1 2013-10-31 Jakub Jelinek <jakub@redhat.com>
3 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
4 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
5 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
6 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
7 uses but in the condition and ASSERT_EXPR and the other successor of
8 the predecessor bb is __builtin_unreachable (), set_range_info of the
9 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
11 2013-10-31 Martin Jambor <mjambor@suse.cz>
13 PR rtl-optimization/58934
15 2013-10-30 Martin Jambor <mjambor@suse.cz>
16 PR rtl-optimization/10474
17 * ira.c (find_moveable_pseudos): Do not calculate dominance info
19 (interesting_dest_for_shprep): New function.
20 (split_live_ranges_for_shrink_wrap): Likewise.
21 (ira): Calculate dominance info and df analysis. Call
22 split_live_ranges_for_shrink_wrap.
24 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
25 Yury Gribov <y.gribov@samsung.com>
28 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
29 shadow pointer to avoid clobbering the main one.
31 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
33 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
35 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
38 * ira-color.c (update_costs_from_copies): Add new parameter. Use
39 it for calling update_costs_from_allocno.
40 (assign_hard_reg): Call restore_costs_from_copies only for
41 !retry_p. Pass new argument to update_costs_from_copies.
42 (color_pass): Pass new argument to update_costs_from_copies.
43 (ira_mark_allocation_change): Ditto.
45 2013-10-30 Sharad Singhai <singhai@google.com>
48 * opts.c (common_handle_option): Remove deprecated option
49 -ftree-vectorizer-verbose.
50 * doc/invoke.texi (Debugging Options): Ditto.
51 * opts-global.c (handle_common_deferred_options): Ditto.
52 (dump_remap_tree_vectorizer_verbose): Delete.
53 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
55 2013-10-30 DJ Delorie <dj@redhat.com>
57 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
58 that take no arguments.
60 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
63 * reload1.c (update_eliminables_and_spill): New function, broken
65 (reload): Use it. Check for frame size change after frame
66 size alignment, and call update_eliminables_and_spill first
69 2013-10-30 Cong Hou <congh@google.com>
72 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
73 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
74 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
76 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
78 PR rtl-optimization/58369
79 * reload1.c (compute_reload_subreg_offset): New function.
80 (choose_reload_regs): Use it to pass endian-correct
81 offset to subreg_regno_offset.
83 2013-10-30 Tobias Burnus <burnus@net-b.de>
86 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
89 2013-10-30 Jason Merrill <jason@redhat.com>
91 * configure.ac (loose_warn): Add -Wno-format if
92 --disable-build-format-warnings.
94 2013-10-30 David Malcolm <dmalcolm@redhat.com>
96 * cgraphunit.c (analyze_functions): Split symtab_node
97 declarations onto multiple lines to make things easier
100 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
101 (symtab_semantically_equivalent_p): Likewise.
103 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
106 * lra.c (check_rtl): Remove address check before LRA work.
108 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
110 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
111 POINTER_PLUS_EXPR in the defining statement.
113 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
116 * tree-pass.h (make_pass_regmove): Remove.
117 * timevar.def (TV_REGMOVE): Remove.
118 * passes.def (pass_regmove): Remove.
119 * opts.c (default_options_table): Remove entry for regmove.
120 * doc/passes.texi: Remove regmove pass description.
121 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
123 (-fdump-rtl-regmove): Ditto.
124 * common.opt (foptimize-register-move, fregmove): Ignore.
125 * Makefile.in (OBJS): Remove regmove.o.
127 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
129 (struct ira_allocno) New member allocno_prefs.
130 (ALLOCNO_PREFS): New macro.
131 (ira_prefs, ira_prefs_num): New external vars.
132 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New
134 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref):
136 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
138 (ira_add_allocno_copy_to_list): Remove prototype.
139 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
140 (ira_pref_iterator): New type.
141 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
142 (FOR_EACH_PREF): New macro.
143 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
144 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
146 (ira_setup_alts): New function.
147 (decrease_live_ranges_number): New function.
148 (ira): Call the above function.
149 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
150 (ira_create_allocno): Initialize allocno prefs.
151 (pref_pool, pref_vec): New static vars.
152 (initiate_prefs, find_allocno_pref, ira_create_pref): New
154 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
155 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
156 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
157 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
158 (ira_add_allocno_copy_to_list): Make static. Rename to
159 add_allocno_copy_to_list.
160 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
161 swap_allocno_copy_ends_if_necessary.
162 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
163 ira_remove_allocno_prefs.
164 (ira_flattening): Ditto.
165 (ira_build): Call initiate_prefs, print_prefs.
166 (ira_destroy): Call finish_prefs.
167 * ira-color.c (struct update_cost_record): New.
168 (struct allocno_color_data): Add new member update_cost_records.
169 (update_cost_record_pool): New static var.
170 (init_update_cost_records, get_update_cost_record): New functions.
171 (free_update_cost_record_list, finish_update_cost_records): Ditto.
172 (struct update_cost_queue_elem): Add member from.
173 (initiate_cost_update): Call init_update_cost_records.
174 (finish_cost_update): Call finish_update_cost_records.
175 (queue_update_cost, get_next_update_cost): Add new param from.
176 (Update_allocno_cost, update_costs_from_allocno): New functions.
177 (update_costs_from_prefs): Ditto.
178 (update_copy_costs): Rename to update_costs_from_copies.
179 (restore_costs_from_copies): New function.
180 (update_conflict_hard_regno_costs): Don't go back.
181 (assign_hard_reg): Call restore_costs_from_copies. Add printing
183 (pop_allocnos): Add priniting more debug info.
184 (color_allocnos): Remove prefs for conflicting hard regs.
185 Call update_costs_from_prefs.
186 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
187 (get_dup_num): Rename, modify, and move to ira.c
188 (process_regs_for_copy): Add prefs.
189 (add_insn_allocno_copies): