): Likewise.
23299 (shorten_branches): Likewise.
23300 (final_scan_insn): Likewise.
23301 * function.c (instantiate_virtual_regs): Likewise.
23302 * gcse.c (insert_insn_end_basic_block): Likewise.
23303 * ira-costs.c (scan_one_insn): Likewise.
23304 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23305 * lra.c (check_rtl): Likewise.
23306 * reload1.c (elimination_costs_in_insn): Likewise.
23307 * reorg.c (follow_jumps): Likewise.
23309 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
23310 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23311 (thumb_far_jump_used_p): Likewise.
23312 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
23313 (workaround_speculation): Likewise.
23314 (add_sched_insns_for_speculation): Likewise.
23315 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
23316 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23317 (frv_for_each_packet): Likewise.
23318 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
23319 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
23320 (final_emit_insn_group_barriers): Likewise.
23321 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23322 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23323 (mips16_insn_length): Likewise.
23324 * config/pa/pa.c (pa_reorg): Likewise.
23325 (pa_combine_instructions): Likewise.
23326 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
23327 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
23328 (sh_reorg): Likewise.
23329 (split_branches): Likewise.
23330 * config/spu/spu.c (get_branch_target): Likewise.
23332 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
23335 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
23337 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
23338 Fix declaration name.
23340 2013-03-28 Lawrence Crowl <crowl@google.com>
23342 * graphds.h (struct graph.indicies): Remove unused.
23343 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
23344 (SCOP_ORIGINAL_PDDRS): Remove unused.
23345 * sese.h (extern insert_loop_close_phis): Removed unused.
23346 (extern insert_guard_phis): Removed unused.
23347 (extern ivtype_map_elt_info): Removed unused.
23348 (new_ivtype_map_elt): Removed unused.
23349 * sese.c (ivtype_map_elt_info): Removed unused.
23351 2013-03-28 Lawrence Crowl <crowl@google.com>
23353 * Makefile.in: Add several missing include dependences.
23355 (test-dump.o): New. This object is not added to any executable,
23356 but is present for ad-hoc testing.
23358 (debug (const bitmap_head_def &)): New.
23359 (debug (const bitmap_head_def *)): New.
23361 (extern debug (const bitmap_head_def &)): New.
23362 (extern debug (const bitmap_head_def *)): New.
23364 (debug (edge_def &)): New.
23365 (debug (edge_def *)): New.
23367 (debug (basic_block_def &)): New.
23368 (debug (basic_block_def *)): New.
23370 (dump_node (const_tree, int, FILE *)): Correct source file.
23372 (debug (die_struct &)): New.
23373 (debug (die_struct *)): New.
23375 (extern debug (die_struct &)): New.
23376 (extern debug (die_struct *)): New.
23377 * gimple-pretty-print.c
23378 (debug (gimple_statement_d &)): New.
23379 (debug (gimple_statement_d *)): New.
23380 * gimple-pretty-print.h
23381 (extern debug (gimple_statement_d &)): New.
23382 (extern debug (gimple_statement_d *)): New.
23384 (debug (ira_allocno_copy &)): New.
23385 (debug (ira_allocno_copy *)): New.
23386 (debug (ira_allocno &)): New.
23387 (debug (ira_allocno *)): New.
23389 (extern debug (ira_allocno_copy &)): New.
23390 (extern debug (ira_allocno_copy *)): New.
23391 (extern debug (ira_allocno &)): New.
23392 (extern debug (ira_allocno *)): New.
23394 (debug (live_range &)): New.
23395 (debug (live_range *)): New.
23397 (debug (lra_live_range &)): New.
23398 (debug (lra_live_range *)): New.
23400 (debug (lra_live_range &)): New.
23401 (debug (lra_live_range *)): New.
23403 (debug (omega_pb_d &)): New.
23404 (debug (omega_pb_d *)): New.
23406 (extern debug (omega_pb_d &)): New.
23407 (extern debug (omega_pb_d *)): New.
23409 (debug (const rtx_def &)): New.
23410 (debug (const rtx_def *)): New.
23412 (debug_tree (tree): Move within file.
23413 (debug_raw (const tree_node &)): New.
23414 (debug_raw (const tree_node *)): New.
23415 (dump_tree_via_hooks (const tree_node *, int)): New.
23416 (debug (const tree_node &)): New.
23417 (debug (const tree_node *)): New.
23418 (debug_verbose (const tree_node &)): New.
23419 (debug_verbose (const tree_node *)): New.
23420 (debug_head (const tree_node &)): New.
23421 (debug_head (const tree_node *)): New.
23422 (debug_body (const tree_node &)): New.
23423 (debug_body (const tree_node *)): New.
23424 (debug_vec_tree (tree): Move and reimplement in terms of dump.
23425 (debug (vec<tree, va_gc> &)): New.
23426 (debug (vec<tree, va_gc> *)): New.
23428 (extern debug (const rtx_def &)): New.
23429 (extern debug (const rtx_def *)): New.
23431 (debug_raw (simple_bitmap_def &)): New.
23432 (debug_raw (simple_bitmap_def *)): New.
23433 (debug (simple_bitmap_def &)): New.
23434 (debug (simple_bitmap_def *)): New.
23436 (extern debug (simple_bitmap_def &)): New.
23437 (extern debug (simple_bitmap_def *)): New.
23438 (extern debug_raw (simple_bitmap_def &)): New.
23439 (extern debug_raw (simple_bitmap_def *)): New.
23441 (debug (vinsn_def &)): New.
23442 (debug (vinsn_def *)): New.
23443 (debug_verbose (vinsn_def &)): New.
23444 (debug_verbose (vinsn_def *)): New.
23445 (debug (expr_def &)): New.
23446 (debug (expr_def *)): New.
23447 (debug_verbose (expr_def &)): New.
23448 (debug_verbose (expr_def *)): New.
23449 (debug (vec<rtx> &)): New.
23450 (debug (vec<rtx> *)): New.
23452 (extern debug (vinsn_def &)): New.
23453 (extern debug (vinsn_def *)): New.
23454 (extern debug_verbose (vinsn_def &)): New.
23455 (extern debug_verbose (vinsn_def *)): New.
23456 (extern debug (expr_def &)): New.
23457 (extern debug (expr_def *)): New.
23458 (extern debug_verbose (expr_def &)): New.
23459 (extern debug_verbose (expr_def *)): New.
23460 (extern debug (vec<rtx> &)): New.
23461 (extern debug (vec<rtx> *)): New.
23463 (_list_iter_cond_expr): Make inline instead of static.
23465 (debug (sreal &)): New.
23466 (debug (sreal *)): New.
23468 (extern debug (sreal &)): New.
23469 (extern debug (sreal *)): New.
23471 (extern debug_raw (const tree_node &)): New.
23472 (extern debug_raw (const tree_node *)): New.
23473 (extern debug (const tree_node &)): New.
23474 (extern debug (const tree_node *)): New.
23475 (extern debug_verbose (const tree_node &)): New.
23476 (extern debug_verbose (const tree_node *)): New.
23477 (extern debug_head (const tree_node &)): New.
23478 (extern debug_head (const tree_node *)): New.
23479 (extern debug_body (const tree_node &)): New.
23480 (extern debug_body (const tree_node *)): New.
23481 (extern debug (vec<tree, va_gc> &)): New.
23482 (extern debug (vec<tree, va_gc> *)): New.
23484 (debug (struct loop &)): New.
23485 (debug (struct loop *)): New.
23486 (debug_verbose (struct loop &)): New.
23487 (debug_verbose (struct loop *)): New.
23488 * tree-dump.c: Add header dependence.
23490 (extern debug (struct loop &)): New.
23491 (extern debug (struct loop *)): New.
23492 (extern debug_verbose (struct loop &)): New.
23493 (extern debug_verbose (struct loop *)): New.
23495 (debug (data_reference &)): New.
23496 (debug (data_reference *)): New.
23497 (debug (vec<data_reference_p> &)): New.
23498 (debug (vec<data_reference_p> *)): New.
23499 (debug (vec<ddr_p> &)): New.
23500 (debug (vec<ddr_p> *)): New.
23502 (extern debug (data_reference &)): New.
23503 (extern debug (data_reference *)): New.
23504 (extern debug (vec<data_reference_p> &)): New.
23505 (extern debug (vec<data_reference_p> *)): New.
23506 (extern debug (vec<ddr_p> &)): New.
23507 (extern debug (vec<ddr_p> *)): New.
23509 (debug (pt_solution &)): New.
23510 (debug (pt_solution *)): New.
23512 (extern debug (pt_solution &)): New.
23513 (extern debug (pt_solution *)): New.
23515 (debug (_var_map &)): New.
23516 (debug (_var_map *)): New.
23517 (debug (tree_live_info_d &)): New.
23518 (debug (tree_live_info_d *)): New.
23520 (extern debug (_var_map &)): New.
23521 (extern debug (_var_map *)): New.
23522 (extern debug (tree_live_info_d &)): New.
23523 (extern debug (tree_live_info_d *)): New.
23525 2013-03-28 Jan Hubicka <jh@suse.cz>
23527 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
23529 2013-03-28 Ian Bolton <ian.bolton@arm.com>
23531 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
23532 record only when desired or required.
23534 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
23536 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
23537 *vec_extractv2di_1_rex64. Use x64 isa attribute.
23539 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23541 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
23542 (*andsi3_compare0_uxtw): New pattern.
23543 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
23544 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
23546 2013-03-28 Jan Hubicka <jh@suse.cz>
23548 * data-streamer-in.c (streamer_read_gcov_count): New function.
23549 * gimple-streamer-out.c: Include value-prof.h.
23550 (output_gimple_stmt): Output histogram.
23551 (output_bb): Use streamer_write_gcov_count.
23552 * value-prof.c: Include data-streamer.h
23553 (dump_histogram_value): Add HIST_TYPE_MAX.
23554 (stream_out_histogram_value): New function.
23555 (stream_in_histogram_value): New function.
23556 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
23557 (stream_out_histogram_value, stream_in_histogram_value): Declare.
23558 * data-streamer-out.c (streamer_write_gcov_count): New function.
23559 (streamer_write_gcov_count_stream): New function.
23560 * lto-cgraph.c (lto_output_edge): Update counter streaming.
23561 (lto_output_node): Likewise.
23562 (input_node, input_edge): Likewise.
23563 * lto-streamer-out.c (output_cfg): Update streaming.
23564 * lto-streamer-in.c (input_cfg): Likewise.
23565 * data-streamer.h (streamer_write_gcov_count,
23566 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
23567 * gimple-streamer-in.c: Include value-prof.h
23568 (input_gimple_stmt): Input histograms.
23569 (input_bb): Update profile streaming.
23571 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
23573 * genmodes.c (emit_max_int): New function.
23574 (emit_insn_modes_h): Added call to emit_max_function.
23575 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
23577 * machmode.def: Fixed comment.
23579 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
23581 * combine.c (try_combine): Removed useless assert.
23582 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
23584 2013-03-28 Marek Polacek <polacek@redhat.com>
23585 Richard Biener <rguenther@suse.de>
23587 PR tree-optimization/56695
23588 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
23589 build signed result of a vector comparison.
23590 * tree-cfg.c (verify_gimple_comparison): Check that a result
23591 of a vector comparison has signed type.
23593 2013-03-28 Richard Biener <rguenther@suse.de>
23595 PR tree-optimization/37021
23596 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
23597 do not restrict gaps between groups.
23598 * tree-vect-stmts.c (vectorizable_load): Properly account for
23599 a gap between groups.
23601 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
23603 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
23604 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
23607 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
23609 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
23610 * doc/extend.texi (Named Address Spaces): Ditto.
23611 (Variable Attributes): Ditto.
23613 2013-03-27 Kai Tietz <ktietz@redhat.com>
23615 * config.build: Add support for cygwin x64 target.
23616 * config.gcc: Likewise.
23617 * config.host: Likewise.
23618 * configure.ac: Likewise
23619 * configure: Regenerated.
23621 2013-03-27 Kai Tietz <ktietz@redhat.com>
23623 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
23624 * config/i386/t-cygwin-w64: New file.
23625 * config/i386/cygwin-w64.h: New file.
23626 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
23627 and add support for x64-cygwin target.
23628 (CPP_SPEC): Likewise.
23629 (CXX_WRAP_SPEC_LIST): Undefine before define.
23630 (LIBGCJ_SONAME): Use 15 as version.
23632 2013-03-27 Richard Biener <rguenther@suse.de>
23634 PR tree-optimization/56716
23635 * tree-ssa-structalias.c (perform_var_substitution): Adjust
23636 dumping for ref nodes.
23638 2013-03-27 Martin Jambor <mjambor@suse.cz>
23640 PR tree-optimization/55334
23641 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
23642 restricted pointers to arrays.
23644 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
23646 * Makefile.in (.SUFFIXES): Add .cc.
23647 (.c.o): Apply same recipe for implicit rule .cc.o.
23649 2013-03-27 Richard Biener <rguenther@suse.de>
23651 PR tree-optimization/37021
23652 * tree-vect-data-refs.c (vect_check_strided_load): Allow
23653 REALPART/IMAGPART_EXPRs around the supported refs.
23654 * tree-ssa-structalias.c (find_func_aliases): Assume that
23655 floating-point values are not used to transfer pointers.
23657 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
23659 * target.def (TARGET_HAS_IFUNC_P): New target hook.
23660 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
23661 * doc/tm.texi: Regenerate.
23662 * targhooks.h (default_has_ifunc_p): New.
23663 * targhooks.c (default_has_ifunc_p): Ditto.
23664 * config/linux-protos.h: New file.
23665 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
23666 hook for linux which disables support of indirect functions in android.
23667 * config/linux-android.c: New file.
23668 * config/t-linux-android.c: Ditto.
23669 * config.gcc: Added new object file linux-android.o.
23670 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
23671 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
23672 * varasm.c (do_assemble_alias): Likewise.
23673 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
23674 doesn't support indirect functions.
23675 * configure: Regenerate.
23677 2013-03-27 Bin Cheng <bin.cheng@arm.com>
23680 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
23681 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
23684 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23686 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
23688 2013-03-27 Terry Guo <terry.guo@arm.com>
23690 * config/arm/arm-cores.def: Added core cortex-r7.
23691 * config/arm/arm-tune.md: Regenerated.
23692 * config/arm/arm-tables.opt: Regenerated.
23693 * doc/invoke.texi: Added entry for core cortex-r7.
23695 2013-03-27 Walter Lee <walt@tilera.com>
23697 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
23698 double-decrement of next_scratch_regno.
23700 2013-03-27 Walter Lee <walt@tilera.com>
23702 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
23704 (insn_v1mulus): Ditto.
23705 (insn_v2muls): Ditto.
23707 2013-03-27 Walter Lee <walt@tilera.com>
23709 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
23710 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
23712 2013-03-27 Walter Lee <walt@tilera.com>
23714 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
23715 (*sibcall_value): Ditto.
23717 2013-03-27 Walter Lee <walt@tilera.com>
23719 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
23720 (insn_mnz_v8qi): ... this ...
23721 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
23723 (insn_v<n>mnz): Replaced by ...
23724 (insn_v1mnz): ... this ...
23725 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
23727 (insn_mz_<mode>): Replaced by ...
23728 (insn_mz_v8qi): ... this ...
23729 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
23731 (insn_v<n>mz): Replaced by ...
23732 (insn_v1mz): ... this ...
23733 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
23736 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
23738 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
23740 2013-03-26 Roland McGrath <mcgrathr@google.com>
23742 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
23743 than fprintf with a non-constant, non-format string.
23745 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
23747 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
23748 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
23749 operand 0 predicate.
23750 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
23751 attribute. Use general_x64nomem_operand as operand 1 predicate.
23752 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
23753 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
23754 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
23755 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
23756 (mov<mode>_insv_1): Remove expander. Merge insn with
23757 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
23758 Use general_x64nomem_operand as operand 1 predicate.
23759 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
23760 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
23761 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
23762 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
23763 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
23764 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
23765 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
23766 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
23767 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
23768 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
23769 (general_x64nomem_operand): Ditto.
23771 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
23773 * config/rtems.opt: Add -pthread option.
23775 2013-03-26 Richard Biener <rguenther@suse.de>
23777 * alias.c (find_base_term): Avoid redundant and not used recursion.
23778 (base_alias_check): Get the initial base term from the caller.
23779 (true_dependence_1): Compute and pass base terms to base_alias_check.
23780 (write_dependence_p): Likewise.
23781 (may_alias_p): Likewise.
23783 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
23785 * config/aarch64/aarch64.c (aarch64_classify_address): Support
23786 PC-relative load in SI modes and above only.
23788 2013-03-26 Xinyu Qi <xyqi@marvell.com>
23790 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
23791 * config/arm/iwmmxt.md (WCGR0): Update.
23792 (WCGR1, WCGR2, WCGR3): Likewise.
23794 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
23796 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
23797 Use x64 and nox64 isa attributes.
23799 2013-03-26 Richard Biener <rguenther@suse.de>
23801 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
23802 alignment computations and rely on get_object_alignment_1
23803 for the !TYPE_P case.
23804 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
23806 2013-03-26 Walter Lee <walt@tilera.com>
23808 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
23809 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
23811 2013-03-25 Jeff Law <law@redhat.com>
23813 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
23814 check for INTEGRAL_TYPE_P that was missing due to checking in
23815 wrong version of prior patch.
23817 2013-03-25 Walter Lee <walt@tilera.com>
23819 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
23820 TILEGX_INSN_SHUFFLEBYTES1.
23821 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
23823 (tilegx_builtins): Ditto.
23824 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
23826 2013-03-25 Walter Lee <walt@tilera.com>
23828 * config/tilegx/tilegx.md (floatsisf2): New pattern.
23829 (floatunssisf2): New pattern.
23830 (floatsidf2): New pattern.
23831 (floatunssidf2): New pattern.
23833 2013-03-25 Walter Lee <walt@tilera.com>
23835 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
23836 tests for constraint J, K, N, P.
23838 2013-03-25 Walter Lee <walt@tilera.com>
23840 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
23841 Use indirect/pcrel encoding.
23842 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
23845 2013-03-25 Steve Ellcey <sellcey@mips.com>
23847 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
23848 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
23849 * config/mips/mips.c (mips_option_override): Set IMADD default.
23850 * config/mips/mips.h (PTF_AVOID_IMADD): New.
23851 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
23852 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
23853 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
23854 * doc/invoke.texi (-mimadd/-mno-imadd): New.
23856 2013-03-25 Jeff Law <law@redhat.com>
23858 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
23859 slightly to avoid creating and folding useless trees. Simplify
23860 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
23862 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
23864 * config/i386/i386.md (*zero_extendsidi2): Merge with
23865 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
23866 * config/i386/predicates.md (x86_64_zext_operand): Rename from
23867 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
23868 targets. Clarify comment.
23870 2013-03-25 Martin Jambor <mjambor@suse.cz>
23872 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
23873 pass-through jump functions differently.
23874 (ipa_read_jump_function): Likewise. Also use setter functions to set
23877 2013-03-25 Martin Jambor <mjambor@suse.cz>
23879 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
23880 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
23882 (ipa_get_indirect_edge_target): New function.
23883 (devirtualization_time_bonus): New parameter known_aggs, pass it to
23884 ipa_get_indirect_edge_target. Update all callers.
23885 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
23886 ipa_get_indirect_edge_target_1 instead of calling
23887 ipa_get_indirect_edge_target.
23888 (create_specialized_node): Pass aggvlas to
23889 ipcp_discover_new_direct_edges.
23891 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23893 * config/arm/arm.md (f_sels, f_seld): New types.
23894 (*cmov<mode>): New pattern.
23895 * config/arm/predicates.md (arm_vsel_comparison_operator): New
23898 2013-03-25 Kai Tietz <ktietz@redhat.com>
23900 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
23901 POSIX-printf for mingw-hosted builds.
23903 2013-03-25 Richard Biener <rguenther@suse.de>
23905 PR middle-end/56694
23906 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
23907 must-not-throw stmt location.
23909 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23911 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
23912 Emit load-acquire versions when acq is true.
23913 (arm_emit_store_exclusive): Add rel parameter.
23914 Emit store-release versions when rel is true.
23915 (arm_split_compare_and_swap): Use acquire-release instructions
23917 of barriers when appropriate.
23918 (arm_split_atomic_op): Likewise.
23919 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
23920 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
23921 (VUNSPEC_SLX): Likewise.
23922 (VUNSPEC_LDA): Likewise.
23923 (VUNSPEC_STL): Likewise.
23924 * config/arm/sync.md (atomic_load<mode>): New pattern.
23925 (atomic_store<mode>): Likewise.
23926 (arm_load_acquire_exclusive<mode>): Likewise.
23927 (arm_load_acquire_exclusivesi): Likewise.
23928 (arm_load_acquire_exclusivedi): Likewise.
23929 (arm_store_release_exclusive<mode>): Likewise.
23931 2013-03-25 Catherine Moore <clm@codesourcery.com>
23933 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
23934 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
23935 * config/mip/predicates.md (lwsp_swsp_operand,
23936 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
23937 sb16_operand, db4_operand, db7_operand, ib3_operand,
23938 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
23939 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
23940 andi16_operand): New predicates.
23941 * config/mips/mips.md (compression): New attribute.
23942 (enabled): New attribute.
23943 (length): Consider compression in computing length.
23944 (shift_compression): New code attribute.
23945 (*add<mode>3): New operands. Record compression.
23946 (sub<mode>3): Likewise.
23947 (one_cmpl<mode>2): Likewise.
23948 (*and<mode>3): Likewise.
23949 (*ior<mode>3): Likewise.
23950 (unnamed pattern for xor): Likewise.
23951 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
23952 (*<optab><mode>3): Likewise.
23953 (*mov<mode>_internal: Likewise.
23954 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
23955 (mips_unsigned_immediate_p): New.
23956 (umips_lwsp_swsp_address_p): New.
23957 (m16_based_address_p): New.
23958 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
23959 (mips_unsigned_immediate_p): New prototype.
23960 (lwsp_swsp_address_p): New prototype.
23961 (m16_based_address_p): New prototype.
23962 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
23963 (mips_signed_immediate_p): New function.
23964 (m16_based_address_p): New function.
23965 (lwsp_swsp_address_p): New function.
23966 (mips_print_operand_punctuation): Recognize short delay slot insns
23967 for microMIPS.add<mode>3"
23969 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23972 * config/arm/iterators.md (v_cmp_result): New mode attribute.
23973 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
23975 2013-03-25 Richard Biener <rguenther@suse.de>
23977 PR tree-optimization/56689
23978 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
23981 2013-03-25 Richard Biener <rguenther@suse.de>
23983 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
23985 (memory_references): Likewise.
23986 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
23987 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
23988 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
23989 (gather_mem_refs_in_loops): Fold into ...
23990 (analyze_memory_references): ... this. Move initialization
23991 to tree_ssa_lim_initialize.
23992 (fill_always_executed_in): Rename to ...
23993 (fill_always_executed_in_1): ... this.
23994 (fill_always_executed_in): Move contains_call computation to
23995 this new function from ...
23996 (tree_ssa_lim_initialize): ... here.
23997 (tree_ssa_lim): Call fill_always_executed_in.
23999 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24001 * postreload.c (reload_combine): Fix code detecting returns.
24003 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24005 * function.c (emit_use_return_register_into_block): On cc0 targets,
24006 do not emit the sequence between cc0 setter and user.
24008 2013-03-25 Kai Tietz <ktietz@redhat.com>
24010 * config/i386/predicates.md (local_symbolic_operand): Interpret
24011 dll-imported symbols as none-local.
24013 2013-03-25 Richard Biener <rguenther@suse.de>
24015 * tree-ssa-loop-im.c (struct depend): Remove.
24016 (struct lim_aux_data): Make depends a vec of gimples.
24017 (free_lim_aux_data): Adjust.
24018 (add_dependency): Likewise.
24019 (set_level): Likewise.
24021 2013-03-25 Richard Biener <rguenther@suse.de>
24023 PR middle-end/56434
24024 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24025 the pointer returned by calls with ECF_MALLOC set.
24027 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24029 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24031 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24033 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24034 using MMXMODE mode iterator.
24035 (*move<mode>_internal): Merge with *movv2sf_internal and
24036 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24038 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24040 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24041 (record_last_mem_set_info): Likewise.
24043 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24044 of XNEWVEC followed by memset.
24045 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24047 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24049 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24050 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24051 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24052 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24053 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24054 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24055 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24056 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24057 BARRIER_P instead of GET_CODE.
24059 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24061 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24062 inaccuracy in the probing code.
24064 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24065 (ctrapdi4): Likewise.
24067 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24069 * calls.c (expand_call): Add missing guard to code handling return
24070 of non-BLKmode structures in MSB.
24071 * function.c (expand_function_end): Likewise.
24073 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24075 * combine.c (try_combine): Adjust comment. Do not add the set of
24076 insn #0 if the destination indirectly is set or dies in insn #2.
24077 Tidy up code to distribute a new note.
24079 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24081 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24082 also for alternatives 16 and 17.
24084 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24086 * config/i386/sse.md (*mov<mode>_internal): Merge with
24087 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24088 Emit insn template depending on type attribute. Use
24089 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24090 movd instead of movq mnemonic for interunit moves. Rewrite mode
24091 attribute calculation. Remove unit attribute calculation.
24092 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24093 Set prefix_data16 attribute for DImode ssemov types.
24094 Use Ym instead of y for SSE-MMX conversion alternatives.
24095 Reorder operand constraints.
24097 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24099 * df.h (df_insn_delete): Adjust prototype.
24100 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24101 and let it decide whether mark the basic block dirty.
24102 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24103 * df-scan.c (df_insn_info_delete): New helper function, split
24104 off from df_insn_delete.
24105 (df_scan_free_bb_info): Use it.
24106 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24108 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24109 that the insn is actually an insn and it has a non-NULL basic block.
24110 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24112 2013-03-22 Richard Biener <rguenther@suse.de>
24114 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24116 (mem_ref_alloc): Do not allocate them.
24117 (refs_independent_p): Do not query or maintain a cache.
24119 2013-03-22 Richard Biener <rguenther@suse.de>
24121 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24122 (gather_mem_refs_in_loops): Do not compute it.
24123 (analyze_memory_references): Do not allocate it.
24124 (tree_ssa_lim_finalize): Do not free it.
24125 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24127 2013-03-22 Richard Biener <rguenther@suse.de>
24129 * is-a.h (as_a): Use gcc_checking_assert.
24131 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24133 * config/aarch64/aarch64.c (aarch64_print_operand): New
24134 format specifier for printing a constant in hex.
24135 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24136 format specifier for printing second operand.
24138 2013-03-22 Richard Biener <rguenther@suse.de>
24140 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24142 (gather_mem_refs_in_loops): Perform store accumulation here.
24143 (create_vop_ref_mapping_loop): Remove.
24144 (create_vop_ref_mapping): Likewise.
24145 (analyze_memory_references): Initialize refs_stored_in_loop.
24146 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24147 (record_indep_loop): Remove.
24148 (record_dep_loop): New function.
24149 (ref_indep_loop_p_1): Adjust to only walk over references
24150 in the loop, not its subloops.
24151 (ref_indep_loop_p): Rename to ...
24152 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24153 maintaining a more fine-grained cache.
24154 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24155 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24157 2013-03-22 Richard Biener <rguenther@suse.de>
24159 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24160 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24161 aggregate mem_ref_loc.
24162 (free_mem_ref_locs): Inline into ...
24163 (memref_free): ... this and adjust.
24164 (mem_ref_alloc): Adjust.
24165 (mem_ref_locs_alloc): Remove.
24166 (record_mem_ref_loc): Adjust.
24167 (get_all_locs_in_loop): Rewrite into ...
24168 (for_all_locs_in_loop): ... this iterator.
24169 (rewrite_mem_ref_loc): New functor.
24170 (rewrite_mem_refs): Use for_all_locs_in_loop.
24171 (sm_set_flag_if_changed): New functor.
24172 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24173 (ref_always_accessed): New functor.
24174 (ref_always_accessed_p): Use for_all_locs_in_loop.
24176 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24178 * tree-pass.h (PROP_gimple_lvec): New.
24179 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24180 (init_optimization_passes): Move pass_lower_vector.
24181 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
24183 (pass_lower_vector): Provide PROP_gimple_lvec.
24184 (pass_lower_vector_ssa): Likewise.
24185 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
24187 2013-03-21 Mark Wielaard <mjw@redhat.com>
24189 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
24191 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24193 * config/i386/i386.md (*movdi_internal): Disparage slightly
24194 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
24195 conversion alternatives.
24197 2013-03-21 Jakub Jelinek <jakub@redhat.com>
24199 PR middle-end/48087
24200 * diagnostic.def (DK_WERROR): New kind.
24201 * diagnostic.h (werrorcount): Define.
24202 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
24203 promoted to DK_ERROR, increment DK_WERROR counter instead of
24205 * toplev.c (toplev_main): Call print_ignored_options even if
24206 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
24207 even if just werrorcount is non-zero.
24210 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
24212 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
24213 (string_cst_pool_decl): New function.
24214 (optimize_one_addr_into_implicit_ptr): New function.
24215 (resolve_addr_in_expr): Optimize DWARF location expression
24216 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
24217 which doesn't live in memory, but has DW_AT_location or
24218 DW_AT_const_value, or refers to a string literal, into
24219 DW_OP_GNU_implicit_pointer.
24220 (optimize_location_into_implicit_ptr): New function.
24221 (resolve_addr): If removing DW_AT_location of a variable because
24222 it was DW_OP_addr of address of the variable, but the variable doesn't
24223 live in memory, try to emit const value attribute for the initializer.
24225 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24227 * tree.h (VECTOR_TYPE_P): New macro.
24228 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
24229 TYPE_MODE): Use it.
24230 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
24231 VEC_COND_EXPR cannot be lvalues.
24232 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
24234 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24236 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
24237 Restrict the transformation to equal modes.
24239 2013-03-21 Richard Biener <rguenther@suse.de>
24241 PR tree-optimization/39326
24242 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
24243 (MEM_ANALYZABLE): Adjust.
24244 (record_mem_ref_loc): Move bitmap ops ...
24245 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
24246 unanalyzable refs, do not record locations for it.
24247 (analyze_memory_references): Allocate ref zero as shared
24249 (refs_independent_p): Do not test for unanalyzed mems here.
24250 (ref_indep_loop_p_1): Special-case disambiguation against
24251 the unanalyzed ref.
24252 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
24254 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
24256 * config/arm/arm-protos.h (tune_params): Add
24257 prefer_neon_for_64bits field.
24258 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
24259 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
24260 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
24261 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
24262 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
24263 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
24264 (arm_option_override): Handle -mneon-for-64bits new option.
24265 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
24266 (prefer_neon_for_64bits): Declare new variable.
24267 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
24268 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
24269 (arch_enabled): Handle new arch types. Remove support for onlya8
24271 (one_cmpldi2): Use new arch names.
24272 (zero_extend<mode>di2, extend<mode>di2): Ditto.
24273 * config/arm/arm.opt (mneon-for-64bits): Add option.
24274 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
24275 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
24276 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
24278 * doc/invoke.texi (-mneon-for-64bits): Document.
24280 2013-03-21 Richard Biener <rguenther@suse.de>
24282 PR tree-optimization/39326
24283 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
24284 (sort_bbs_in_loop_postorder_cmp): New function.
24285 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
24287 2013-03-21 Richard Biener <rguenther@suse.de>
24289 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
24290 (vect_insert_into_interleaving_chain): Likewise.
24291 (vect_drs_dependent_in_basic_block): Inline ...
24292 (vect_slp_analyze_data_ref_dependence): ... here. New function,
24294 (vect_analyze_data_ref_dependence): ... here. Simplify.
24295 (vect_check_interleaving): Simplify.
24296 (vect_analyze_data_ref_dependences): Likewise. Split out ...
24297 (vect_slp_analyze_data_ref_dependences): ... this new function.
24298 (dr_group_sort_cmp): New function.
24299 (vect_analyze_data_ref_accesses): Compute data-reference groups
24300 here instead of in vect_analyze_data_ref_dependence. Use
24301 a more efficient algorithm.
24302 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
24303 vect_slp_analyze_data_ref_dependences. Call
24304 vect_analyze_data_ref_accesses earlier.
24305 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
24306 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
24307 (vect_slp_analyze_data_ref_dependences): New prototype.
24309 2013-03-21 Richard Biener <rguenther@suse.de>
24311 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
24312 ref is stored in the loop.
24313 (find_refs_for_sm): Walk only over all stores.
24314 (store_motion_loop): Allocate from lim_bitmap_obstack.
24315 (store_motion): Likewise.
24317 2013-03-21 Richard Biener <rguenther@suse.de>
24319 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
24320 Update virtual SSA form.
24322 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24324 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
24325 * configure: Regenerate.
24326 * config.in: Regenerate.
24327 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
24328 if !HAVE_LD_EH_FRAME_CIEV3.
24330 2013-03-21 Richard Biener <rguenther@suse.de>
24332 * tree-cfg.c (verify_expr_no_block): New function.
24333 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
24334 nor DECL_VALUE_EXPR have locations with associated blocks.
24335 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
24336 (clear_unused_block_pointer): Remove code dealing with
24337 blocks in DECL_DEBUG_EXPR locations.
24339 2013-03-21 Richard Biener <rguenther@suse.de>
24341 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
24342 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
24343 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
24344 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
24345 instead of DECL_DEBUG_EXPR_IS_FROM.
24346 * gimplify.c (gimplify_modify_expr): Likewise.
24347 * tree-cfg.c (verify_expr_location_1): Likewise.
24348 * tree-complex.c (create_one_component_var): Likewise.
24349 * tree-sra.c (create_access_replacement): Likewise.
24350 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
24351 (clear_unused_block_pointer): Likewise.
24352 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
24353 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
24354 * var-tracking.c (var_debug_decl): Likewise.
24355 (track_expr_p): Likewise.
24356 * tree-inline.c (add_local_variables): Likewise. Set
24357 DECL_HAS_DEBUG_EXPR_P after copying it.
24358 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
24359 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
24361 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24364 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
24365 * configure: Regenerate.
24366 * config.in: Regenerate.
24367 * config/i386/i386.md (*movdf_internal): Use
24368 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24369 movd instead of movq mnemonic for interunit moves.
24370 (*movdi_internal): Ditto.
24372 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24374 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
24375 (abd<mode>_3): New pattern.
24376 (aba<mode>_3): New pattern.
24377 (fabd<mode>_3): New pattern.
24379 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24381 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
24382 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
24383 occurrence of REGISTER_PREFIX as its empty string.
24385 2013-03-20 Jeff Law <law@redhat.com>
24387 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
24388 addititional equivalences for equality comparisons between an SSA_NAME
24389 and a constant where the SSA_NAME was set from a widening conversion.
24391 2013-03-20 Walter Lee <walt@tilera.com>
24393 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
24395 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24397 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
24398 depending on type attribute.
24399 (*movti_internal): Ditto.
24400 (*movtf_internal): Ditto.
24401 (*movxf_internal): Ditto.
24402 (*movdf_internal): Ditto.
24403 (*movsf_internal): Ditto.
24405 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24407 * config/i386/i386.md (*movti_internal): Set prefix attribute to
24408 maybe_vex for sselog1 and ssemov types.
24409 (*movdi_internal): Reorder operand constraints.
24410 (*movsi_internal): Ditto. Set prefix attribute to
24411 maybe_vex for sselog1 and ssemov types.
24412 (*movtf_internal): Set prefix attribute to maybe_vex
24413 for sselog1 and ssemov types.
24414 (*movdf_internal): Ditto. Set prefix_data16 attribute for
24415 DImode ssemov types. Reorder operand constraints.
24416 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
24417 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
24418 attribute for SImode ssemov types. Reorder operand constraints.
24420 2013-03-20 Martin Jambor <mjambor@suse.cz>
24422 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
24423 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
24425 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
24427 * config/rs6000/predicates.md (indexed_address, update_address_mem
24428 update_indexed_address_mem): New predicates.
24429 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
24430 attribute for load/store instructions.
24431 * config/rs6000/dfp.md (movsd_store): Likewise.
24432 (movsd_load): Likewise.
24433 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
24434 (unnamed HI->DI extend define_insn): Likewise.
24435 (unnamed SI->DI extend define_insn): Likewise.
24436 (unnamed QI->SI extend define_insn): Likewise.
24437 (unnamed QI->HI extend define_insn): Likewise.
24438 (unnamed HI->SI extend define_insn): Likewise.
24439 (unnamed HI->SI extend define_insn): Likewise.
24440 (extendsfdf2_fpr): Likewise.
24441 (movsi_internal1): Likewise.
24442 (movsi_internal1_single): Likewise.
24443 (movhi_internal): Likewise.
24444 (movqi_internal): Likewise.
24445 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
24446 attribute for load/store instructions.
24447 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
24449 (mov<mode>_softfloat): Likewise.
24450 (mov<mode>_hardfloat32): Likewise.
24451 (mov<mode>_hardfloat64): Likewise.
24452 (mov<mode>_softfloat64): Likewise.
24453 (movdi_internal32): Likewise.
24454 (movdi_internal64): Likewise.
24455 (probe_stack_<mode>): Likewise.
24457 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
24459 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
24460 floating point, and decimal floating point to reload iterator.
24462 * config/rs6000/constraints.md (wl constraint): New constraints to
24463 return FLOAT_REGS if certain options are used to reduce the number
24464 of separate patterns that exist in the file.
24465 (wx constraint): Likewise.
24466 (wz constraint): Likewise.
24468 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
24469 -mdebug=reg, print wg, wl, wx, and wz constraints.
24470 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
24471 Initialize the reload functions for 64-bit binary/decimal floating
24473 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
24474 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
24475 create the buffer on the stack to overcome not having a 32-bit
24477 (rs6000_emit_move): Likewise.
24478 (rs6000_secondary_memory_needed_rtx): Likewise.
24479 (rs6000_alloc_sdmode_stack_slot): Likewise.
24480 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
24481 via xxlxor, just like DFmode 0.0.
24483 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
24484 define as 1 if we are running on a power7 or newer.
24485 (enum r6000_reg_class_enum): Add new constraints.
24487 * config/rs6000/dfp.md (movsd): Delete, combine with binary
24488 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
24489 with other moves by using conditional constraits (wg). Use LFIWZX
24490 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
24491 (movsd splitter): Likewise.
24492 (movsd_hardfloat): Likewise.
24493 (movsd_softfloat): Likewise.
24495 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
24496 binary and decimal floating point moves.
24497 (fmove_ok): New attributes to combine binary and decimal floating
24498 point moves, and to combine power6x (mfpgpr) moves along normal
24500 (real_value_to_target): Likewise.
24501 (f32_lr): Likewise.
24502 (f32_lm): Likewise.
24503 (f32_li): Likewise.
24504 (f32_sr): Likewise.
24505 (f32_sm): Likewise.
24506 (f32_si): Likewise.
24507 (movsf): Combine binary and decimal floating point moves. Combine
24508 power6x (mfpgpr) moves with other moves by using conditional
24509 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
24510 (mov<mode> for SFmode/SDmode); Likewise.
24511 (SFmode/SDmode splitters): Likewise.
24512 (movsf_hardfloat): Likewise.
24513 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
24514 (movsf_softfloat): Likewise.
24515 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
24517 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
24518 wx and wz constraints.
24520 * config/rs6000/constraints.md (wg constraint): New constraint to
24521 return FLOAT_REGS if -mmfpgpr (power6x) was used.
24523 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
24526 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
24527 -mdebug=reg, print wg, wl, wx, and wz constraints.
24528 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
24529 Initialize the reload functions for 64-bit binary/decimal floating
24531 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
24532 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
24533 create the buffer on the stack to overcome not having a 32-bit
24535 (rs6000_emit_move): Likewise.
24536 (rs6000_secondary_memory_needed_rtx): Likewise.
24537 (rs6000_alloc_sdmode_stack_slot): Likewise.
24538 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
24539 via xxlxor, just like DFmode 0.0.
24541 * config/rs6000/dfp.md (movdd): Delete, combine with binary
24542 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
24543 with other moves by using conditional constraits (wg). Use LFIWZX
24544 and STFIWX for loading SDmode on power7.
24545 (movdd splitters): Likewise.
24546 (movdd_hardfloat32): Likewise.
24547 (movdd_softfloat32): Likewise.
24548 (movdd_hardfloat64_mfpgpr): Likewise.
24549 (movdd_hardfloat64): Likewise.
24550 (movdd_softfloat64): Likewise.
24552 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
24553 64-bit binary and decimal floating point moves.
24554 (FMOVE64X): Likewise.
24555 (movdf): Combine 64-bit binary and decimal floating point moves.
24556 Combine power6x (mfpgpr) moves with other moves by using
24557 conditional constraits (wg).
24558 (mov<mode> for DFmode/DDmode): Likewise.
24559 (DFmode/DDmode splitters): Likewise.
24560 (movdf_hardfloat32): Likewise.
24561 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
24562 (movdf_softfloat32): Likewise.
24563 (movdf_hardfloat64_mfpgpr): Likewise.
24564 (movdf_hardfloat64): Likewise.
24565 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
24566 (movdf_softfloat64): Likewise.
24567 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
24568 (reload_<mode>_load): Move to later in the file so they aren't in
24569 the middle of the floating point move insns.
24570 (reload_<mode>_store): Likewise.
24572 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
24575 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
24576 constraint if -mdebug=reg.
24577 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
24578 Enable using dd reload support if needed.
24580 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
24581 binary and decimal floating point moves in rs6000.md.
24582 (movtd_internal): Likewise.
24584 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
24585 decimal floating point moves.
24587 (movtf_internal): Likewise.
24588 (mov<mode>_internal, TDmode/TFmode): Likewise.
24589 (movtf_softfloat): Likewise.
24590 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
24592 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
24593 movdi_internal64, using wg constraint for move direct operations.
24594 (movdi_internal64): Likewise.
24596 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
24597 MODES_TIEABLE_P for selected modes. Print the numerical value of
24598 the various virtual registers. Use GPR/FPR first/last values,
24599 instead of hard coding the register numbers. Print which modes
24600 have reload functions registered.
24601 (rs6000_option_override_internal): If -mdebug=reg, trace the options
24602 settings before/after setting cpu, target and subtarget settings.
24603 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
24604 and for secondary reload failures in rs6000_secondary_reload_inner.
24605 (rs6000_secondary_reload_fail): Likewise.
24606 (rs6000_secondary_reload_inner): Likewise.
24608 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
24609 macros for first/last GPR and FPR registers.
24610 (LAST_GPR_REGNO): Likewise.
24611 (FIRST_FPR_REGNO): Likewise.
24612 (LAST_FPR_REGNO): Likewise.
24614 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
24615 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
24616 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
24617 (vcond<mode><mode>): Likewise.
24618 (vcondu<mode><mode>): Likewise.
24619 (vector_gtu<mode>): Likewise.
24620 (vector_gte<mode>): Likewise.
24621 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
24622 to prevent the compiler from converting DImode operations to TImode.
24623 (ior<mode>3): Likewise.
24624 (and<mode>3): Likewise.
24625 (one_cmpl<mode>2): Likewise.
24626 (nor<mode>3): Likewise.
24627 (andc<mode>3): Likewise.
24629 * config/rs6000/constraints.md (wt constraint): New constraint
24630 that returns VSX_REGS if TImode is allowed in VSX registers.
24632 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
24633 constant under VSX.
24635 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
24636 similar to TImode, but it is restricted to being in the GPRs.
24638 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
24639 TImode to occupy a single VSX register.
24641 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
24642 -mvsx-timode for power7/power8.
24643 (power7 cpu): Likewise.
24644 (power8 cpu): Likewise.
24646 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
24647 sure that TFmode/TDmode take up two registers if they are ever
24648 allowed in the upper VSX registers.
24649 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
24651 (rs6000_init_hard_regno_mode_ok): Likewise.
24652 (rs6000_debug_reg_global): Add debugging for PTImode and wt
24653 constraint. Print if LRA is turned on.
24654 (rs6000_option_override_internal): Give an error if -mvsx-timode
24655 and VSX is not enabled.
24656 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
24657 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
24658 to reg+offset addressing. Use PTImode when checking offset
24659 addresses for validity.
24660 (reg_offset_addressing_ok_p): Likewise.
24661 (rs6000_legitimate_offset_address_p): Likewise.
24662 (rs6000_legitimize_address): Likewise.
24663 (rs6000_legitimize_reload_address): Likewise.
24664 (rs6000_legitimate_address_p): Likewise.
24665 (rs6000_eliminate_indexed_memrefs): Likewise.
24666 (rs6000_emit_move): Likewise.
24667 (rs6000_secondary_reload): Likewise.
24668 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
24669 reloads to fpr registers to continue to use reg+offset addressing,
24670 but 64-bit reloads to altivec registers need reg+reg addressing.
24671 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
24672 it. Treat LO_SUM like a PLUS operation.
24673 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
24674 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
24675 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
24676 registers to share a register with a smaller sized type, since VSX
24677 puts scalars in the upper 64-bits.
24678 (print_operand): Add support for PTImode.
24679 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
24680 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
24681 registers, but don't have arithmetic support.
24682 (rs6000_memory_move_cost): Add test for VSX.
24683 (rs6000_opt_masks): Add -mvsx-timode.
24685 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
24688 (VSr): Use wt constraint for TImode.
24689 (VSv): Drop TImode support.
24690 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
24691 (vsx_movti_64bit): Likewise.
24692 (vsx_movti_32bit): Likewise.
24693 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
24694 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
24695 one '?' on the appropriate output constraint. Do not allow TImode
24696 logical operations on 32-bit systems.
24697 (vsx_ior<mode>3): Likewise.
24698 (vsx_xor<mode>3): Likewise.
24699 (vsx_one_cmpl<mode>2): Likewise.
24700 (vsx_nor<mode>3): Likewise.
24701 (vsx_andc<mode>3): Likewise.
24702 (vsx_concat_<mode>): Likewise.
24703 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
24705 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
24706 OPTION_MASK_VSX_TIMODE.
24707 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
24708 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
24710 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
24711 (TI2 iterator): New iterator for TImode, PTImode.
24712 (wd mode attribute): Add values for vector types.
24713 (movti_string): Replace TI move operations with operations for TImode
24714 and PTImode. Add support for TImode being allowed in VSX registers.
24715 (mov<mode>_string, TImode/PTImode): Likewise.
24716 (movti_ppc64): Likewise.
24717 (mov<mode>_ppc64, TImode/PTImode): Likewise.
24718 (TI mode splitters): Likewise.
24720 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
24723 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
24725 PR tree-optimization/56355
24726 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
24727 Also handle integers with undefined overflow.
24729 2013-03-20 Catherine Moore <clm@codesourcery.com>
24730 Maciej W. Rozycki <macro@codesourcery.com>
24731 Tom de Vries <tom@codesourcery.com>
24732 Nathan Sidwell <nathan@codesourcery.com>
24733 Iain Sandoe <iain@codesourcery.com>
24734 Nathan Froyd <froydnj@codesourcery.com>
24735 Chao-ying Fu <fu@mips.com>
24737 * doc/extend.texi: (micromips, nomicromips, nocompression):
24738 Document new function attributes.
24739 * doc/invoke.texi (minterlink-compressed, mmicromips,
24740 m14k, m14ke, m14kec): Document new options.
24741 (minterlink-mips16): Update documentation.
24742 * doc/md.texi (ZC, ZD): Document new constraints.
24743 * configure.ac (gcc_cv_as_micromips): Check if linker
24744 supports the .set micromips directive.
24745 * configure: Regenerate.
24746 * config.in: Regenerate.
24747 * config/mips/mips-tables.opt: Regenerate.
24748 * config/mips/micromips.md: New file.
24749 * constraints.md (ZC, ZD): New constraints.
24750 * config/mips/predicates.md (movep_src_register): New predicate.
24751 (movep_src_operand): New predicate.
24752 (non_volatile_mem_operand): New predicate.
24753 * config/mips/mips.md (multimem): New type.
24754 (length): Differentiate between 17-bit and 18-bit branch offsets.
24755 (MOVEP1, MOVEP2): New mode iterator.
24756 (mov_<load>l): Use ZC constraint.
24757 (mov_<load>r): Likewise.
24758 (mov_<store>l): Likewise.
24759 (mov_<store>r): Likewise.
24760 (*branch_equality<mode>_inverted): Add microMIPS support.
24761 (*branch_equality<mode>): Likewise.
24762 (*jump_absolute): Likewise.
24763 (indirect_jump_<mode>): Likewise.
24764 (tablejump_<mode>): Likewise.
24765 (<optab>_internal): Likewise.
24766 (sibcall_internal): Likewise.
24767 (sibcall_value_internal): Likewise.
24768 (prefetch): Use constraint ZD.
24769 * config/mips/mips.opt (minterlink-compressed): New option.
24770 (minterlink-mips16): Now an alias for minterlink-compressed.
24771 (mmicromips): New option.
24772 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24773 (compare_and_swap_12): Likewise.
24774 (sync_add<mode>): Likewise.
24775 (sync_<optab>_12): Likewise.
24776 (sync_old_<optab>_12): Likewise.
24777 (sync_new_<optab>_12): Likewise.
24778 (sync_nand_12): Likewise.
24779 (sync_old_nand_12): Likewise.
24780 (sync_new_nand_12): Likewise.
24781 (sync_sub<mode>): Likewise.
24782 (sync_old_add<mode>): Likewise.
24783 (sync_old_sub<mode>): Likewise.
24784 (sync_new_add<mode>): Likewise.
24785 (sync_new_sub<mode>): Likewise.
24786 (sync_<optab><mode>): Likewise.
24787 (sync_old_<optab><mode>): Likewise.
24788 (sync_new_<optab><mode>): Likewise.
24789 (sync_nand<mode>): Likewise.
24790 (sync_old_nand<mode>): Likewise.
24791 (sync_new_nand<mode>): Likewise.
24792 (sync_lock_test_and_set<mode>): Likewise.
24793 (test_and_set_12): Likewise.
24794 (atomic_compare_and_swap<mode>): Likewise.
24795 (atomic_exchange<mode>_llsc): Likewise.
24796 (atomic_fetch_add<mode>_llsc): Likewise.
24797 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24798 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24799 (umips_save_restore_pattern_p): Likewise.
24800 (umips_load_store_pair_p): Likewise.
24801 (umips_output_load_store_pair): Likewise.
24802 (umips_movep_target_p): Likewise.
24803 (umips_12bit_offset_address_p): Likewise.
24804 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24805 (mips_base_mips16): Rename this...
24806 (mips_base_compression_flags): ...to this. Update all uses.
24807 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24808 (mips_mips16_decl_p): Delete.
24809 (mips_nomips16_decl_p): Delete.
24810 (mips_get_compress_on_flags): New function.
24811 (mips_get_compress_off_flags): New function.
24812 (mips_get_compress_mode): New function.
24813 (mips_get_compress_on_name): New function.
24814 (mips_get_compress_off_name): New function.
24815 (mips_insert_attributes): Support multiple compression types.
24816 (mips_merge_decl_attributes): Likewise.
24817 (umips_12bit_offset_address_p): New function.
24818 (mips_start_function_definition): Emit .set micromips directive.
24819 (mips_call_may_need_jalx_p): New function.
24820 (mips_function_ok_for_sibcall): Add microMIPS support.
24821 (mips_print_operand_punctuation): Support short delay slots and
24823 (umips_swm_mask, umips_swm_encoding): New.
24824 (umips_build_save_restore): New function.
24825 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24826 (was_mips16_p): Remove.
24827 (old_compression_mode): New.
24828 (mips_set_compression_mode): New function.
24829 (mips_set_current_function): Add microMIPS support.
24830 (mips_option_override): Likewise.
24831 (umips_save_restore_pattern_p): New function.
24832 (umips_output_save_restore): New function.
24833 (umips_load_store_pair_p_1): New function.
24834 (umips_load_store_pair_p): New function.
24835 (umips_output_load_store_pair_1): New function.
24836 (umips_output_load_store_pair): New function.
24837 (umips_movep_target_p) New function.
24838 (mips_prepare_pch_save): Add microMIPS support.
24839 * config/mips/mips.h (TARGET_COMPRESSION): New.
24840 (TARGET_CPU_CPP_BUILTINS): Update macro
24841 to use new compression flags and to support microMIPS.
24842 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24843 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24844 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24845 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24846 (ASM_SPEC): Support mmicromips and mno-micromips.
24847 (M16STORE_REG_P): New macro.
24848 (MIPS_CALL): Support TARGET_MICROMIPS.
24849 (MICROMIPS_J): New macro.
24850 (mips_base_mips16): Rename this...
24851 (mips_base_compression_flags): ...to this.
24852 (UMIPS_12BIT_OFFSET_P): New macro.
24853 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24854 (MULTILIB_DIRNAMES): Likewise.
24855 2013-03-20 Richard Biener <rguenther@suse.de>
24857 PR tree-optimization/56661
24858 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
24859 the result does not have to be distinct.
24861 2013-03-20 Richard Biener <rguenther@suse.de>
24863 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
24866 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24867 Steven Bosscher <steven@gcc.gnu.org>
24869 PR rtl-optimization/56605
24870 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
24872 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24875 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
24876 that require movd instead of movq.
24878 2013-03-20 Richard Biener <rguenther@suse.de>
24880 * tree-ssa-structalias.c (struct variable_info): Add pointer
24881 to the first field of an aggregate with sub-vars. Make
24882 this and the pointer to the next subfield its ID.
24883 (vi_next): New function.
24884 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
24885 storedanything_id, integer_id): Increment by one.
24886 (new_var_info, get_call_vi, lookup_call_clobber_vi,
24887 get_call_clobber_vi): Adjust.
24888 (solution_set_expand): Simplify and speedup.
24889 (solution_set_add): Inline into ...
24890 (set_union_with_increment): ... this. Adjust accordingly.
24891 (do_sd_constraint): Likewise.
24892 (do_ds_constraint): Likewise.
24893 (do_complex_constraint): Simplify.
24894 (build_pred_graph): Adjust.
24895 (solve_graph): Likewise. Simplify and speedup.
24896 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
24897 get_constraint_for_component_ref, get_constraint_for_1,
24898 first_vi_for_offset, first_or_preceding_vi_for_offset,
24899 create_function_info_for, create_variable_info_for_1,
24900 create_variable_info_for, intra_create_variable_infos): Adjust.
24901 (init_base_vars): Push NULL for ID zero.
24902 (compute_points_to_sets): Adjust.
24904 2013-03-20 Richard Biener <rguenther@suse.de>
24906 * cfgloop.c (verify_loop_structure): Streamline and avoid
24907 ICEing on corrupt loop tree.
24908 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
24911 2013-03-20 Richard Biener <rguenther@suse.de>
24913 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
24914 check whether an SSA update is needed.
24916 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
24918 * config/mips/constraints.md (T): Rename to...
24922 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
24923 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
24925 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24927 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
24928 (*subsi3_carryin_uxtw): Likewise.
24930 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24932 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
24933 (*rorsi3_insn_uxtw): Likewise.
24935 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24937 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
24938 (*extrsi5_insn_uxtw): Likewise.
24940 2013-03-19 Richard Biener <rguenther@suse.de>
24942 PR tree-optimization/56273
24943 * passes.c (init_optimization_passes): Move second VRP after DOM.
24945 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
24947 * config/i386/i386.md (*movti_internal): Merge from
24948 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
24949 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
24950 nox64 isa attributes.
24952 2013-03-18 Richard Biener <rguenther@suse.de>
24954 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
24956 (merge_node_constraints): Likewise.
24957 (build_succ_graph): Likewise.
24958 (valid_graph_edge): Inline into single caller.
24959 (unify_nodes): Likewise. Use bitmap_set_bit return value
24961 (scc_visit): Fix formatting and variable use.
24962 (do_sd_constraint): Use gcc_checking_assert.
24963 (do_ds_constraint): Likewise.
24964 (do_complex_constraint): Likewise.
24965 (condense_visit): Likewise. Cleanup.
24966 (dump_pred_graph): New function.
24967 (perform_var_substitution): Dump the pred-graph before
24968 variable substitution.
24969 (find_equivalent_node): Use gcc_checking_assert.
24970 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
24972 2013-03-18 Richard Biener <rguenther@suse.de>
24974 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
24975 Remove cond_expr_stmt_list argument and do not gimplify the
24977 (vect_loop_versioning): Adjust.
24978 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
24979 Cleanup to use less temporaries.
24980 (vect_create_data_ref_ptr): Cleanup.
24982 2013-03-18 Jakub Jelinek <jakub@redhat.com>
24984 PR tree-optimization/56635
24985 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
24986 require types_compatible_p types.
24988 2013-03-18 Nick Clifton <nickc@redhat.com>
24990 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
24991 spurious backslash.
24993 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
24994 Add missing line to comment describing function.
24996 2013-03-18 Richard Biener <rguenther@suse.de>
24998 PR tree-optimization/56210
24999 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25000 Handle string / character search functions.
25001 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25003 2013-03-18 Richard Biener <rguenther@suse.de>
25005 PR middle-end/56483
25006 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25007 and implement properly.
25008 * gimple.h (gimple_cond_single_var_p): Remove.
25010 2013-03-18 Richard Biener <rguenther@suse.de>
25012 * tree-data-ref.h (find_data_references_in_loop): Declare.
25013 * tree-data-ref.c (get_references_in_stmt): Use a stack
25014 vector pre-allocated in the callers.
25015 (find_data_references_in_stmt): Adjust.
25016 (graphite_find_data_references_in_stmt): Likewise.
25017 (create_rdg_vertices): Likewise.
25018 (find_data_references_in_loop): Export.
25019 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25020 Compute dependences here...
25021 (vect_analyze_data_refs): ...not here. When we encounter
25022 a non-vectorizable data reference in basic-block vectorization
25023 truncate the data reference vector. Do not bother to
25024 fixup data-dependence information for gather loads.
25025 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25026 of data references, as reported.
25028 2013-03-18 Richard Biener <rguenther@suse.de>
25030 PR tree-optimization/3713
25031 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25032 has_constants and expr.
25033 (stmt_has_constants): Properly valueize SSA names when deciding
25034 whether the stmt has constants.
25036 2013-03-18 Richard Biener <rguenther@suse.de>
25038 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25039 whole function when there is nothing to do.
25040 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25041 * tree-vectorizer.c (vectorize_loops): Update virtual and
25042 loop-closed SSA once.
25043 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25045 2013-03-18 Richard Biener <rguenther@suse.de>
25047 PR middle-end/56113
25048 * domwalk.c (bb_postorder): New global static.
25049 (cmp_bb_postorder): New function.
25050 (walk_dominator_tree): Replace scheme imposing an order for
25051 visiting dominator sons by one sorting them at the time they
25052 are pushed on the stack.
25054 2013-03-18 Richard Biener <rguenther@suse.de>
25056 PR tree-optimization/39326
25057 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25058 (struct mem_ref): Replace mem member with ao_ref typed member.
25059 (MEM_ANALYZABLE): Adjust.
25060 (memref_eq): Likewise.
25061 (mem_ref_alloc): Likewise.
25062 (gather_mem_refs_stmt): Likewise.
25063 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25064 (execute_sm_if_changed_flag_set): Adjust.
25065 (execute_sm): Likewise.
25066 (ref_always_accessed_p): Likewise.
25067 (refs_independent_p): Likewise.
25068 (can_sm_ref_p): Likewise.
25070 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25073 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25074 return 1 even for !unsignedp.
25076 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25078 * config/i386/i386.md (isa): Add x64 and nox64.
25079 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25080 (*pushtf): Enable *roF alternative for x64 isa only.
25081 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25082 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25083 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25084 (*movtf_internal): Merge from *movtf_internal_rex64 and
25085 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25086 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25087 nox64 isa attributes.
25088 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25089 nox64 isa attributes.
25090 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25092 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25094 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25096 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25099 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25100 class names. Remove trailing comma after "ALL_REGS".
25102 2013-03-16 Jan Hubicka <jh@suse.cz>
25104 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25105 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25106 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25107 of cgraph_get_create_node.
25108 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25110 2013-03-16 Jason Merrill <jason@redhat.com>
25113 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25114 with DW_AT_default_value.
25116 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25118 * BASE-VER: Set to 4.9.0.
25120 2013-03-14 Andi Kleen <ak@linux.intel.com>
25123 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25124 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25125 Document _x* TSX intrinsics.
25127 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25128 David Holsgrove <david.holsgrove@xilinx.com>
25130 * configure.ac: Add MicroBlaze TLS support detection.
25131 * configure: Regenerate.
25132 * config/microblaze/microblaze-protos.h
25133 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25134 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25135 * config/microblaze/microblaze.c (microblaze_address_type): Add
25136 ADDRESS_TLS and tls_reloc address types.
25137 (microblaze_address_info): Add tls_reloc.
25138 (TARGET_HAVE_TLS): Define.
25139 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25140 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25141 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25142 load_tls_operand, microblaze_call_tls_get_addr,
25143 microblaze_legitimize_tls_address): New functions.
25144 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25145 (get_base_reg): Use microblaze_tls_symbol_p.
25146 (microblaze_classify_address): Handle TLS.
25147 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25148 label_mentioned_p and microblaze_tls_referenced_p.
25149 (microblaze_legitimize_address): Handle TLS.
25150 (microblaze_address_insns): Handle ADDRESS_TLS.
25151 (pic_address_needs_scratch): Handle TLS.
25152 (print_operand_address): Handle TLS.
25153 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25154 (microblaze_expand_move): Handle TLS.
25155 (microblaze_legitimate_constant_p): Check
25156 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25157 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25158 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25159 (PIC_OFFSET_TABLE_REGNUM): Set.
25160 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25161 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25162 (addsi3, movsi_internal2, movdf_internal): Update constraints
25163 * config/microblaze/predicates.md (arith_plus_operand): Define
25164 (move_operand): Redefine as move_src_operand,
25165 check microblaze_tls_referenced_p.
25167 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25169 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25170 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25172 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25174 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25177 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25179 PR tree-optimization/53265
25180 * common.opt (Waggressive-loop-optimizations): New option.
25181 * tree-ssa-loop-niter.c: Include tree-pass.h.
25182 (do_warn_aggressive_loop_optimizations): New function.
25183 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
25184 if number_of_latch_executions returned constant.
25185 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
25186 early. If number_of_latch_executions returned constant, set
25187 nb_iterations_upper_bound back to it.
25188 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
25190 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
25191 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
25193 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
25194 (MULTILIB_OSDIRNAMES): Set.
25195 * genmultilib: If defaultosdirname doesn't start with :: , set
25196 defaultosdirname2 instead, clear it and emit two . multilib_raw
25197 entries instead of just one.
25199 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
25201 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25202 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
25203 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25204 (SUBTARGET_OVERRIDE_OPTIONS): New.
25206 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
25209 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
25210 (musermode): Convert to Var(TARGET_USERMODE).
25211 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
25212 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
25213 * config/sh/sh.c (sh_option_override): Use
25214 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
25215 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
25217 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
25219 (udivsi3_i4_single, divsi3_i4_single): Use
25220 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
25222 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
25224 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
25227 2013-03-13 Richard Biener <rguenther@suse.de>
25229 PR tree-optimization/56608
25230 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
25231 calls when vectorizing basic-blocks.
25233 2013-03-13 Jakub Jelinek <jakub@redhat.com>
25236 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
25239 2013-03-12 Jakub Jelinek <jakub@redhat.com>
25241 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
25243 2013-03-11 Jan Hubicka <jh@suse.cz>
25246 * lto-streamer-out.c (output_symbol_p): Skip references from
25247 constructors of external variables.
25249 2013-03-11 Jan Hubicka <jh@suse.cz>
25251 PR middle-end/56571
25252 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
25254 * emit-rtl.c (verify_rtx_sharing): Likewise.
25255 (copy_insn_1): Likewise.
25256 * rtl.c (copy_rtx): Likewise.
25258 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
25261 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
25262 output_operand_lossage message.
25264 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
25267 * arm.c (shift_op): Validate RTL pattern on the fly.
25268 (arm_print_operand, case 'S'): Don't use shift_operator to validate
25271 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25274 * config/pa/pa.md (call_value): Check for calls to powf and direct to
25275 new call patterns that clobber %fr12.
25276 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
25277 split and postreload patterns.
25278 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
25279 registers %fr12 and %fr12R as call used.
25281 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
25283 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
25284 (canon_address, record_store, replace_read, check_mem_read_rtx,
25285 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
25286 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
25287 rest_of_handle_dse): Likewise.
25289 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
25291 PR middle-end/56524
25292 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
25294 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
25295 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
25296 (save_optabs_if_changed): Replace with...
25297 (init_tree_optimization_optabs): ...this.
25298 * optabs.c (save_optabs_if_changed): Rename to...
25299 (init_tree_optimization_optabs): ...this. Take the optimization node
25300 as argument. Do nothing if the base optabs are already correct.
25301 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
25302 to recompute optabs.
25303 * function.h (function): Remove optabs field.
25304 * function.c (invoke_set_current_function_hook): Call
25305 init_tree_optimization_optabs. Use the result to initialize
25308 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
25310 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
25311 if GTMA_HAS_NO_INSTRUMENTATION.
25312 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
25313 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
25314 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
25315 * gimple-pretty-print.c (dump_gimple_transaction): Handle
25316 GTMA_HAS_NO_INSTRUMENTATION.
25318 2013-03-08 Jakub Jelinek <jakub@redhat.com>
25320 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
25323 2013-03-08 Marek Polacek <polacek@redhat.com>
25324 Jakub Jelinek <jakub@redhat.com>
25326 PR tree-optimization/56478
25327 * predict.c (is_comparison_with_loop_invariant_p): Change the
25328 type of loop_step to tree.
25329 (predict_loops): Adjust.
25330 (predict_iv_comparison): Perform the computations on double_ints.
25332 2013-03-08 Richard Biener <rguenther@suse.de>
25334 PR tree-optimization/56570
25335 * tree-cfg.c (verify_expr_location_1): Verify locations for
25337 * tree-sra.c (create_access_replacement): Strip locations
25338 from DECL_DEBUG_EXPRs.
25340 2013-03-08 Richard Biener <rguenther@suse.de>
25342 * tree-inline.c (expand_call_inline): Do not associate
25343 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
25344 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
25346 2013-03-08 Richard Biener <rguenther@suse.de>
25348 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
25349 or block changes with -Og. Fix for location / block encoding
25350 changes and PHI arguments with locations.
25352 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
25354 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
25356 (struct output_info): Likewise.
25357 (register_overhead): Remove bad gcc_assert.
25358 (bitmap_find_bit): If there is only a single bitmap element, do not
25359 count a miss as a search.
25360 (print_statistics): Update for counter type changes.
25361 (dump_bitmap_statistics): Likewise. Print headers such that they
25362 are properly lined up with the printed counters.
25364 2013-03-07 Jakub Jelinek <jakub@redhat.com>
25366 PR tree-optimization/56559
25367 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
25368 check that it has only a single use.
25370 2013-03-07 Richard Biener <rguenther@suse.de>
25372 * doc/invoke.texi (fwhole-program): Discourage use in combination
25375 2013-03-06 Jakub Jelinek <jakub@redhat.com>
25377 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
25379 PR tree-optimization/56539
25380 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
25381 instead of GSI_CONTINUE_LINKING as last argument to
25382 force_gimple_operand_gsi. Adjust function comment.
25384 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
25387 PR middle-end/56548
25388 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
25389 promoted mode, convert the result back to the original mode.
25391 2013-03-06 Richard Biener <rguenther@suse.de>
25393 PR middle-end/56294
25394 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
25395 (insert_updated_phi_nodes_compare_uids): New function.
25396 (update_ssa): Sort symbols_to_rename after UID before
25397 traversing it to insert PHI nodes.
25399 2013-03-06 Richard Biener <rguenther@suse.de>
25401 PR middle-end/50494
25402 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
25403 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
25406 2013-02-13 Richard Biener <rguenther@suse.de>
25409 * varasm.c (output_constant_def_1): Get the decl representing
25410 the constant as argument.
25411 (output_constant_def): Wrap output_constant_def_1.
25412 (make_decl_rtl): Use output_constant_def_1 with the decl
25413 representing the constant.
25414 (build_constant_desc): Optionally re-use a decl already
25415 representing the constant.
25416 (tree_output_constant_def): Adjust.
25418 2013-03-06 Joey Ye <joey.ye@arm.com>
25421 * gcc.c (convert_white_space): New function.
25422 (main): Handles white space in function name.
25424 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
25427 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
25428 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
25429 to SH_DIV_CALL_TABLE for TARGET_SH2.
25430 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
25432 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
25433 call-table options.
25435 2013-03-05 Sterling Augustine <saugustine@google.com>
25436 Cary Coutant <ccoutant@google.com>
25439 * dwarf2out.c (resolve_addr): Don't call
25440 remove_loc_list_addr_table_entries a second time for the same
25443 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25446 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
25447 (avoid_complex_debug_insns): New function.
25448 (expand_debug_locations): Call it.
25450 PR rtl-optimization/56484
25451 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
25452 lifetimes of hard registers on small register class machines.
25454 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
25456 * config/microblaze/microblaze-protos.h: Rename
25457 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
25458 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
25460 (microblaze_fast_interrupt_function_p): New function.
25461 (microblaze_is_interrupt_handler): Rename to
25462 microblaze_is_interrupt_variant and add fast_interrupt check.
25463 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
25464 (save_restore_insns): Likewise.
25465 (compute_frame_size): Likewise.
25466 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
25467 (microblaze_globalize_label): Likewise.
25468 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
25469 * config/microblaze/microblaze.md: Use wrapper
25470 microblaze_is_interrupt_variant.
25472 2013-03-05 Kai Tietz <ktietz@redhat.com>
25474 * sdbout.c (sdbout_one_type): Switch to current function's section
25475 supporting cold/hot.
25477 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
25479 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
25482 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25484 PR middle-end/56461
25485 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
25486 if VALGRIND_GET_VBITS is defined, temporarily make object
25487 memory all defined, and restore previous valgrind addressability
25488 and definability afterwards. Free this_object at the end.
25490 PR middle-end/56461
25491 * lra.c (lra): Call lra_clear_live_ranges if live_p,
25492 right before calling lra_create_live_ranges, also call it
25493 when clearing live_p. Only call lra_clear_live_ranges
25494 at the end if live_p.
25496 PR middle-end/56461
25497 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
25499 2013-03-05 Richard Biener <rguenther@suse.de>
25501 PR tree-optimization/56521
25502 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
25505 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
25508 * except.h (remove_unreachable_eh_regions): New prototype.
25509 * except.c (remove_eh_handler_splicer): New function, split out
25510 of remove_eh_handler.
25511 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
25512 warning about running it on many EH regions one at a time.
25513 (remove_unreachable_eh_regions_worker): New function, walk the
25514 EH tree in depth-first order and remove non-marked regions.
25515 (remove_unreachable_eh_regions): New function.
25516 * tree-eh.c (mark_reachable_handlers): New function, split out
25517 from remove_unreachable_handlers.
25518 (remove_unreachable_handlers): Use mark_reachable_handlers and
25519 remove_unreachable_eh_regions.
25520 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
25521 and remove_unreachable_eh_regions.
25523 2013-03-05 Richard Biener <rguenther@suse.de>
25525 PR middle-end/56525
25526 * loop-init.c (fix_loop_structure): Remove loops in two stages,
25527 not freeing them until the end.
25529 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25531 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
25533 2013-03-05 Richard Biener <rguenther@suse.de>
25535 PR tree-optimization/56270
25536 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
25537 of loads after scheduling an SLP instance.
25539 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25541 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
25543 (check_gcc_parallelize): Run guality.exp as a separate job from
25544 vect.exp with unsorted.exp and $(dg_target_exps) separately from
25545 struct-layout-1.exp with stackalign.exp.
25547 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
25549 PR middle-end/56461
25550 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
25551 load_index sbitmap even if some bit in it isn't set.
25553 PR middle-end/56461
25554 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
25555 (discover_iteration_bound_by_body_walk): Change queues to
25556 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
25557 spelling in comment. Call safe_push on queues[bound_index] directly.
25558 Release queues[queue_index] in every iteration unconditionally.
25559 Release bounds vector.
25561 PR middle-end/56461
25562 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
25563 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
25564 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
25567 2013-03-05 Richard Biener <rguenther@suse.de>
25570 * tree-inline.c (remap_blocks_to_null): New function.
25571 (expand_call_inline): When expanding a call stmt without
25572 an associated block inline remap all callee blocks to NULL.
25574 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25576 PR rtl-optimization/56494
25577 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
25578 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
25579 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
25581 PR middle-end/56461
25582 * sel-sched-ir.c (free_sched_pools): Release
25583 succs_info_pool.stack[succs_info_pool.max_top] vectors too
25584 if succs_info_pool.max_top isn't -1.
25587 * opts.c (opts_obstack, opts_concat): Moved to...
25588 * opts-common.c (opts_obstack, opts_concat): ... here.
25590 2013-03-04 Jakub Jelinek <jakub@redhat.com>
25592 PR middle-end/56461
25593 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
25595 2013-03-04 Martin Jambor <mjambor@suse.cz>
25597 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
25598 all appropriate places.
25600 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
25602 PR tree-optimization/56424
25603 * ipa-split.c (split_function): Do not set the RSO flag if result is
25604 not by reference and its type is a register type.
25606 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
25608 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
25609 (microblaze_legitimate_pic_operand): Likewise
25610 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
25611 new function microblaze_legitimate_pic_operand
25612 * config/microblaze/microblaze-protos.h
25613 (microblaze_legitimate_pic_operand): Declare.
25615 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25617 * config/microblaze/predicates.md (call_insn_simple_operand):
25618 New predicate for supported rtx code types.
25619 * config/microblaze/microblaze.md (call_internal1): Use
25620 call_insn_simple_operand predicate.
25622 2013-03-04 Jakub Jelinek <jakub@redhat.com>
25624 PR middle-end/56461
25625 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
25626 partitions.ordered_remove.
25628 PR middle-end/56461
25629 * tree-vect-stmts.c (vectorizable_conversion): Don't call
25630 vec_oprnds0.create (1) for modifier == NONE.
25632 PR middle-end/56461
25633 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
25634 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
25635 vec_oprnds1 right before pushing anything to it for
25638 PR middle-end/56461
25639 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
25640 set nbbs to 0 instead of having separate code path.
25641 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
25642 instead of false as last argument if returning NULL.
25644 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
25646 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
25647 the attribute is now called "target" instead of "option".
25648 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
25649 * doc/tm.texi.in (Target Attributes): Likewise document the correct
25650 attribute/pragma name for TARGET_OPTION_VALID_P and
25651 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
25652 * doc/tm.texi: Regenerated.
25654 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
25656 * config/microblaze/microblaze.c:
25657 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
25658 * config/microblaze/microblaze.h: Add -mxl-reorder to
25660 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
25661 instructions emitted if TARGET_REORDER.
25662 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
25663 or 0 for -m/-mno case, but initialises as 2 to detect default use case
25666 2013-03-01 Xinliang David Li <davidxl@google.com>
25668 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
25671 2013-03-01 Jakub Jelinek <jakub@redhat.com>
25673 PR middle-end/56461
25674 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
25675 vector even when returning true. Fix up function comment formatting.
25677 PR middle-end/56461
25678 * ira-build.c (ira_loop_nodes_count): New variable.
25679 (create_loop_tree_nodes): Initialize it.
25680 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
25682 PR middle-end/56461
25683 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
25684 method on dr_chain and result_chain.
25685 * tree-vect-stmts.c (vectorizable_store): Only call
25686 result_chain.create if j == 0.
25688 PR middle-end/56461
25689 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
25690 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
25691 before overwriting it.
25693 2013-03-01 Tobias Burnus <burnus@net-b.de>
25695 * doc/extended.texi (C Extensions): Change order in @menu
25697 (Other MIPS Built-in Functions): Move last MIPS entry before
25698 "picoChip Built-in Functions".
25699 (SH Built-in Functions): Move after RX Built-in Functions.
25700 * doc/gcc.texi (Introduction): Change order in @menu
25702 * doc/md.texi (Constraints): Ditto.
25703 * gty.texi (Type Information): Ditto.
25704 (User-provided marking routines for template types): Make
25706 * doc/invoke.texi (AArch64 Options): Move before
25707 "Adapteva Epiphany Options".
25709 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
25710 Jakub Jelinek <jakub@redhat.com>
25713 * asan.c (gate_asan): Lookup no_sanitize_address instead of
25714 no_address_safety_analysis attribute.
25715 * doc/extend.texi (no_address_safety_attribute): Rename to
25716 no_sanitize_address attribute, mention no_address_safety_analysis
25717 attribute as deprecated alias.
25719 2013-02-28 Jakub Jelinek <jakub@redhat.com>
25721 PR middle-end/56461
25722 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
25723 type to vec<vec<tree> > *.
25724 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
25725 to be vec<tree> instead of vec<tree> *, set vec_defs
25726 to vNULL and call vec_defs.create (number_of_vects), adjust other
25728 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
25729 vectorizable_condition): Adjust vect_get_slp_defs callers.
25731 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
25733 * config/aarch64/aarch64.c
25734 (aarch64_float_const_representable): Remove unused variable.
25736 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
25738 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
25740 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
25742 * config/aarch64/aarch64-builtins.c
25743 (aarch64_init_simd_builtins): Make static.
25745 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
25747 * config/aarch64/aarch64.c
25748 (aarch64_simd_make_constant): Make static.
25750 2013-02-28 Martin Jambor <mjambor@suse.cz>
25752 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
25753 with no initialization to the RHS of debug statements.
25755 2013-02-28 Martin Jambor <mjambor@suse.cz>
25757 PR tree-optimization/56294
25758 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
25760 (get_access_replacement): Do not call create_access_replacement.
25761 Assert a replacement exists.
25762 (get_repl_default_def_ssa_name): Create the replacement declaration
25765 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25767 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
25768 final_end_function.
25770 2013-02-28 Marek Polacek <polacek@redhat.com>
25772 PR rtl-optimization/56466
25773 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
25774 if we're changing a loop.
25775 (peel_loops_completely): Likewise.
25777 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
25780 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
25782 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
25785 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
25786 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
25787 INTX_FTYPE_FX, FX_FTYPE_INTX.
25788 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
25790 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
25792 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
25793 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
25794 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
25795 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
25796 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
25797 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
25798 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
25799 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
25800 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
25801 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
25802 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
25803 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
25804 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
25805 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
25806 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
25807 (avrxmega6): Increase max flash segments from 5 to 6.
25808 * config/avr/t-multilib: Regenerate.
25809 * config/avr/avr-tables.opt: Regenerate.
25810 * doc/avr-mmcu.texi: Regenerate.
25812 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
25814 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
25815 (avr_device_to_arch): Rename to avr_device_to_ld.
25816 (avr_device_to_as): New prototype.
25817 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
25818 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
25819 * config/avr/driver-avr.c (avr_device_to_as): New.
25820 (avr_device_to_arch): Rename to avr_device_to_ld.
25822 2013-02-27 Jakub Jelinek <jakub@redhat.com>
25824 PR middle-end/56461
25825 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
25826 method on dr_chain and result_chain.
25828 PR middle-end/56461
25829 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
25830 pointer_set_destroy on not_executed_last_iteration.
25832 PR middle-end/56461
25833 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
25835 PR middle-end/56461
25836 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
25837 FOR_EACH_DEFINED_FUNCTION when freeing state.
25839 PR middle-end/56461
25840 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
25842 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
25845 PR middle-end/56461
25846 * ipa-cp.c (decide_whether_version_node): Call vec_free on
25847 known_aggs[i].items and release known_aggs vector.
25849 PR middle-end/56461
25850 * ipa-reference.c (propagate): Free node_info even for alias nodes.
25852 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25854 * config/microblaze/microblaze.c (microblaze_emit_compare):
25855 Use xor for EQ/NE comparisions.
25856 * config/microblaze/microblaze.md (cstoresf4): Add constraints
25857 (cbranchsf4): Adjust operator to comparison_operator.
25859 2013-02-27 Jakub Jelinek <jakub@redhat.com>
25861 PR middle-end/56461
25862 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
25864 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
25865 vec_safe_push, always update *slot.
25866 (redirect_edge_var_map_clear): Use vec_free.
25867 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
25868 (free_var_map_entry): Use vec_free.
25869 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
25870 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
25872 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
25874 PR middle-end/45472
25875 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
25876 when the may_trap_p bit of the exprs being merged differs.
25877 Reorder tests for speculativeness in the logical and operator.
25879 2013-02-27 Jakub Jelinek <jakub@redhat.com>
25881 * incpath.c (add_standard_paths): Use reconcat instead of concat
25882 where appropriate and avoid leaking memory.
25884 * opts.h: Include obstack.h.
25885 (opts_concat): New prototype.
25886 (opts_obstack): New declaration.
25887 * opts.c (opts_concat): New function.
25888 (opts_obstack): New variable.
25889 (init_options_struct): Call gcc_init_obstack on opts_obstack.
25890 (finish_options): Use opts_concat instead of concat
25891 and XOBNEWVEC instead of XNEWVEC.
25892 * opts-common.c (generate_canonical_option, decode_cmdline_option,
25893 generate_option): Likewise.
25894 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
25895 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
25898 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
25899 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
25901 2013-02-26 Jakub Jelinek <jakub@redhat.com>
25903 PR middle-end/56461
25904 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
25906 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
25908 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
25909 (arm_block_move_unaligned_straight): Likewise.
25910 (arm_adjust_block_mem): Likewise.
25912 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
25915 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
25916 temp, cond and label.
25917 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
25920 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
25921 * config/c6x/c6x.h (dbx_register_map): Update declaration.
25924 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
25925 of prologue/epilogue functions.
25928 * config/tilegx/tilegx.c (tilegx_expand_prologue):
25929 Remove unused variable cfa_offset.
25930 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
25933 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
25934 type promotion to unsigned.
25937 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
25938 for HOST_WIDE_INT of 32 bit / same size as int.
25939 (arm_block_move_unaligned_straight): Likewise.
25940 (arm_adjust_block_mem): Likewise.
25943 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
25946 2013-02-26 Marek Polacek <polacek@redhat.com>
25948 PR tree-optimization/56426
25949 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
25951 2013-02-26 Richard Biener <rguenther@suse.de>
25954 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
25955 unused variable loops.
25957 2013-02-26 Jakub Jelinek <jakub@redhat.com>
25959 PR tree-optimization/56448
25960 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
25961 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
25962 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
25963 later operands of the references, or even first operand for
25964 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
25966 PR tree-optimization/56443
25967 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
25968 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
25969 to type_for_mode langhook.
25971 2013-02-25 Matt Turner <mattst88@gmail.com>
25973 * doc/invoke.texi: Document r4700.
25975 2013-02-25 Richard Biener <rguenther@suse.de>
25977 PR tree-optimization/56175
25978 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
25980 (simplify_bitwise_binary): ... here. Also guard the conversion
25981 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
25983 2013-02-25 Catherine Moore <clm@codesourcery.com>
25986 2013-02-24 Catherine Moore <clm@codesourcery.com>
25987 Maciej W. Rozycki <macro@codesourcery.com>
25988 Tom de Vries <tom@codesourcery.com>
25989 Nathan Sidwell <nathan@codesourcery.com>
25990 Iain Sandoe <iain@codesourcery.com>
25991 Nathan Froyd <froydnj@codesourcery.com>
25992 Chao-ying Fu <fu@mips.com>
25994 * doc/extend.texi: (micromips, nomicromips, nocompression):
25995 Document new function attributes.
25996 * doc/invoke.texi (minterlink-compressed, mmicromips,
25997 m14k, m14ke, m14kec): Document new options.
25998 (minterlink-mips16): Update documentation.
25999 * doc/md.texi (ZC, ZD): Document new constraints.
26000 * configure.ac (gcc_cv_as_micromips): Check if linker
26001 supports the .set micromips directive.
26002 * configure: Regenerate.
26003 * config.in: Regenerate.
26004 * config/mips/mips-tables.opt: Regenerate.
26005 * config/mips/micromips.md: New file.
26006 * constraints.md (ZC, AD): New constraints.
26007 * config/mips/predicates.md (movep_src_register): New predicate.
26008 (movep_src_operand): New predicate.
26009 (non_volatile_mem_operand): New predicate.
26010 * config/mips/mips.md (multimem): New type.
26011 (length): Differentiate between 17-bit and 18-bit branch offsets.
26012 (MOVEP1, MOVEP2): New mode iterator.
26013 (mov_<load>l): Use ZC constraint.
26014 (mov_<load>r): Likewise.
26015 (mov_<store>l): Likewise.
26016 (mov_<store>r): Likewise.
26017 (*branch_equality<mode>_inverted): Add microMIPS support.
26018 (*branch_equality<mode>): Likewise.
26019 (*jump_absolute): Likewise.
26020 (indirect_jump_<mode>): Likewise.
26021 (tablejump_<mode>): Likewise.
26022 (<optab>_internal): Likewise.
26023 (sibcall_internal): Likewise.
26024 (sibcall_value_internal): Likewise.
26025 (prefetch): Use constraint ZD.
26026 * config/mips/mips.opt (minterlink-compressed): New option.
26027 (minterlink-mips16): Now an alias for minterlink-compressed.
26028 (mmicromips): New option.
26029 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26030 (compare_and_swap_12): Likewise.
26031 (sync_add<mode>): Likewise.
26032 (sync_<optab>_12): Likewise.
26033 (sync_old_<optab>_12): Likewise.
26034 (sync_new_<optab>_12): Likewise.
26035 (sync_nand_12): Likewise.
26036 (sync_old_nand_12): Likewise.
26037 (sync_new_nand_12): Likewise.
26038 (sync_sub<mode>): Likewise.
26039 (sync_old_add<mode>): Likewise.
26040 (sync_old_sub<mode>): Likewise.
26041 (sync_new_add<mode>): Likewise.
26042 (sync_new_sub<mode>): Likewise.
26043 (sync_<optab><mode>): Likewise.
26044 (sync_old_<optab><mode>): Likewise.
26045 (sync_new_<optab><mode>): Likewise.
26046 (sync_nand<mode>): Likewise.
26047 (sync_old_nand<mode>): Likewise.
26048 (sync_new_nand<mode>): Likewise.
26049 (sync_lock_test_and_set<mode>): Likewise.
26050 (test_and_set_12): Likewise.
26051 (atomic_compare_and_swap<mode>): Likewise.
26052 (atomic_exchange<mode>_llsc): Likewise.
26053 (atomic_fetch_add<mode>_llsc): Likewise.
26054 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26055 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26056 (umips_save_restore_pattern_p): Likewise.
26057 (umips_load_store_pair_p): Likewise.
26058 (umips_output_load_store_pair): Likewise.
26059 (umips_movep_target_p): Likewise.
26060 (umips_12bit_offset_address_p): Likewise.
26061 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26062 (mips_base_mips16): Rename this...
26063 (mips_base_compression_flags): ...to this. Update all uses.
26064 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26065 (mips_mips16_decl_p): Delete.
26066 (mips_nomips16_decl_p): Delete.
26067 (mips_get_compress_on_flags): New function.
26068 (mips_get_compress_off_flags): New function.
26069 (mips_get_compress_mode): New function.
26070 (mips_get_compress_on_name): New function.
26071 (mips_get_compress_off_name): New function.
26072 (mips_insert_attributes): Support multiple compression types.
26073 (mips_merge_decl_attributes): Likewise.
26074 (umips_12bit_offset_address_p): New function.
26075 (mips_start_function_definition): Emit .set micromips directive.
26076 (mips_call_may_need_jalx_p): New function.
26077 (mips_function_ok_for_sibcall): Add microMIPS support.
26078 (mips_print_operand_punctuation): Support short delay slots and
26080 (umips_swm_mask, umips_swm_encoding): New.
26081 (umips_build_save_restore): New function.
26082 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26083 (was_mips16_p): Remove.
26084 (old_compression_mode): New.
26085 (mips_set_compression_mode): New function.
26086 (mips_set_current_function): Add microMIPS support.
26087 (mips_option_override): Likewise.
26088 (umips_save_restore_pattern_p): New function.
26089 (umips_output_save_restore): New function.
26090 (umips_load_store_pair_p_1): New function.
26091 (umips_load_store_pair_p): New function.
26092 (umips_output_load_store_pair_1): New function.
26093 (umips_output_load_store_pair): New function.
26094 (umips_movep_target_p) New function.
26095 (mips_prepare_pch_save): Add microMIPS support.
26096 * config/mips/mips.h (TARGET_COMPRESSION): New.
26097 (TARGET_CPU_CPP_BUILTINS): Update macro
26098 to use new compression flags and to support microMIPS.
26099 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26100 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26101 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26102 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26103 (ASM_SPEC): Support mmicromips and mno-micromips.
26104 (M16STORE_REG_P): New macro.
26105 (MIPS_CALL): Support TARGET_MICROMIPS.
26106 (MICROMIPS_J): New macro.
26107 (mips_base_mips16): Rename this...
26108 (mips_base_compression_flags): ...to this.
26109 (UMIPS_12BIT_OFFSET_P): New macro.
26110 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26111 (MULTILIB_DIRNAMES): Likewise.
26113 2013-02-25 Tom de Vries <tom@codesourcery.com>
26115 PR rtl-optimization/56131
26116 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26117 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26118 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26120 2013-02-25 Tobias Burnus <burnus@net-b.de>
26122 * doc/invoke.texi (-fsanitize=): Move from optimization
26123 to debugging options.
26125 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26127 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26129 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26130 Alexander Monakov <amonakov@ispras.ru>
26132 PR middle-end/56077
26133 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26134 flush pending lists also on non-jumps. Adjust comment.
26136 2013-02-24 Catherine Moore <clm@codesourcery.com>
26137 Maciej W. Rozycki <macro@codesourcery.com>
26138 Tom de Vries <tom@codesourcery.com>
26139 Nathan Sidwell <nathan@codesourcery.com>
26140 Iain Sandoe <iain@codesourcery.com>
26141 Nathan Froyd <froydnj@codesourcery.com>
26142 Chao-ying Fu <fu@mips.com>
26144 * doc/extend.texi: (micromips, nomicromips, nocompression):
26145 Document new function attributes.
26146 * doc/invoke.texi (minterlink-compressed, mmicromips,
26147 m14k, m14ke, m14kec): Document new options.
26148 (minterlink-mips16): Update documentation.
26149 * doc/md.texi (ZC, ZD): Document new constraints.
26150 * configure.ac (gcc_cv_as_micromips): Check if linker
26151 supports the .set micromips directive.
26152 * configure: Regenerate.
26153 * config.in: Regenerate.
26154 * config/mips/mips-tables.opt: Regenerate.
26155 * config/mips/micromips.md: New file.
26156 * constraints.md (ZC, AD): New constraints.
26157 * config/mips/predicates.md (movep_src_register): New predicate.
26158 (movep_src_operand): New predicate.
26159 (non_volatile_mem_operand): New predicate.
26160 * config/mips/mips.md (multimem): New type.
26161 (length): Differentiate between 17-bit and 18-bit branch offsets.
26162 (MOVEP1, MOVEP2): New mode iterator.
26163 (mov_<load>l): Use ZC constraint.
26164 (mov_<load>r): Likewise.
26165 (mov_<store>l): Likewise.
26166 (mov_<store>r): Likewise.
26167 (*branch_equality<mode>_inverted): Add microMIPS support.
26168 (*branch_equality<mode>): Likewise.
26169 (*jump_absolute): Likewise.
26170 (indirect_jump_<mode>): Likewise.
26171 (tablejump_<mode>): Likewise.
26172 (<optab>_internal): Likewise.
26173 (sibcall_internal): Likewise.
26174 (sibcall_value_internal): Likewise.
26175 (prefetch): Use constraint ZD.
26176 * config/mips/mips.opt (minterlink-compressed): New option.
26177 (minterlink-mips16): Now an alias for minterlink-compressed.
26178 (mmicromips): New option.
26179 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26180 (compare_and_swap_12): Likewise.
26181 (sync_add<mode>): Likewise.
26182 (sync_<optab>_12): Likewise.
26183 (sync_old_<optab>_12): Likewise.
26184 (sync_new_<optab>_12): Likewise.
26185 (sync_nand_12): Likewise.
26186 (sync_old_nand_12): Likewise.
26187 (sync_new_nand_12): Likewise.
26188 (sync_sub<mode>): Likewise.
26189 (sync_old_add<mode>): Likewise.
26190 (sync_old_sub<mode>): Likewise.
26191 (sync_new_add<mode>): Likewise.
26192 (sync_new_sub<mode>): Likewise.
26193 (sync_<optab><mode>): Likewise.
26194 (sync_old_<optab><mode>): Likewise.
26195 (sync_new_<optab><mode>): Likewise.
26196 (sync_nand<mode>): Likewise.
26197 (sync_old_nand<mode>): Likewise.
26198 (sync_new_nand<mode>): Likewise.
26199 (sync_lock_test_and_set<mode>): Likewise.
26200 (test_and_set_12): Likewise.
26201 (atomic_compare_and_swap<mode>): Likewise.
26202 (atomic_exchange<mode>_llsc): Likewise.
26203 (atomic_fetch_add<mode>_llsc): Likewise.
26204 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26205 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26206 (umips_save_restore_pattern_p): Likewise.
26207 (umips_load_store_pair_p): Likewise.
26208 (umips_output_load_store_pair): Likewise.
26209 (umips_movep_target_p): Likewise.
26210 (umips_12bit_offset_address_p): Likewise.
26211 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26212 (mips_base_mips16): Rename this...
26213 (mips_base_compression_flags): ...to this. Update all uses.
26214 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26215 (mips_mips16_decl_p): Delete.
26216 (mips_nomips16_decl_p): Delete.
26217 (mips_get_compress_on_flags): New function.
26218 (mips_get_compress_off_flags): New function.
26219 (mips_get_compress_mode): New function.
26220 (mips_get_compress_on_name): New function.
26221 (mips_get_compress_off_name): New function.
26222 (mips_insert_attributes): Support multiple compression types.
26223 (mips_merge_decl_attributes): Likewise.
26224 (umips_12bit_offset_address_p): New function.
26225 (mips_start_function_definition): Emit .set micromips directive.
26226 (mips_call_may_need_jalx_p): New function.
26227 (mips_function_ok_for_sibcall): Add microMIPS support.
26228 (mips_print_operand_punctuation): Support short delay slots and
26230 (umips_swm_mask, umips_swm_encoding): New.
26231 (umips_build_save_restore): New function.
26232 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26233 (was_mips16_p): Remove.
26234 (old_compression_mode): New.
26235 (mips_set_compression_mode): New function.
26236 (mips_set_current_function): Add microMIPS support.
26237 (mips_option_override): Likewise.
26238 (umips_save_restore_pattern_p): New function.
26239 (umips_output_save_restore): New function.
26240 (umips_load_store_pair_p_1): New function.
26241 (umips_load_store_pair_p): New function.
26242 (umips_output_load_store_pair_1): New function.
26243 (umips_output_load_store_pair): New function.
26244 (umips_movep_target_p) New function.
26245 (mips_prepare_pch_save): Add microMIPS support.
26246 * config/mips/mips.h (TARGET_COMPRESSION): New.
26247 (TARGET_CPU_CPP_BUILTINS): Update macro
26248 to use new compression flags and to support microMIPS.
26249 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26250 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26251 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26252 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26253 (ASM_SPEC): Support mmicromips and mno-micromips.
26254 (M16STORE_REG_P): New macro.
26255 (MIPS_CALL): Support TARGET_MICROMIPS.
26256 (MICROMIPS_J): New macro.
26257 (mips_base_mips16): Rename this...
26258 (mips_base_compression_flags): ...to this.
26259 (UMIPS_12BIT_OFFSET_P): New macro.
26260 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26261 (MULTILIB_DIRNAMES): Likewise.
26263 2013-02-24 Jakub Jelinek <jakub@redhat.com>
26266 * target-globals.c (save_target_globals): For init_reg_sets and
26267 target_reinit remporarily set this_fn_optabs to this_target_optabs.
26269 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
26271 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
26272 * config/aarch64/t-aarch64
26273 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
26275 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
26277 PR inline-asm/56148
26278 * lra-constraints.c (process_alt_operands): Reload operand
26279 conflicting with earlier clobber only if no more other conflicting
26282 2013-02-22 Jakub Jelinek <jakub@redhat.com>
26285 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
26286 if not linking a shared library.
26288 2013-02-22 Seth LaForge <sethml@google.com>
26290 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
26292 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
26294 * config/arm/arm.md (split for extendsidi): Update condition.
26295 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
26296 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
26297 (qhs_zextenddi_cstr): Likewise.
26299 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26301 PR middle-end/56420
26302 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
26303 avoid signed wrapping.
26304 (expand_mult): Handle properly multiplication by
26305 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
26306 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
26307 in the compiler if coeff is HOST_WIDE_INT_MIN.
26308 (expand_divmod): Don't make ext_op1 static, change it's type to
26309 uhwi. Avoid undefined behavior in -INTVAL (op1).
26311 PR rtl-optimization/50339
26312 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
26314 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
26315 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
26316 into splitting_ashiftrt field.
26317 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
26319 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
26322 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
26324 PR middle-end/56108
26325 * trans-mem.c (execute_tm_mark): Do not expand transactions that
26326 are sure to go irrevocable.
26328 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
26330 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
26331 scalars are valid operands.
26333 2013-02-21 Martin Jambor <mjambor@suse.cz>
26335 PR tree-optimization/56310
26336 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
26337 only matching indices and non-negative final offsets.
26338 (intersect_aggregates_with_edge): Pass src_idx to
26339 agg_replacements_to_vector. Pass src_idx insstead of index to
26340 intersect_with_agg_replacements.
26342 2013-02-21 Martin Jambor <mjambor@suse.cz>
26344 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
26345 instead of hard-wired defaults.
26347 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
26349 * doc/invoke.texi (MIPS Options): Update documentation of the
26350 floating-point multiply-accumulate instruction restrictions.
26352 2013-02-21 Kostya Serebryany <kcc@google.com>
26354 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
26355 asan_shadow_offset on x86_64 linux.
26357 2013-02-21 Richard Biener <rguenther@suse.de>
26359 PR tree-optimization/56415
26361 2013-02-11 Richard Biener <rguenther@suse.de>
26363 PR tree-optimization/56273
26364 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
26367 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26370 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
26373 PR inline-asm/56405
26374 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
26375 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
26377 2013-02-20 Jan Hubicka <jh@suse.cz>
26379 PR tree-optimization/56265
26380 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
26381 when target is referenced for first time.
26383 2013-02-20 Richard Biener <rguenther@suse.de>
26385 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
26386 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
26387 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
26388 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
26389 not return anything.
26390 (rename_ssa_copies): Do not remove unused locals.
26391 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
26392 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
26393 * passes.c (execute_function_todo): Do not schedule unused locals
26394 removal if cleanup_tree_cfg did something.
26395 * tree-ssa-live.c (remove_unused_locals): Dump statistics
26396 about the number of removed locals.
26398 2013-02-20 Richard Biener <rguenther@suse.de>
26400 PR tree-optimization/56398
26401 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
26403 2013-02-20 Martin Jambor <mjambor@suse.cz>
26405 PR tree-optimization/55334
26406 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
26407 restricted pointers to arrays.
26409 2013-02-20 Richard Biener <rguenther@suse.de>
26410 Jakub Jelinek <jakub@redhat.com>
26412 PR tree-optimization/56396
26413 * tree-ssa-ccp.c (n_const_val): New static variable.
26414 (get_value): Return NULL for SSA names we don't have a lattice
26416 (ccp_initialize): Initialize n_const_val.
26417 * tree-ssa-copy.c (n_copy_of): New static variable.
26418 (init_copy_prop): Initialize n_copy_of.
26419 (get_value): Return NULL_TREE for SSA names we don't have a
26422 2013-02-20 Martin Jambor <mjambor@suse.cz>
26424 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
26426 2013-02-20 Richard Biener <rguenther@suse.de>
26428 * genpreds.c (write_lookup_constraint): Do not compare first
26429 letter of the constraint again.
26431 2013-02-20 Richard Biener <rguenther@suse.de>
26433 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
26435 (get_use_iv_cost): Terminate hashtable walk when coming across
26438 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
26440 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
26441 reassociation for avx2 targets.
26443 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26445 * config/microblaze/microblaze.c: microblaze_has_clz = 0
26446 Add version check for v8.10.a to enable microblaze_has_clz
26447 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
26448 version and TARGET_PATTERN_COMPARE check
26449 * config/microblaze/microblaze.md: New clzsi2 instruction
26451 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26453 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
26454 function before branching.
26456 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
26458 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
26460 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
26462 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
26464 PR middle-end/55889
26465 * sel-sched.c: Include ira.h.
26466 (implicit_clobber_conflict_p): New function.
26467 (moveup_expr): Use it.
26468 * Makefile.in (sel-sched.o): Depend on ira.h.
26470 2013-02-19 Richard Biener <rguenther@suse.de>
26472 PR tree-optimization/56384
26473 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
26474 (vn_hash_type): Split out from ...
26475 (vn_hash_constant_with_type): ... here.
26476 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
26477 (vn_phi_eq): Compare types from vn_phi_s structure.
26478 (vn_phi_lookup): Populate vn_phi_s type.
26479 (vn_phi_insert): Likewise.
26481 2013-02-19 Jakub Jelinek <jakub@redhat.com>
26483 PR tree-optimization/56350
26484 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
26485 if haven't found reduction or nested cycle operand, rather than
26486 asserting we must find it.
26488 PR tree-optimization/56381
26489 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
26492 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
26493 Jakub Jelinek <jakub@redhat.com>
26496 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
26497 (swap_optab_enable): Same.
26498 (init_all_optabs): Use argument instead of global.
26499 * tree.h (struct tree_optimization_option): New field target_optabs.
26500 * expr.h (init_all_optabs): Add argument to prototype.
26501 (TREE_OPTIMIZATION_OPTABS): New.
26502 (save_optabs_if_changed): Protoize.
26503 * optabs.h: Declare this_fn_optabs.
26504 * optabs.c (save_optabs_if_changed): New.
26505 Declare this_fn_optabs.
26506 (init_optabs): Add argument to init_all_optabs() call.
26507 * function.c (invoke_set_current_function_hook): Handle per
26509 * function.h (struct function): New field optabs.
26510 * config/mips/mips.c (mips_set_mips16_mode): Handle when
26511 optimization_current_node has changed.
26512 * target-globals.h (save_target_globals_default_opts): Protoize.
26513 * target-globals.c (save_target_globals_default_opts): New.
26515 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26518 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
26519 registers %fr12 and %fr12R as call used.
26522 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
26523 and HImode, require all displacements to be an integer multiple of
26525 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
26526 only allow QImode and HImode when reload is in progress and strict is
26527 true. Likewise for symbolic addresses. Use base14_operand to check
26528 displacements in REG+BASE addresses.
26530 2013-02-18 Richard Biener <rguenther@suse.de>
26532 PR tree-optimization/56366
26533 * tree-vect-loop.c (get_initial_def_for_induction): Properly
26534 handle sign-conversion of outer-loop initial induction value.
26536 2013-02-18 Richard Biener <rguenther@suse.de>
26538 PR middle-end/56349
26539 * cfghooks.c (merge_blocks): If we merge a latch into another
26540 block adjust references to it.
26541 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
26542 (verify_loop_structure): Verify that a recorded latch is in fact
26545 2013-02-18 Richard Biener <rguenther@suse.de>
26547 PR tree-optimization/56321
26548 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
26549 order SSA name release and virtual operand unlinking.
26551 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26553 * config/microblaze/microblaze.md (save_stack_block): Define.
26554 (restore_stack_block): Likewise.
26556 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26558 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
26559 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
26560 * config/microblaze/microblaze.c (microblaze_option_override):
26561 Bail out early for PIC modes when target does not support PIC.
26563 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26565 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
26566 Replace with a microblaze version.
26567 (microblaze_trampoline_init): Adapt for microblaze.
26568 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
26571 2013-02-16 Jakub Jelinek <jakub@redhat.com>
26572 Dodji Seketeli <dodji@redhat.com>
26575 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
26576 (instrument_mem_region_access): Do not forget to always put
26577 instrumentation of the of 'base' and 'base + len' in a "if (len !=
26578 0) statement, even for cases where either 'base' or 'base + len'
26579 are not instrumented -- because they have been previously
26580 instrumented. Simplify the logic by putting all the statements
26581 instrument 'base + len' inside a sequence, and then insert that
26582 sequence right before the current insertion point. Then, to
26583 instrument 'base + len', just get an iterator on that statement.
26584 And do not forget to update the pointer to iterator the function
26585 received as argument.
26587 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
26589 PR rtl-optimization/56348
26590 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
26592 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
26594 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
26595 (clean_graph_dump_file): Pass base to start_graph_dump.
26597 2013-02-14 Richard Henderson <rth@redhat.com>
26600 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
26602 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
26604 * collect2-aix.h: Define F_LOADONLY.
26606 2013-02-14 Richard Biener <rguenther@suse.de>
26609 * varasm.c (output_constant_def_1): Get the decl representing
26610 the constant as argument.
26611 (output_constant_def): Wrap output_constant_def_1.
26612 (make_decl_rtl): Use output_constant_def_1 with the decl
26613 representing the constant.
26614 (build_constant_desc): Optionally re-use a decl already
26615 representing the constant.
26616 (tree_output_constant_def): Adjust.
26618 2013-02-14 Dodji Seketeli <dodji@redhat.com>
26621 * asan.c (instrument_builtin_call): Really put the length of the
26622 second source argument into src1_len.
26624 2013-02-13 Jakub Jelinek <jakub@redhat.com>
26626 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
26627 argument. If it is false, don't create edge from then_bb to
26629 (insert_if_then_before_iter): Pass true to it.
26630 (build_check_stmt): Pass false to it.
26631 (transform_statements): Flush hash table only on extended basic
26632 block boundaries, rather than at the beginning of every bb.
26633 Don't flush hash table on nonfreeing_call_p calls.
26634 * tree-flow.h (nonfreeing_call_p): New prototype.
26635 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
26637 2013-02-13 David S. Miller <davem@davemloft.net>
26639 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
26641 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
26644 * ira.c (max_regno_before_ira): Move from ...
26646 (fix_reg_equiv_init): Use max_regno_before_ira instead of
26649 2013-02-13 Jakub Jelinek <jakub@redhat.com>
26651 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
26653 2013-02-13 Richard Biener <rguenther@suse.de>
26656 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
26657 globals in MEM_REFs.
26659 2013-02-13 Richard Biener <rguenther@suse.de>
26661 * loop-init.c (loop_optimizer_init): Clear loop state when
26662 re-initializing preserved loops.
26663 * loop-unswitch.c (unswitch_single_loop): Return whether
26664 we unswitched the loop. Do not verify loop state here.
26665 (unswitch_loops): When we unswitched a loop discover new loops.
26667 2013-02-13 Kostya Serebryany <kcc@google.com>
26669 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
26671 * sanitizer.def: Rename __asan_init to __asan_init_v1.
26673 2013-02-12 Dodji Seketeli <dodji@redhat.com>
26675 Avoid instrumenting duplicated memory access in the same basic block
26676 * Makefile.in (asan.o): Add new dependency on hash-table.h
26677 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
26678 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
26679 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
26680 (free_mem_ref_resources, has_mem_ref_been_instrumented)
26681 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
26682 (get_mem_ref_of_assignment): New functions.
26683 (get_mem_refs_of_builtin_call): Extract from
26684 instrument_builtin_call and tweak a little bit to make it fit with
26686 (instrument_builtin_call): Use the new
26687 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
26688 of is_gimple_builtin_call.
26689 (instrument_derefs, instrument_mem_region_access): Insert the
26690 instrumented memory reference into the hash table.
26691 (maybe_instrument_assignment): Renamed instrument_assignment into
26692 this, and change it to advance the iterator when instrumentation
26693 actually happened and return true in that case. This makes it
26694 homogeneous with maybe_instrument_assignment, and thus give a
26695 chance to callers to be more 'regular'.
26696 (transform_statements): Clear the memory reference hash table
26697 whenever we enter a new BB, when we cross a function call, or when
26698 we are done transforming statements. Use
26699 maybe_instrument_assignment instead of instrumentation. No more
26700 need to special case maybe_instrument_assignment and advance the
26701 iterator after calling it; it's now handled just like
26702 maybe_instrument_call. Update comment.
26704 2013-02-13 Richard Biener <rguenther@suse.de>
26706 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
26707 Fix loop discovery code.
26709 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
26711 PR inline-asm/56148
26712 * lra-constraints.c (process_alt_operands): Match early clobber
26713 operand with itself. Check conflicts with earlyclobber only if
26714 the operand is not reloaded. Prefer to reload conflicting operand
26715 if earlyclobber and matching operands are the same.
26717 2013-02-12 Richard Biener <rguenther@suse.de>
26720 * lto-streamer-out.c (write_symbol): Do not output symbols
26721 for hard register variables.
26723 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
26726 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
26727 (umulsidi3_insn, mulsidi3_insn): New insns.
26729 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
26731 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
26732 (struct tune_params): Add vec_costs field.
26733 * config/arm/arm.c (arm_builtin_vectorization_cost)
26734 (arm_add_stmt_cost): New functions.
26735 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
26736 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
26737 (arm_default_vec_cost): New struct of type cpu_vec_costs.
26738 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
26739 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
26740 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
26741 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
26743 2013-02-12 Richard Biener <rguenther@suse.de>
26746 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
26747 decls again if possible.
26749 2013-02-12 Richard Biener <rguenther@suse.de>
26751 PR middle-end/56288
26752 * tree-ssa.c (verify_ssa_name): Fix check, move
26753 SSA_NAME_IN_FREE_LIST check up.
26755 2013-02-12 Jakub Jelinek <jakub@redhat.com>
26756 Steven Bosscher <steven@gcc.gnu.org>
26758 PR rtl-optimization/56151
26759 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
26760 equal to op0 or op1, and last_insn pattern is CODE operation
26761 with MEM dest and one of the operands matches that MEM.
26763 2013-02-11 Sriraman Tallam <tmsriram@google.com>
26765 * doc/extend.texi: Document Function Multiversioning and "default"
26766 parameter string to target attribute.
26767 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
26768 target attribute parameter is "default".
26769 (ix86_compare_version_priority): Remove checks for target attribute.
26770 (ix86_mangle_function_version_assembler_name): Change error to sorry.
26771 Remove check for target attribute equal to NULL. Add assert.
26772 (ix86_generate_version_dispatcher_body): Change error to sorry.
26774 2013-02-11 Iain Sandoe <iain@codesourcery.com>
26775 Jack Howarth <howarth@bromo.med.uc.edu>
26776 Patrick Marlier <patrick.marlier@gmail.com>
26779 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
26780 define ENDFILE_SPEC as TM_DESTRUCTOR.
26781 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
26783 2013-02-11 Alexander Potapenko <glider@google.com>
26784 Jack Howarth <howarth@bromo.med.uc.edu>
26785 Jakub Jelinek <jakub@redhat.com>
26788 * config/darwin.c (cdtor_record): Rename ctor_record.
26789 (sort_cdtor_records): Rename sort_ctor_records.
26790 (finalize_dtors): New routine to sort destructors by
26791 priority before use in assemble_integer.
26792 (machopic_asm_out_destructor): Use finalize_dtors if needed.
26794 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
26796 PR rtl-optimization/56275
26797 * simplify-rtx.c (avoid_constant_pool_reference): Check that
26798 offset is non-negative and less than cmode size before
26799 calling simplify_subreg.
26801 2013-02-11 Richard Biener <rguenther@suse.de>
26803 PR tree-optimization/56264
26804 * cfgloop.h (fix_loop_structure): Adjust prototype.
26805 * loop-init.c (fix_loop_structure): Return the number of
26806 newly discovered loops.
26807 * tree-cfgcleanup.c (repair_loop_structures): When new loops
26808 are discovered, do a full loop-closed SSA rewrite.
26810 2013-02-11 Richard Biener <rguenther@suse.de>
26812 PR tree-optimization/56273
26813 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
26815 (check_array_ref): Fix missing newline in dumps.
26816 (search_for_addr_array): Likewise.
26818 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
26820 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
26822 2013-02-09 Jakub Jelinek <jakub@redhat.com>
26825 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
26827 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
26829 PR rtl-optimization/56246
26830 * lra-constraints.c (simplify_operand_subreg): Try to reuse
26832 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
26833 constraints are satisfied.
26835 2013-02-08 Jeff Law <law@redhat.com>
26838 * emit-rtl.c (reg_is_parm_p): New function.
26839 * regs.h (reg_is_parm_p): New prototype.
26840 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
26841 callee-clobbered registers.
26843 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
26846 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
26847 If there is no implicit builtin declaration, just return NULL.
26849 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
26851 * config/i386/sse.md (FMAMODEM): New mode iterator.
26852 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
26853 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
26855 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
26857 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
26858 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
26859 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
26861 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26863 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
26864 (microblaze*-*-elf): Likewise.
26865 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
26867 * config/microblaze/microblaze-c.c: Add builtin defines for
26868 _LITTLE_ENDIAN and _BIG_ENDIAN.
26869 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
26870 add to TARGET_DEFAULT flags.
26871 Expand ASM_SPEC and LINK_SPEC.
26872 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
26873 * config/microblaze/microblaze.md: Update extendsidi2 and
26874 movdi_internal instructions to use low-order / high-order reg
26876 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
26877 options and inversemask / mask of LITTLE_ENDIAN.
26878 * config/microblaze/t-microblaze: Expand multilib options to
26879 include mlittle-endian (le) and update exceptions patterns.
26881 2013-02-08 Jakub Jelinek <jakub@redhat.com>
26883 PR rtl-optimization/56195
26884 * lra-constraints.c (get_reload_reg): Don't reuse regs
26885 if they have smaller mode than requested, if they have
26886 wider mode than requested, try to return a SUBREG.
26888 PR tree-optimization/56250
26889 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
26890 if type is unsigned and code isn't MULT_EXPR.
26892 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
26894 PR tree-optimization/56064
26895 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
26896 bits according to mode.
26897 * fixed-value.h (fixed_from_double_int)
26898 (const_fixed_from_double_int): Adjust comments.
26900 2013-02-08 Richard Biener <rguenther@suse.de>
26903 * lto-streamer.h (struct data_in): Remove current_file, current_line
26904 and current_col members.
26905 * lto-streamer-out.c (lto_output_location): Stream changed bits
26906 en-block for efficiency.
26907 * lto-streamer-in.c (clear_line_info): Remove.
26908 (lto_input_location): Cache current file, line and column
26909 globally via local statics. Read changed bits en-block.
26910 (input_function): Do not call clear_line_info.
26911 (lto_read_body): Likewise.
26912 (lto_input_toplevel_asms): Likewise.
26914 2013-02-08 Michael Matz <matz@suse.de>
26916 PR tree-optimization/52448
26917 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
26918 (nt_call_phase): New static.
26919 (add_or_mark_expr): Only mark accesses with newer phase than any
26921 (nonfreeing_call_p): New.
26922 (nt_init_block): Update nt_call_phase, mark blocks as visited.
26923 (nt_fini_block): Keep blocks marked as visited.
26924 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
26926 2013-02-08 Richard Biener <rguenther@suse.de>
26928 * ira.c (ira): Free broken dominator information.
26930 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
26932 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
26934 2013-02-08 Marek Polacek <polacek@redhat.com>
26936 * cfgloop.c (verify_loop_structure): Add more checking of headers.
26938 2013-02-08 Richard Biener <rguenther@suse.de>
26940 PR middle-end/56181
26941 * cfgloop.h (flow_loops_find): Adjust.
26942 (bb_loop_header_p): Declare.
26943 * cfgloop.c (bb_loop_header_p): New function split out from ...
26944 (flow_loops_find): ... here. Adjust function signature,
26945 support incremental loop structure update.
26946 (verify_loop_structure): Cleanup. Verify a loop is a loop.
26947 * cfgloopmanip.c (fix_loop_structure): Move ...
26948 * loop-init.c (fix_loop_structure): ... here.
26949 (apply_loop_flags): Split out from ...
26950 (loop_optimizer_init): ... here.
26951 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
26952 in incremental mode, only remove dead loops here.
26954 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
26957 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
26958 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
26959 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
26960 (*round<mode>3.libgcc): New insns for fixed-modes.
26961 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
26962 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
26963 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
26964 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
26965 implementations. Define to __builtin_avr_absFX,
26966 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
26967 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
26968 __builtin_avr_countlsFX, respectively.
26969 * config/avr/avr-c.c (target.h): Include it.
26970 (enum avr_builtin_id): New enum.
26971 (avr_resolve_overloaded_builtin): New static function.
26972 (avr_register_target_pragmas): Use it to set
26973 targetm.resolve_overloaded_builtin.
26974 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
26975 tree nodes used by DEF_BUILTIN.
26976 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
26977 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
26978 <AVR_BUILTIN_xxBITS>: Same.
26980 2013-02-08 Richard Biener <rguenther@suse.de>
26982 * cfgloop.c (verify_loop_structure): Properly handle
26983 a loop exiting to another loop header.
26984 * ira-int.h (ira_loops): Remove.
26985 * ira.c (ira_loops): Remove.
26986 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
26987 (do_reload): Use loop_optimizer_finalize.
26988 * ira-build.c (create_loop_tree_nodes): Use get_loops and
26989 number_of_loops to access the loop tree.
26990 (more_one_region_p): Likewise.
26991 (finish_loop_tree_nodes): Likewise.
26992 (rebuild_regno_allocno_maps): Likewise.
26993 (mark_loops_for_removal): Likewise.
26994 (mark_all_loops_for_removal): Likewise.
26995 (remove_unnecessary_regions): Likewise.
26996 (ira_build): Likewise.
26997 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
26999 2013-02-08 Richard Biener <rguenther@suse.de>
27001 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27002 * ipa-pure-const.c (analyze_function): Avoid calling
27003 mark_irreducible_loops twice.
27004 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27006 2013-02-07 David S. Miller <davem@davemloft.net>
27008 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27010 * var-tracking.c (vt_add_function_parameter): Test the presence of
27011 HAVE_window_save properly and do not remap argument registers when
27012 we have a leaf function.
27014 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27017 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27019 * config/i386/i386.c (ix86_print_operand): Ditto.
27021 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27023 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27025 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27027 PR rtl-optimization/56225
27028 * lra-constraints.c (process_alt_operands): Check that reload hard
27029 reg can hold value for strict_low_part.
27031 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27034 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27035 dwarf2out_end_function.
27036 (in_first_function_p, maybe_at_text_label_p,
27037 first_loclabel_num_not_at_text_label): New variables.
27038 (dwarf2out_var_location): In the first function find out
27039 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27040 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27043 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27045 PR rtl-optimization/56178
27046 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27047 SUBREG of a register. Tidy up related block of code.
27048 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27049 note if the source is a register or a SUBREG of a register.
27051 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27054 * config/rs6000/rs6000.md (ptrm): New mode attr.
27055 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27056 call_value_indirect_aix<pttrsize>,
27057 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27060 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27062 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27063 if -bnortl. Convert to strcmp and strncmp.
27065 2013-02-07 Alan Modra <amodra@gmail.com>
27068 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27069 addresses won't wrap when offsetting.
27070 (rs6000_secondary_reload): Provide secondary reloads needed for
27071 wrapping LO_SUM addresses.
27073 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27075 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27076 MACH, just __MACH__.
27078 2013-02-06 Richard Biener <rguenther@suse.de>
27080 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27081 instead of calling fix_loop_structure.
27083 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27085 PR middle-end/56217
27086 * omp-low.c (use_pointer_for_field): Return false if
27087 lower_send_shared_vars doesn't generate any copy-out code.
27089 2013-02-06 Tom de Vries <tom@codesourcery.com>
27091 PR rtl-optimization/56131
27092 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27093 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27094 of the label is NULL. Add comment.
27096 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27098 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27101 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27102 (STATIC_LIBTSAN_LIBS): Likewise.
27103 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27104 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27105 is defined, don't add anything else beyond that.
27106 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27107 (LINK_COMMAND_SPEC): Use them.
27109 PR tree-optimization/56205
27110 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27111 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27112 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27114 2013-02-05 Richard Biener <rguenther@suse.de>
27116 PR tree-optimization/53342
27117 PR tree-optimization/53185
27118 * tree-vectorizer.h (vect_check_strided_load): Remove.
27119 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27120 not disallow peeling for vectorized strided loads.
27121 (vect_check_strided_load): Make static and simplify.
27122 (vect_analyze_data_refs): Adjust.
27123 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27124 correctly when vectorizing strided loads.
27126 2013-02-05 Richard Biener <rguenther@suse.de>
27128 * doc/install.texi: Refer to ISL, not PPL.
27130 2013-02-05 Jan Hubicka <jh@suse.cz>
27132 PR tree-optimization/55789
27133 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27135 2013-02-05 Jan Hubicka <jh@suse.cz>
27137 PR tree-optimization/55789
27138 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27139 the dead call anyway.
27141 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27144 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27146 2013-02-04 Alexander Potapenko <glider@google.com>
27147 Jack Howarth <howarth@bromo.med.uc.edu>
27148 Jakub Jelinek <jakub@redhat.com>
27151 * config/darwin.c (sort_ctor_records): Stabilized qsort
27152 on constructor priority by using original position.
27153 (finalize_ctors): New routine to sort constructors by
27154 priority before use in assemble_integer.
27155 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27157 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27160 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27161 about visibility on artificial decls.
27162 * config/sol2.c (solaris_assemble_visibility): Likewise.
27164 2013-02-04 Kai Tietz <ktietz@redhat.com>
27167 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27168 argument and improve checking of return-argument types for 16-byte
27170 (ix86_function_value_1): Add additional valtype argument on call
27171 of function_value_64.
27172 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27173 handling infunction_value_64 function.
27175 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27177 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27179 2013-02-04 Richard Biener <rguenther@suse.de>
27181 PR tree-optimization/56188
27182 * tree-ssa-structalias.c (label_visit): Consider case with
27183 initially non-empty points-to set.
27184 (perform_var_substitution): Dump node mapping and clean up.
27186 2013-02-04 Richard Guenther <rguenther@suse.de>
27189 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
27190 node prevail as last resort.
27191 (lto_symtab_merge_decls): Remove guard on LTRANS here.
27192 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
27194 2013-02-04 Richard Biener <rguenther@suse.de>
27196 PR tree-optimization/56113
27197 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
27199 (equiv_class_lookup_or_add): ... this.
27200 (label_visit): Adjust and fix error in previous patch.
27201 (perform_var_substitution): Adjust.
27203 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
27205 * config/sh/divtab.c: Fix formatting and comments throughout the file.
27206 * config/sh/sh4-300.md: Likewise.
27207 * config/sh/sh4a.md: Likewise.
27208 * config/sh/constraints.md: Likewise.
27209 * config/sh/sh.md: Likewise.
27210 * config/sh/netbsd-elf.h: Likewise.
27211 * config/sh/predicates.md: Likewise.
27212 * config/sh/sh-protos.h: Likewise.
27213 * config/sh/ushmedia.h: Likewise.
27214 * config/sh/linux.h: Likewise.
27215 * config/sh/sh.c: Likewise.
27216 * config/sh/superh.h: Likewise.
27217 * config/sh/elf.h: Likewise.
27218 * config/sh/sh4.md: Likewise.
27219 * config/sh/sh.h: Likewise.
27221 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27223 * config/pa/constraints.md: Adjust unused letters. Change "T"
27224 constraint to match_test floating_point_store_memory_operand().
27225 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
27226 (base14_operand): New.
27227 (floating_point_store_memory_operand): New.
27228 (integer_store_memory_operand): Revise to use base14_operand and
27229 reg_plus_base_memory_operand.
27230 (move_dest_operand): Allow symbolic_memory_operands.
27231 (symbolic_memory_operand): Check for LO_SOM.
27232 (symbolic_operand): Change default case to break.
27233 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
27234 CONST_DOUBLE values to be reloaded by putting them into memory when
27235 the destination is a floating point register.
27236 (movdf): Remove code to handle CONST_DOUBLE.
27238 (reload_indf_r1): New.
27239 (reload_insf_r1): New.
27240 Consistently use "Q" and "T" constraints with integer and floating
27241 point move instructions, respectively.
27242 (movdi): Remove FAIL.
27243 Change predicate for source operand unamed DImode move from
27244 general_operand to move_src_operand.
27245 (umulsidi3): Change predicate for destination operand to
27247 Likewise for similar unamed patterns.
27248 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
27249 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
27250 (hppa_legitimize_address): Simplify mask calculation.
27251 (pa_emit_move_sequence): Revised handling of secondary reloads from
27252 REG+D addresses for floating point loads and stores. Directly handle
27253 loading CONST0_RTX (mode) to a floating point register.
27254 (pa_secondary_reload): Handle reloading DF and SFmode constant values
27255 to floating point registers. Don't restrict secondary reloads to
27256 floating point registers to integer modes. Revise some comments and
27258 (TARGET_LEGITIMATE_ADDRESS_P): Define.
27259 (pa_legitimate_address_p): New.
27260 (pa_legitimize_reload_address): New.
27261 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
27262 (STRICT_REG_OK_FOR_BASE_P): New.
27263 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
27264 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
27266 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
27267 Andrew Dixie <andrewd@gentrack.com>
27269 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
27272 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
27274 * expmed.c (extract_bit_field_1): Pass the full width of the
27275 structure to get_best_reg_extraction_insn.
27277 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27280 * configure.ac (use_cxa_atexit): Add AIX.
27281 * configure: Regenerate.
27283 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
27285 2013-02-01 Jakub Jelinek <jakub@redhat.com>
27288 * final.c (need_profile_function): New variable.
27289 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
27290 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
27291 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
27292 notes, targetm.asm_out.function_prologue doesn't emit anything,
27293 HAVE_prologue and profiler should be emitted before prologue,
27294 set need_profile_function instead of emitting it.
27295 (final_scan_insn): If need_profile_function, emit
27296 profile_function on the first NOTE_INSN_BASIC_BLOCK or
27297 NOTE_INSN_FUNCTION_BEG note.
27299 2013-02-01 Richard Henderson <rth@redhat.com>
27301 * config/rs6000/rs6000.md (smulditi3): New.
27304 * config/alpha/alpha.md (umulditi3): New.
27306 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27308 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
27309 (ASM_OUTPUT_ALIGNED_LOCAL): New.
27311 2013-02-01 Richard Biener <rguenther@suse.de>
27313 PR tree-optimization/56113
27314 * tree-ssa-structalias.c (label_visit): Reduce work for
27315 single-predecessor nodes.
27317 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
27319 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
27320 range isn't testing for zero.
27322 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
27324 PR middle-end/56113
27325 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
27327 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
27328 Nick Clifton <nickc@redhat.com>
27330 * config/v850/constraints.md (Q): Define as a memory constraint.
27331 * config/v850/predicates.md (label_ref_operand): New predicate.
27332 (e3v5_shift_operand): New predicate.
27333 (ior_operator): New predicate.
27334 * config/v850/t-v850: Add e3v5 multilib.
27335 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
27336 (v850_gen_movdi): Prototype.
27337 * config/v850/v850.c: Add support for e3v5 architecture.
27338 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
27340 (construct_save_jarl): Add e3v5 long JARL support.
27341 (v850_adjust_insn_length): New function. Adjust length of call
27342 insns when using e3v5 instructions.
27343 (v850_gen_movdi): New function: Generate instructions to move a
27345 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
27346 (CPP_SPEC): Define __v850e3v5__ as appropriate.
27347 (TARGET_USE_FPU): Enable for e3v5.
27348 (CONST_OK_FOR_W): New macro.
27349 (ADJUST_INSN_LENGTH): Define.
27350 * config/v850/v850.md (UNSPEC_LOOP): Define.
27351 (attr cpu): Add v850e3v5.
27352 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
27353 (movdi): New pattern.
27354 (movdi_internal): New pattern.
27355 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
27356 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
27357 (cstoresf4): Likewise.
27358 (cstoredf4): Likewise.
27359 (insv): New pattern.
27360 (rotlso3_a): New pattern.
27361 (rotlsi3_b): New pattern
27362 (rotlsi3_v850e3v5): New pattern.
27363 (doloop_begin): New pattern.
27364 (fix_loop_counter): New pattern.
27365 (doloop_end): New pattern.
27366 (branch_normal): Add e3v5 long branch support.
27367 (branch_invert): Likewise.
27368 (branch_z_normal): Likewise.
27369 (branch_z_invert): Likewise.
27370 (branch_nz_normal): Likewise.
27371 (branch_nz_invert): Likewise.
27372 (call_internal_short): Add e3v5 register-indirect JARL support.
27373 (call_internal_long): Likewise.
27374 (call_value_internal_short): Likewise.
27375 (call_value_internal_long): Likewise.
27376 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
27377 (mloop): New option.
27378 * config.gcc: Add support for configuring v840e3v5 target.
27379 * doc/invoke.texi: Document new v850 specific command line options.
27381 2013-01-31 Paul Koning <ni1d@arrl.net>
27385 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
27386 children if parent is a class.
27387 (prune_unused_types_prune): Don't add DW_AT_declaration.
27389 2013-01-31 Richard Biener <rguenther@suse.de>
27391 PR tree-optimization/56157
27392 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
27393 match up operand with SLP child.
27395 2013-01-31 Jason Merrill <jason@redhat.com>
27398 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
27399 parameters the first time.
27400 (gen_scheduled_generic_parms_dies): Check completeness here.
27402 2013-01-31 Richard Biener <rguenther@suse.de>
27404 PR middle-end/53073
27405 * common.opt (faggressive-loop-optimizations): New flag,
27406 enabled by default.
27407 * doc/invoke.texi (faggressive-loop-optimizations): Document.
27408 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
27409 infer_loop_bounds_from_undefined by it.
27411 2013-01-31 Richard Biener <rguenther@suse.de>
27413 PR tree-optimization/56150
27414 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
27415 visit virtual operands.
27416 (find_uses_to_rename_bb): Likewise.
27418 2013-01-31 Richard Biener <rguenther@suse.de>
27420 PR tree-optimization/56150
27421 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
27422 mixed store non-store stmts.
27424 2013-01-30 Jakub Jelinek <jakub@redhat.com>
27427 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
27428 LIBASAN_EARLY_SPEC is defined.
27429 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
27430 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
27432 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
27435 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
27436 invalid args instead of ICEing on it.
27437 (ix86_valid_target_attribute_tree): Return error_mark_node if
27438 ix86_valid_target_attribute_inner_p failed.
27439 (ix86_valid_target_attribute_p): Return false only if
27440 ix86_valid_target_attribute_tree returned error_mark_node. Allow
27441 target("default") attribute.
27442 (sorted_attr_string): Change argument from const char * to tree,
27443 merge in all target attribute arguments rather than just one.
27444 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
27445 instead of free. Avoid using strcat.
27446 (ix86_mangle_function_version_assembler_name): Mangle
27447 target("default") as if no target attribute is present. Adjust
27448 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
27449 instead of xmalloc and XDELETEVEC instead of free.
27450 (ix86_function_versions): Don't return true if one of the decls
27451 doesn't have target attribute. If they don't and one of the decls
27452 is DECL_FUNCTION_VERSIONED, report an error. Adjust
27453 sorted_attr_string caller. Use XDELETEVEC instead of free.
27454 (ix86_supports_function_versions): Remove.
27455 (make_name): Fix up formatting.
27456 (make_dispatcher_decl): Remove resolver_name and its initialization.
27457 Avoid leaking memory.
27458 (is_function_default_version): Return true if there is
27459 target("default") attribute rather than no target attribute at all.
27460 (make_resolver_func): Avoid leaking memory.
27461 (ix86_generate_version_dispatcher_body): Likewise.
27462 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
27463 * target.def (supports_function_versions): Remove.
27464 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
27465 * doc/tm.texi: Regenerated.
27467 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
27469 PR rtl-optimization/56144
27470 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
27471 for values with side effects.
27473 2013-01-30 Richard Biener <rguenther@suse.de>
27475 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
27476 (sparseset_pop): Likewise.
27477 * cfganal.c (compute_idf): Likewise. Increase work-stack size
27478 to be able to use quick_push in the worker loop.
27480 2013-01-30 Marek Polacek <polacek@redhat.com>
27482 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
27484 2013-01-30 Richard Biener <rguenther@suse.de>
27487 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
27489 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
27491 PR tree-optimization/56064
27492 * fixed-value.c (fixed_from_double_int): New function.
27493 * fixed-value.h (fixed_from_double_int): New prototype.
27494 (const_fixed_from_double_int): New static inline function.
27495 * fold-const.c (native_interpret_fixed): New static function.
27496 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
27497 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
27498 (native_encode_fixed): New static function.
27499 (native_encode_expr) <FIXED_CST>: Use it.
27500 (native_interpret_int): Move double_int worker code to...
27501 * double-int.c (double_int::from_buffer): ...this new static method.
27502 * double-int.h (double_int::from_buffer): Prototype it.
27504 2013-01-30 Richard Biener <rguenther@suse.de>
27506 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
27507 New pointer-map and obstack.
27508 (init_alias_vars): Allocate pointer-map and obstack.
27509 (delete_points_to_sets): Free them.
27510 (find_what_var_points_to): Cache result.
27511 (find_what_p_points_to): Adjust for changed interface of
27512 find_what_var_points_to.
27513 (compute_points_to_sets): Likewise.
27514 (ipa_pta_execute): Likewise.
27516 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
27518 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
27519 * configure: Regenerate.
27520 * config.in: Regenerate.
27521 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
27522 #nobits/#progbits if supported.
27524 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
27527 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
27528 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
27529 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
27531 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27533 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
27534 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
27536 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27538 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
27539 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
27541 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27543 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
27545 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
27546 * config/arm/cortex-a7.md: New bypasses using
27547 arm_mac_accumulator_is_result.
27549 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27551 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
27552 (cortex_a7_neon_mla): Likewise.
27553 (cortex_a7_fpfmad): New reservation.
27554 (cortex_a7_fpmacs): Use ffmas and update required units.
27555 (cortex_a7_fpmuld): Update required units and latency.
27556 (cortex_a7_fpmacd): Likewise.
27557 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
27558 (cortex_a7_neon). Likewise.
27559 (bypass) Update participating units.
27561 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27563 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
27564 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
27566 * config/arm/vfp11.md (vfp_farith): Use ffmas.
27567 (vfp_fmul): Use ffmad.
27568 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
27569 (cortex_r4_fmacd): Use ffmad.
27570 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
27571 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
27572 (cortex_a9_fmacd): Use ffmad.
27573 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
27574 (cortex_a8_vfp_macd): Use ffmad.
27575 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
27576 (cortex_a5_fpmacd): Use ffmad.
27577 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
27578 (cortex_a15_vfp_macd): Use ffmad.
27579 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
27581 2013-01-29 Jason Merrill <jason@redhat.com>
27584 * varasm.c (default_assemble_visibility): Don't warn about
27585 visibility on artificial decls.
27587 2013-01-29 Richard Biener <rguenther@suse.de>
27589 PR tree-optimization/56113
27590 * tree-ssa-structalias.c (equiv_class_lookup): Also return
27592 (label_visit): Free duplicate bitmaps and record the leader instead.
27593 (perform_var_substitution): Adjust.
27595 2013-01-29 Richard Biener <rguenther@suse.de>
27597 PR tree-optimization/55270
27598 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
27599 the CFG, schedule loops for fixup.
27601 2013-01-29 Nick Clifton <nickc@redhat.com>
27603 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
27606 2013-01-28 Leif Ekblad <leif@rdos.net>
27608 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
27609 * config/i386/i386.h (TARGET_RDOS): New macro.
27610 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
27611 * config/i386/i386.c (ix86_option_override_internal): For 64bit
27612 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
27613 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
27614 DEFAULT_LARGE_SECTION_THRESHOLD.
27615 * config/i386/i386.md (R14_REG, R15_REG): New constants.
27616 * config/i386/rdos.h: New file.
27617 * config/i386/rdos64.h: New file.
27619 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
27622 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
27625 2013-01-28 Jakub Jelinek <jakub@redhat.com>
27627 PR rtl-optimization/56117
27628 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
27629 call cselib_lookup_from_insn on the MEM before calling
27630 add_insn_mem_dependence.
27632 2013-01-28 Richard Biener <rguenther@suse.de>
27634 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
27635 to a stmt that didn't have one.
27636 (copy_phis_for_bb): Likewise for PHI arguments.
27637 (copy_debug_stmt): Likewise for debug stmts.
27639 2013-01-28 Richard Biener <rguenther@suse.de>
27641 PR tree-optimization/56034
27642 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
27643 (partition_builtin_p): Adjust.
27644 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
27645 it is the last partition.
27646 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
27647 up the vertex for the definition.
27648 (classify_partition): Classify whether a partition is a
27649 PKIND_REDUCTION, thus has uses outside of the loop.
27650 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
27651 Merge all PKIND_REDUCTION partitions into the last partition.
27652 (tree_loop_distribution): Seed partitions from reductions as well.
27654 2013-01-28 Jakub Jelinek <jakub@redhat.com>
27656 PR tree-optimization/56125
27657 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
27658 pow(x,c) into sqrt(x) * powi(x, n/2) or
27659 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
27660 optimizing for size.
27661 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
27662 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
27665 PR tree-optimization/56094
27666 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
27667 to UNKNOWN_LOCATION while gimplifying expr.
27669 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
27672 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
27673 operand 0 in movabs insn template for -masm=intel asm alternative.
27674 (*movabs<mode>_2): Ditto for operand 1.
27676 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
27679 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
27682 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27684 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
27687 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27689 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
27690 Undef to avoid warning.
27692 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27694 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
27695 * configure: Regenerate.
27697 2013-01-25 Jakub Jelinek <jakub@redhat.com>
27699 PR tree-optimization/56098
27700 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
27701 for stmts with volatile ops.
27702 (cond_store_replacement): Don't optimize if assign has volatile ops.
27703 (cond_if_else_store_replacement_1): Don't optimize if either
27704 then_assign or else_assign have volatile ops.
27705 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
27708 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
27710 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
27712 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
27714 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
27715 missing ':' in asm example.
27717 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
27719 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
27720 entries into lane and laneq entries.
27721 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
27722 Remove AdvSIMD scalar modes.
27723 (aarch64_sq<r>dmulh_laneq<mode>): New.
27724 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
27726 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
27727 builtin implementations to relfect changes in RTL in aarch64-simd.md.
27728 * config/aarch64/iterators.md (VCOND): New.
27731 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
27734 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
27735 Add NULL LIBNAME argument to existing definitions.
27736 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
27737 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
27738 * config/avr/avr.c (DEF_BUILTIN): Same.
27739 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
27740 (avr_expand_builtin): Expand to a vanilla call if a libgcc
27741 implementation is available (DECL_ASSEMBLER_NAME is set).
27742 (avr_fold_absfx): New static function.
27743 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
27744 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
27745 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
27746 AVR_BUILTIN_ABSLLK.
27747 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
27748 (abshk, absk, abslk, absllk): Provide as static inline functions.
27750 2013-01-25 Marek Polacek <polacek@redhat.com>
27752 PR tree-optimization/56035
27753 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
27755 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
27757 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
27758 (*movtf_internal_rex64): Add (!o,C) alternative
27759 (*movxf_internal_rex64): Ditto.
27760 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
27762 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
27764 * doc/invoke.texi: fix typo.
27765 * doc/objc.texi: fix typo.
27767 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
27769 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
27770 for the first two alternatives.
27772 2013-01-24 Diego Novillo <dnovillo@google.com>
27774 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
27775 (ggc-zone.o): Remove.
27776 * configure.ac: Remove option --with-gc.
27777 * configure: Re-generate.
27778 * doc/install.texi: Remove documentation for --with-gc.
27779 * gengtype.c (write_enum_defn): Remove. Update all users.
27780 (write_Types_process_field): Remove generation of gt_e_* argument.
27781 (output_type_enum): Remove. Update all users.
27782 (write_enum_defn): Remove. Update all users.
27783 (enum alloc_zone): Remove. Update all users.
27784 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
27785 * ggc-common.c (ggc_splay_alloc): Remove first argument.
27786 Update all callers.
27787 (struct ptr_data): Remove field TYPE. Update all users.
27788 (gt_pch_note_object): Remove argument TYPE. Update all users.
27789 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
27791 * ggc-none.c (ggc_alloc_typed_stat): Remove.
27792 (struct alloc_zone): Remove.
27793 (ggc_internal_alloc_zone_stat): Remove.
27794 (ggc_internal_cleared_alloc_zone_stat): Remove.
27795 * ggc-page.c (ggc_alloc_typed_stat): Remove.
27796 (ggc_pch_count_object): Remove last argument. Update all users.
27797 (ggc_pch_alloc_object): Remove last argument. Update all users.
27798 (struct alloc_zone): Remove.
27799 * ggc-zone.c: Remove.
27800 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
27801 (struct alloc_zone): Remove.
27802 (ggc_alloc_typed_stat): Remove.
27803 (ggc_alloc_typed): Remove.
27804 (ggc_splay_alloc): Remove first argument.
27805 (rtl_zone): Remove. Update all users.
27806 (tree_zone): Remove. Update all users.
27807 (tree_id_zone): Remove. Update all users.
27808 (ggc_internal_zone_alloc_stat): Remove. Update all users.
27809 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
27810 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
27811 * tree-ssanames.c: Remove references to zone allocator in comments.
27813 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
27815 * config/avr/avr.c (avr_out_fract): Make register numbers that
27816 might be outside of source operand signed.
27818 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
27820 * config/i386/constraints.md (Yf): New constraint.
27821 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
27822 of f constraint to conditionaly disable x87 register preferences.
27823 (*movdf_internal): Ditto.
27824 (*movsf_internal): Ditto.
27826 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
27828 PR inline-asm/55934
27829 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
27830 that have operands with impossible constraints.
27831 Add a FIXME for a speed-up opportunity.
27832 * lra-constraints.c (process_alt_operands): Verify that a class
27833 selected from constraints on asms is valid for the operand mode.
27834 (curr_insn_transform): Remove incorrect comment.
27836 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
27838 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
27839 TOC operand is a valid symbol ref in the constant pool.
27841 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27843 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
27845 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
27848 * config/avr/stdfix.h: New file.
27849 * t-avr (stdfix-gcc.h): New rule to build it.
27850 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
27852 2013-01-23 Kostya Serebryany <kcc@google.com>
27854 * config/darwin.h: remove dependency on
27855 CoreFoundation (asan on Mac OS).
27857 2013-01-23 Jakub Jelinek <jakub@redhat.com>
27860 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
27861 instead of cmpdi_operand for first comparison operand.
27862 Don't assert that comparison operands aren't both constants.
27864 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
27866 * doc/install.texi (Downloading the Source): Update references to
27867 downloading separate components.
27869 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
27871 * doc/extend.texi (__int128): Improve grammar.
27873 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
27876 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
27877 alternative to (o,r).
27878 (*movdi_internal_rex64): Remove (!o,n) alternative.
27879 (DImode immediate->memory splitter): Remove.
27880 (DImode immediate->memory peephole2): Remove.
27881 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
27882 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
27883 alternative to (!o,*r).
27884 (*movtf_internal_sse): New pattern.
27885 (*movxf_internal_rex64): New pattern.
27886 (*movxf_internal): Disable for TARGET_64BIT.
27887 (*movdf_internal_rex64): Remove (!o,F) alternative.
27889 2013-01-22 Jakub Jelinek <jakub@redhat.com>
27891 PR middle-end/56074
27892 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
27893 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
27894 * tree-vect-loop-manip.c (find_loop_location): Also ignore
27895 stmt locations where LOCATION_LOCUS of the stmt location is
27896 UNKNOWN_LOCATION or BUILTINS_LOCATION.
27899 * config/i386/i386.md (UNSPEC_STOS): New.
27900 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
27901 *strsetqi_1): Add UNSPEC_STOS.
27903 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
27906 * doc/invoke.texi: Remove left over -Wsynth example.
27908 2013-01-21 Jakub Jelinek <jakub@redhat.com>
27910 PR tree-optimization/56051
27911 * fold-const.c (fold_binary_loc): Don't fold
27912 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
27913 a narrowing conversion, or widening conversion from signed
27916 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
27918 PR rtl-optimization/56023
27919 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
27920 dependent on debug instruction.
27922 2013-01-21 Martin Jambor <mjambor@suse.cz>
27924 PR middle-end/56022
27925 * function.c (allocate_struct_function): Call
27926 invoke_set_current_function_hook earlier.
27928 2013-01-21 Jakub Jelinek <jakub@redhat.com>
27930 * reload1.c (init_reload): Only initialize reload_obstack
27931 during the first call.
27933 2013-01-21 Marek Polacek <polacek@redhat.com>
27935 * cfgloop.c (verify_loop_structure): Fix up grammar.
27937 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
27939 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
27940 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
27942 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27945 * config/arm/marvell-pj4.md: Update copyright year.
27946 Fix up use of alu to alu_reg and simple_alu_imm.
27948 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
27950 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
27952 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
27955 * lra-constraints.c (curr_insn_transform): Don't reuse original
27956 insn for secondary memory move when memory mode should be different.
27958 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27960 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
27961 atomic_storedi_1): New patterns.
27963 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27965 btver2 pipeline descriptions.
27966 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
27968 * config/i386/i386.md (btver2_decode): New type attributes.
27969 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
27971 * config/i386/btver2.md: New file describing btver2 pipelines.
27973 2013-01-19 Andrew Pinski <apinski@cavium.com>
27975 PR tree-optimization/52631
27976 * tree-ssa-sccvn (visit_use): Before looking up the original
27977 statement, try looking up the simplified expression.
27979 2013-01-19 Anthony Green <green@moxielogic.com>
27981 * config/moxie/moxie.c (moxie_expand_prologue): Set
27982 current_function_static_stack_size.
27984 2013-01-18 Jakub Jelinek <jakub@redhat.com>
27986 PR tree-optimization/56029
27987 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
27988 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
27990 2013-01-18 Sharad Singhai <singhai@google.com>
27992 PR tree-optimization/55995
27993 * dumpfile.c (dump_loc): Print location only if available.
27994 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
27996 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
27999 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28000 secondary memory move.
28001 (inherit_reload_reg): Use rclass instead of cl for
28002 check_secondary_memory_needed_p.
28004 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28006 PR middle-end/56015
28007 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28008 the case where writing real complex part of target modifies op1.
28010 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28012 * config/aarch64/aarch64-simd.md
28013 (aarch64_vcond_internal<mode>): Handle unordered cases.
28014 * config/aarch64/iterators.md (v_cmp_result): New.
28016 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28017 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28019 * config/arm/marvell-pj4.md: New file.
28020 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28021 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28022 (generic_vfp): Likewise.
28023 * config/arm/arm-cores.def: Add marvell-pj4.
28024 * config/arm/arm-tune.md: Regenerate.
28025 * config/arm/arm-tables.opt: Regenerate.
28026 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28027 * doc/invoke.texi: Document marvell-pj4.
28029 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28031 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28033 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28038 * var-tracking.c (negative_power_of_two_p): New.
28039 (global_get_addr_cache, local_get_addr_cache): New.
28040 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28041 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28043 (vt_stack_offset_p): Remove.
28044 (vt_canon_true_dep): Always canonicalize loc's address.
28045 (clobber_overlapping_mems): Make sure we have a MEM.
28046 (local_get_addr_clear_given_value): New.
28047 (val_reset): Clear local cached entries.
28048 (compute_bb_dataflow): Create and release the local cache.
28049 Disable duplicate MEMs clobbering.
28050 (emit_notes_in_bb): Clobber MEMs likewise.
28051 (vt_emit_notes): Create and release the local cache.
28052 (vt_initialize, vt_finalize): Create and release the global
28053 cache, respectively.
28054 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28056 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28058 PR libmudflap/53359
28059 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28060 not found in the symtab.
28062 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28065 PR rtl-optimization/55547
28066 PR rtl-optimization/53827
28069 * alias.c (offset_overlap_p): New, factored out of...
28070 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28071 the conservative special case for symbolic constants. Don't
28072 adjust zero sizes on alignment.
28074 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28076 PR rtl-optimization/52573
28077 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28078 REG_UNUSED for the same register.
28080 2013-01-17 Richard Biener <rguenther@suse.de>
28081 Marek Polacek <polacek@redhat.com>
28083 PR rtl-optimization/55833
28084 * loop-unswitch.c (unswitch_loops): Move loop verification...
28085 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28086 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28087 Set it to true when we're removing a loop from hierarchy tree in
28088 an irreducible region.
28089 (fix_bb_placements): Adjust caller.
28090 (fix_loop_placements): Likewise.
28092 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28094 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28095 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28096 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28097 Remove ID. Adjust comments.
28098 * config/avr/avr-c.c (avr_builtin_name): Remove.
28099 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28100 * config/avr/avr.c (avr_tolower): New static function.
28101 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28102 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28103 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28106 2013-01-17 Jan Hubicka <jh@suse.cz>
28108 PR tree-optimization/55273
28109 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28111 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28114 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28115 store through atomic_store<mode>_1.
28116 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28118 2013-01-17 Martin Jambor <mjambor@suse.cz>
28120 PR tree-optimizations/55264
28121 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28122 for virtual methods.
28123 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28124 virtual methods before inlining is over.
28125 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28127 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28130 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28132 PR rtl-optimization/56005
28133 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28134 pending reads for prefetch.
28136 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28138 * config/aarch64/aarch64.md
28139 (*cstoresi_neg_uxtw): New pattern.
28140 (*cmovsi_insn_uxtw): New pattern.
28141 (*<optab>si3_uxtw): New pattern.
28142 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28143 (*<optab>si3_insn_uxtw): New pattern.
28144 (*bswapsi2_uxtw): New pattern.
28146 2013-01-16 Richard Biener <rguenther@suse.de>
28148 * tree-inline.c (tree_function_versioning): Remove set but
28149 never used variable.
28151 2013-01-16 Richard Biener <rguenther@suse.de>
28153 PR tree-optimization/55964
28154 * tree-flow.h (rename_variables_in_loop): Remove.
28155 (rename_variables_in_bb): Likewise.
28156 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28157 (copy_loop_before): Adjust and delete update-ssa status.
28158 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28159 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28160 (rename_variables_in_loop): Remove.
28161 (slpeel_update_phis_for_duplicate_loop): Likewise.
28162 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28163 use available cfg machinery instead of duplicating it.
28164 Update PHI nodes and perform poor-mans SSA update here.
28165 (slpeel_tree_peel_loop_to_edge): Adjust.
28167 2013-01-16 Richard Biener <rguenther@suse.de>
28169 PR tree-optimization/54767
28170 PR tree-optimization/53465
28171 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28172 (vrp_visit_phi_node): For PHI arguments coming via backedges
28173 drop all symbolical range information.
28174 (execute_vrp): Compute backedges.
28176 2013-01-16 Richard Biener <rguenther@suse.de>
28178 * doc/install.texi: Update CLooG and ISL requirements to
28181 2013-01-16 Christian Bruel <christian.bruel@st.com>
28184 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
28185 (broken_move): Handle UNSPECV_SP_SWITCH_B.
28186 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
28188 2013-01-16 DJ Delorie <dj@redhat.com>
28190 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
28191 (UNSPECV_SP_SWITCH_E): New.
28192 (sp_switch_1): Change to an unspec.
28193 (sp_switch_2): Change to an unspec. Don't use post-inc when we
28196 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
28198 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
28199 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
28200 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
28201 (expand_mem_thread_fence): Ditto.
28202 (expand_mem_signal_fence): Ditto.
28203 (expand_atomic_load): Ditto.
28204 (expand_atomic_store): Ditto.
28206 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
28208 PR rtl-optimization/55547
28209 PR rtl-optimization/53827
28212 * alias.c (memrefs_conflict_p): Set sizes to negative after
28215 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28218 * function.c (thread_prologue_and_epilogue_insns): Always
28219 add crtl->drap_reg to set_up_by_prologue.set, even if
28220 stack_realign_drap is false.
28222 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
28224 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
28225 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
28226 *call): Fix indention.
28228 2013-01-15 Tom de Vries <tom@codesourcery.com>
28231 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
28234 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
28236 PR rtl-optimization/55153
28237 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
28239 2013-01-15 Martin Jambor <mjambor@suse.cz>
28241 PR tree-optimization/55920
28242 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
28243 accesses as grp_to_be_debug_replaced.
28245 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28247 PR tree-optimization/55920
28248 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
28249 there is non-useless type conversion needed from debug rhs to lhs,
28250 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
28252 2013-01-15 Joseph Myers <joseph@codesourcery.com>
28253 Mikael Pettersson <mikpe@it.uu.se>
28256 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
28258 (ASM_OUTPUT_CASE_LABEL): Remove.
28259 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
28260 * final.c (shorten_branches): Update alignment of labels before
28261 jump tables if CASE_VECTOR_SHORTEN_MODE.
28263 2013-01-15 Richard Biener <rguenther@suse.de>
28266 * system.h: Do not include gmp.h for building host tools.
28268 2013-01-15 Richard Biener <rguenther@suse.de>
28270 PR middle-end/55882
28271 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
28272 account for bitpos when computing alignment.
28274 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
28276 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
28277 (ix86_target_macros_internal): Likewise.
28279 * config/i386/i386.c (m_CORE2I7): Removed.
28280 (m_CORE_HASWELL): New macro.
28281 (m_CORE_ALL): Likewise.
28282 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
28283 (initial_ix86_arch_features): Likewise.
28284 (processor_target_table): Initializations for Core avx2.
28285 (cpu_names): New names "core-avx2".
28286 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
28287 PROCESSOR_CORE_HASWELL.
28288 (ix86_issue_rate): New case.
28289 (ia32_multipass_dfa_lookahead): Likewise.
28290 (ix86_sched_init_global): Likewise.
28292 * config/i386/i386.h (TARGET_HASWELL): New macro.
28293 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
28294 (processor_type): New PROCESSOR_HASWELL.
28296 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28298 PR tree-optimization/55955
28299 * tree-vect-loop.c (vectorizable_reduction): Give up early on
28300 *SHIFT_EXPR and *ROTATE_EXPR codes.
28302 PR tree-optimization/48766
28303 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
28304 -ftrapv disable -fwrapv.
28306 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28309 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
28310 etc. to 1 and not to __flash.
28311 Use LL suffix for __INT24_MAX__ with -mint8.
28312 Use ULL suffix for __UINT24_MAX__ with -mint8.
28314 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28316 * config/avr/avr-arch.h
28317 (struct base_arch_s): Use typedef avr_arch_t instead.
28318 (struct arch_info_s): Use typedef avr_arch_info_t instead.
28319 (struct mcu_type_s): Use typedef avr_mcu_t instead.
28320 * config/avr/avr.c: Same.
28321 * config/avr/avr-devices.c: Same.
28322 * config/avr/driver-avr.c: Same.
28323 * config/avr/gen-avr-mmcu-texi.c: Same.
28324 * config/avr/avr-mcus.def: Adjust comment.
28326 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
28328 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
28329 * config/aarch64/iterators.md (VALLDI): New.
28331 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
28332 Andi Kleen <ak@linux.intel.com>
28335 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
28336 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
28339 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28341 * config/avr/avr-stdint.h: Remove trailing blanks.
28342 * config/avr/avr-log.h: Same.
28343 * config/avr/avr-arch.h: Same.
28344 * config/avr/avr-devices.c: Same.
28345 * config/avr/avr-dimode.md: Same.
28346 * config/avr/predicates.md: Same.
28347 * config/avr/avr-c.c: Same. And fix typo.
28349 * config/avr/avr-protos.h: Same. And:
28350 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
28351 (init_cumulative_args): Rename to avr_init_cumulative_args.
28352 (expand_prologue): Rename to avr_expand_prologue.
28353 (expand_epilogue): Rename to avr_expand_epilogue.
28354 (adjust_insn_length): Rename to avr_adjust_insn_length.
28355 (notice_update_cc): Rename to avr_notice_update_cc.
28356 (final_prescan_insn): Rename to avr_final_prescan_insn.
28357 * config/avr/avr.c: Same.
28358 * config/avr/avr.h: Same.
28359 * config/avr/avr.md: Remove trailing blanks.
28360 (prologue): Use avr_expand_prologue.
28361 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
28363 2013-01-14 Richard Biener <rguenther@suse.de>
28365 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
28366 verify_location, collect_subblocks): New functions.
28367 (verify_gimple_in_cfg): Verify that locations only reference
28368 BLOCKs in the functions BLOCK tree.
28370 2013-01-14 Richard Biener <rguenther@suse.de>
28372 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
28374 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
28376 (insert_out_of_ssa_copy_on_edge): Likewise.
28377 (rewrite_close_phi_out_of_ssa): Likewise.
28378 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
28380 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
28381 propagated constants.
28382 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
28385 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28387 * config/avr/avr-modes.def: Add GPL copyright notice.
28389 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
28391 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
28392 MEMMODEL_MASK to determine memory model.
28393 (atomic_store<mode>): Ditto from operands[2].
28394 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
28396 2013-01-13 Jakub Jelinek <jakub@redhat.com>
28399 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
28400 (fold_gimple_assign): Don't call unshare_expr here.
28401 (fold_ctor_reference): Call unshare_expr.
28403 2013-01-13 Terry Guo <terry.guo@arm.com>
28405 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
28406 * doc/fragments.texi: Document MULTILIB_REUSE.
28407 * gcc.c (multilib_reuse): New internal spec.
28408 (set_multilib_dir): Also search multilib from multilib_reuse.
28409 * genmultilib (tmpmultilib3): Refactor code.
28410 (tmpmultilib4): Ditto.
28411 (multilib_reuse): New multilib argument.
28413 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
28415 * Makefile.in: Update copyright.
28417 2013-01-12 Tom de Vries <tom@codesourcery.com>
28419 PR middle-end/55890
28420 * calls.c (expand_call): Check if arg_nr is valid.
28422 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
28424 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
28425 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
28426 documentation. Add missing '__' in front of
28427 __builtin_ia32_packssdw256.
28429 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28432 * config/s390/s390.c (s390_preferred_reload_class): Do not return
28433 NO_REGS for larl operands.
28434 (s390_reload_larl_operand): Use s390_load_address instead of
28437 2013-01-11 Richard Biener <rguenther@suse.de>
28439 * tree-cfg.c (verify_node_sharing_1): Split out from ...
28440 (verify_node_sharing): ... here.
28441 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
28443 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
28445 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
28446 Substitute TREECHECKING.
28447 * configure: Regenerate.
28448 * Makefile.in (TREECHECKING): New.
28450 2013-01-11 Richard Guenther <rguenther@suse.de>
28452 PR tree-optimization/44061
28453 * tree-vrp.c (extract_range_basic): Compute zero as
28454 value-range for __builtin_constant_p of function parameters.
28456 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
28458 Update copyright years.
28460 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
28462 PR rtl-optimization/55672
28463 * lra-eliminations.c (mark_not_eliminable): Permit addition with
28464 const to be eliminable.
28466 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
28468 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
28469 * configure: Regenerate.
28471 2013-01-10 Richard Biener <rguenther@suse.de>
28473 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
28475 2013-01-10 Richard Biener <rguenther@suse.de>
28478 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
28479 locations for virtual PHI arguments.
28480 (rewrite_update_phi_arguments): Likewise.
28482 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
28484 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
28487 2013-01-10 Jakub Jelinek <jakub@redhat.com>
28489 PR tree-optimization/55921
28490 * tree-complex.c (expand_complex_asm): New function.
28491 (expand_complex_operations_1): Call it for GIMPLE_ASM.
28493 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28496 * config/s390/s390.c (s390_symref_operand_p)
28497 (s390_loadrelative_operand_p): Merge the two functions.
28498 (s390_check_qrst_address, print_operand_address): Add parameters
28499 to s390_loadrelative_operand_p invokation.
28500 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
28501 (s390_reload_larl_operand, s390_secondary_reload): Use
28502 s390_loadrelative_operand_p instead of s390_symref_operand_p.
28503 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
28505 2013-01-09 Mike Stump <mikestump@comcast.net>
28507 * dse.c (record_store): Remove unnecessary assert.
28509 2013-01-09 Jan Hubicka <jh@suse.cz>
28511 PR tree-optimization/55569
28512 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
28513 * cfgloop.h (scale_loop_profile): Likewise.
28515 2013-01-09 Jan Hubicka <jh@suse.cz>
28518 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
28520 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
28522 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
28524 PR middle-end/55114
28525 * expr.h (maybe_emit_group_store): Declare.
28526 * expr.c (maybe_emit_group_store): New function.
28527 * builtins.c (expand_builtin_int_roundingfn): Call it.
28528 (expand_builtin_int_roundingfn_2): Likewise.
28530 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
28532 PR rtl-optimization/55829
28533 * lra-constraints.c (match_reload): Add code for absent output.
28534 (curr_insn_transform): Add code for reloads of matched inputs
28537 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
28539 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
28540 attribute of movddup insn to DF.
28541 (*vec_interleave_lowv2df): Ditto.
28542 (vec_dupv2df): Ditto.
28544 2013-01-09 Jan Hubicka <jh@suse.cz>
28546 PR tree-optimiation/55875
28547 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
28548 EVERY_ITERATION parameter.
28549 (number_of_iterations_exit): Check if exit is executed every iteration.
28550 (idx_infer_loop_bounds): Similarly here.
28551 (n_of_executions_at_most): Simplify
28552 to only test for cases where statement is dominated by the
28553 particular bound; handle correctly the "postdominance" test.
28554 (scev_probably_wraps_p): Use max loop iterations info
28555 as a global bound first.
28557 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
28558 Nick Clifton <nickc@redhat.com>
28560 * config/v850/v850.md (cbranchsf4): New pattern.
28561 (cstoresf4): New pattern.
28562 (cbranchdf4): New pattern.
28563 (cstoredf4): New pattern.
28564 (movsicc): Disallow floating point comparisons.
28565 (cmpsf_le_insn): Fix order of operators.
28566 (cmpsf_lt_insn): Likewise.
28567 (cmpsf_eq_insn): Likewise.
28568 (cmpdf_le_insn): Likewise.
28569 (cmpdf_lt_insn): Likewise.
28570 (cmpdf_eq_insn): Likewise.
28571 (cmpsf_ge_insn): Use LE comparison.
28572 (cmpdf_ge_insn): Likewise.
28573 (cmpsf_gt_insn): Use LT comparison.
28574 (cmpdf_gt_insn): Likewise.
28575 (cmpsf_ne_insn): Delete pattern.
28576 (cmpdf_ne_insn): Delete pattern.
28577 * config/v850/v850.c (v850_gen_float_compare): Use
28578 gen_cmpdf_eq_insn for NE comparison.
28579 (v850_float_z_comparison_operator)
28580 (v850_float_nz_comparison_operator): Move from here ...
28581 * config/v850/predicates.md: ... to here. Move GT and GE
28582 comparisons into v850_float_z_comparison_operator.
28583 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
28585 (v850_float_nz_comparison_operator): Likewise.
28587 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28589 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
28590 with calls to gen_insvsi/gen_insvdi.
28592 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28594 * config/i386/i386.c (initial_ix86_tune_features): Set up
28595 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
28597 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
28598 Jakub Jelinek <jakub@redhat.com>
28600 PR tree-optimization/48189
28601 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
28602 If nitercst is 0, don't predict the exit edge.
28604 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28606 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
28607 in asm_fprintf with reg_names.
28608 (aarch64_print_operand_address): Likewise.
28609 (aarch64_return_addr): Likewise.
28610 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
28612 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28614 * config/pa/pa.h (VAL_U6_BITS_P): Define.
28615 (INT_U6_BITS): Likewise.
28616 * config/pa/predicates.md (uint6_operand): New predicate.
28617 (shift5_operand, shift6_operand): Likewise.
28618 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
28620 (lshrdi3): Use shift6_operand.
28621 (shrpsi4, shrpdi4): New insn patterns.
28622 (extzv): Delete expander.
28623 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
28624 predicates in unamed zero extract patterns. Tighten common constraint.
28625 (extv): Delete expander.
28626 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
28627 predicates in unamed sign extract patterns. Tighten common constraint.
28628 (insv): Delete expander.
28629 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
28630 predicates in unamed insert patterns. Tighten common constraint.
28631 Change uint32_operand predicate to uint6_operand predicate in unamed
28632 DImode pattern to insert constant values of type 1...1xxxx.
28634 2013-01-04 Jan Hubicka <jh@suse.cz>
28636 PR tree-optimization/55823
28637 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
28640 2013-01-08 Jakub Jelinek <jakub@redhat.com>
28641 Uros Bizjak <ubizjak@gmail.com>
28643 PR rtl-optimization/55845
28644 * df-problems.c (can_move_insns_across): Stop scanning at
28645 volatile_insn_p source instruction or give up if
28646 across_from .. across_to range contains any volatile_insn_p
28649 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
28651 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
28652 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
28654 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
28655 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
28657 2013-01-08 Jakub Jelinek <jakub@redhat.com>
28660 * asan.c (asan_clear_shadow): New function.
28661 (asan_emit_stack_protection): Use it.
28663 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
28665 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
28666 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
28667 with tab instead of space.
28669 2013-01-08 Nick Clifton <nickc@redhat.com>
28671 * config/rl78/rl78.c (rl78_expand_prologue): Always select
28672 register bank 0 at the start of an interrupt handler.
28673 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
28676 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
28678 * config/aarch64/aarch64-simd.md
28679 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
28680 (aarch64_simd_bsl): Likewise.
28681 (aarch64_vcond_internal<mode>): Likewise.
28682 (vcond<mode><mode>): Likewise.
28683 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
28684 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
28686 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
28688 * config/aarch64/aarch64-builtins.c
28689 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
28691 2013-01-08 Martin Jambor <mjambor@suse.cz>
28694 * tree-sra.c (analyze_access_subtree): Return true also after
28695 potentially creating a debug-only replacement.
28697 2013-01-08 Jakub Jelinek <jakub@redhat.com>
28699 PR middle-end/55890
28700 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
28702 PR tree-optimization/54120
28703 * tree-vrp.c (range_fits_type_p): Don't allow
28704 src_precision < precision from signed vr to unsigned_p
28705 if vr->min or vr->max is negative.
28706 (simplify_float_conversion_using_ranges): Test can_float_p
28707 against CODE_FOR_nothing.
28709 2013-01-08 Jakub Jelinek <jakub@redhat.com>
28710 Richard Biener <rguenther@suse.de>
28712 PR middle-end/55851
28713 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
28714 types instead of just INTEGER_TYPE types.
28716 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
28718 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
28721 2013-01-07 Steve Ellcey <sellcey@mips.com>
28724 * config/mips/mips.opt: Change mad to mmad to match documentation.
28726 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28729 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
28730 .progmemx.data now.
28732 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28735 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
28736 (avr_addrspace_t): Add .section_name field.
28737 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
28739 (avr_addrspace): Same. Initialize .section_name. Remove last
28740 NULL entry. Put __memx into .progmemx.data.
28741 (progmem_section_prefix): Remove.
28742 (avr_asm_init_sections): No need to initialize progmem_section.
28743 (avr_asm_named_section): Use avr_addrspace[].section_name to get
28744 section name prefix.
28745 (avr_asm_select_section): Ditto. And use get_unnamed_section to
28746 retrieve the progmem section.
28747 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
28748 boundary to run over avr_addrspace[].
28749 (avr_register_target_pragmas): Ditto.
28751 2013-01-06 Jakub Jelinek <jakub@redhat.com>
28753 * varasm.c (output_constant_def_contents): For asan_protect_global
28754 protected strings, adjust DECL_ALIGN if needed, before testing for
28756 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
28757 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
28759 (output_object_block): For asan protected decls, emit asan padding
28760 after their contents.
28761 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
28762 (asan_finish_file): Test it here instead.
28764 2013-01-07 Nick Clifton <nickc@redhat.com>
28765 Matthias Klose <doko@debian.org>
28766 Doug Kwan <dougkwan@google.com>
28767 H.J. Lu <hongjiu.lu@intel.com>
28770 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
28772 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
28774 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
28776 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
28778 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
28780 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28783 * doc/install.texi (Cross-Compiler-Specific Options): Document
28786 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
28788 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
28789 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
28790 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
28791 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
28792 vqmovun_high_s64): Fix source operand number and update copyright.
28794 2013-01-07 Richard Biener <rguenther@suse.de>
28796 PR middle-end/55890
28797 * gimple.h (gimple_call_builtin_p): New overload.
28798 * gimple.c (validate_call): New function.
28799 (gimple_call_builtin_p): Likewise.
28800 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
28801 Use gimple_call_builtin_p.
28802 (find_func_clobbers): Likewise.
28803 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
28804 (strlen_optimize_stmt): Likewise.
28806 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
28808 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
28809 (vld1q_dup_*): Likewise.
28810 (vld1_*): Likewise.
28811 (vld1q_*): Likewise.
28812 (vld1_lane_*): Likewise.
28813 (vld1q_lane_*): Likewise.
28815 2013-01-07 Richard Biener <rguenther@suse.de>
28817 * lto-streamer.h (LTO_minor_version): Bump to 2.
28819 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
28821 * config/aarch64/aarch64-protos.h
28822 (aarch64_const_double_zero_rtx_p): Rename to...
28823 (aarch64_float_const_zero_rtx_p): ...this.
28824 (aarch64_float_const_representable_p): New.
28825 (aarch64_output_simd_mov_immediate): Likewise.
28826 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
28827 move immediate case.
28828 * config/aarch64/aarch64.c
28829 (aarch64_const_double_zero_rtx_p): Rename to...
28830 (aarch64_float_const_zero_rtx_p): ...this.
28831 (aarch64_print_operand): Allow printing of new constants.
28832 (aarch64_valid_floating_const): New.
28833 (aarch64_legitimate_constant_p): Check for valid floating-point
28835 (aarch64_simd_valid_immediate): Likewise.
28836 (aarch64_vect_float_const_representable_p): New.
28837 (aarch64_float_const_representable_p): Likewise.
28838 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
28839 (aarch64_output_simd_mov_immediate): New.
28840 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
28841 (*movdf_aarch64): Likewise.
28842 * config/aarch64/constraints.md (Ufc): New.
28843 (Y): call aarch64_float_const_zero_rtx.
28844 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
28846 2013-01-07 Richard Biener <rguenther@suse.de>
28848 PR tree-optimization/55888
28849 PR tree-optimization/55862
28850 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
28851 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
28852 not if it is contained therein.
28854 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28856 * config/avr/t-avr: Typo.
28858 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28861 * config/avr/t-avr: Don't automatically rebuild
28862 $(srcdir)/config/avr/t-multilib
28863 $(srcdir)/config/avr/avr-tables.opt
28864 $(srcdir)/doc/avr-mmcu.texi
28865 (avr-mcus): New phony target to build them on request.
28866 (s-avr-mlib, s-avr-mmcu-texi): Remove.
28867 * avr/avr-mcus.def: Adjust comments.
28869 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
28871 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
28873 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
28875 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
28877 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
28879 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
28881 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
28883 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
28884 to generate profiling.
28885 * config/rs6000/aix64.h (LIB_SPEC): Same.
28887 2013-01-04 Andrew Pinski <apinski@cavium.com>
28889 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
28891 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
28893 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
28895 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
28897 (ix86_expand_move): Ditto.
28898 (ix86_zero_extend_to_Pmode): Ditto.
28899 (ix86_expand_call): Ditto.
28900 (ix86_expand_special_args_builtin): Ditto.
28901 (ix86_expand_builtin): Ditto.
28903 2013-01-04 Richard Biener <rguenther@suse.de>
28905 PR tree-optimization/55862
28906 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
28907 translating them through PHI nodes.
28909 2013-01-04 Martin Jambor <mjambor@suse.cz>
28911 PR tree-optimization/55755
28912 * tree-sra.c (sra_modify_assign): Do not check that an access has no
28913 children when trying to avoid producing a VIEW_CONVERT_EXPR.
28915 2013-01-04 Marek Polacek <polacek@redhat.com>
28917 PR middle-end/55859
28918 * opts.c (default_options_optimization): Clarify error message.
28920 2013-01-04 Richard Biener <rguenther@suse.de>
28922 PR middle-end/55863
28923 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
28926 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28929 * config/pa/pa.md (movsi): Revert previous change.
28930 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
28933 2013-01-03 Richard Henderson <rth@redhat.com>
28935 * config/i386/i386.c (ix86_expand_move): Always assign to op1
28936 after eliminating TLS symbols.
28938 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
28941 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
28942 * graphite-poly.c (debug_gmp_value): Likewise.
28944 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
28947 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
28948 selected code model, define __code_mode_small__, __code_model_medium__,
28949 __code_model_large__, __code_model_32__ or __code_model_kernel__.
28950 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
28951 xchg temporary register with %k. Declare temporary register as
28953 [__x86_64__]: For medium and large code models, preserve %rbx register.
28955 2013-01-03 Richard Biener <rguenther@suse.de>
28957 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
28958 (dump_subscript): Adjust.
28959 (finalize_ddr_dependent): Do not dump redundant info.
28960 (analyze_siv_subscript): Adjust.
28961 (subscript_dependence_tester): Likewise.
28962 (compute_affine_dependence): Likewise.
28964 2013-01-03 Richard Biener <rguenther@suse.de>
28967 2013-01-03 Richard Biener <rguenther@suse.de>
28969 PR tree-optimization/55857
28970 * tree-vect-stmts.c (vectorizable_load): Do not setup
28971 re-alignment for invariant loads.
28973 2013-01-02 Richard Biener <rguenther@suse.de>
28975 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
28976 invariant load do not generate a vector load from the scalar location.
28978 2013-01-03 Richard Biener <rguenther@suse.de>
28980 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
28981 for not vectorizing.
28982 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
28983 not build INDIRECT_REFs, call get_name once only.
28984 (vect_create_data_ref_ptr): Likewise. Dump base object kind
28985 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
28987 2013-01-03 Richard Biener <rguenther@suse.de>
28989 PR tree-optimization/55857
28990 * tree-vect-stmts.c (vectorizable_load): Do not setup
28991 re-alignment for invariant loads.
28993 2013-01-03 Richard Biener <rguenther@suse.de>
28996 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
28997 prefer a built-in decl.
28999 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29001 * gcc.c (process_command): Update copyright notice dates.
29002 * gcov.c (print_version): Likewise.
29003 * gcov-dump.c (print_version): Likewise.
29005 PR rtl-optimization/55838
29006 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29007 iv0.step, iv1.step and step.
29009 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29010 Marc Glisse <marc.glisse@inria.fr>
29012 PR tree-optimization/55832
29013 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29014 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29015 integer_{one,zero}_node.
29017 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29020 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29021 * var-tracking.c (reverse_op): Don't add reverse ops to
29022 VALUEs that have already
29023 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29025 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29027 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29029 2013-01-02 Teresa Johnson <tejohnson@google.com>
29031 * dumpfile.c (dump_loc): Print filename with location.
29032 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29033 new location_t parameter to emit complete unroll message with
29034 new dump framework.
29035 (canonicalize_loop_induction_variables): Compute loops location
29036 and pass to try_unroll_loop_completely.
29037 * loop-unroll.c (report_unroll_peel): New function.
29038 (peel_loops_completely): Use new dump format with location
29039 for main dumpfile message, and invoke report_unroll_peel on success.
29040 (decide_unrolling_and_peeling): Ditto.
29041 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29042 by report_unroll_peel.
29043 (decide_peel_completely): Ditto.
29044 (decide_unroll_constant_iterations): Ditto.
29045 (decide_unroll_runtime_iterations): Ditto.
29046 (decide_peel_simple): Ditto.
29047 (decide_unroll_stupid): Ditto.
29048 * cfgloop.c (get_loop_location): New function.
29049 * cfgloop.h (get_loop_location): Declare.
29051 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29053 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29056 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29058 PR middle-end/55198
29059 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29060 BLKmode objects when EXPAND_MEMORY is specified.
29062 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29064 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29066 (fold_builtin_cpu): Do not share cpu model decls across statements.
29068 2013-01-02 Jason Merrill <jason@redhat.com>
29071 * tree.c (build_array_type_1): Revert earlier change.
29073 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29075 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29077 * config/aarch64/aarch64-tune.md: Re-generate.
29079 2013-01-02 Richard Biener <rguenther@suse.de>
29081 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29082 invariant load do not generate a vector load from the scalar location.
29084 2013-01-02 Richard Biener <rguenther@suse.de>
29087 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29088 * configure: Regenerate.
29090 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29092 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29093 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29094 (expand_builtin_int_roundingfn_2): Keep the original target around
29095 for the fallback case.
29097 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29099 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29100 to be clear for sign changes.
29102 2013-01-01 Jan Hubicka <jh@suse.cz>
29104 * ipa-inline-analysis.c: Fix formatting.
29106 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29108 PR tree-optimization/55831
29109 * tree-vect-loop.c (get_initial_def_for_induction): Use
29110 gsi_after_labels instead of gsi_start_bb.
29112 Copyright (C) 2013 Free Software Foundation, Inc.
29114 Copying and distribution of this file, with or without modification,
29115 are permitted in any medium without royalty provided the copyright
29116 notice and this notice are preserved.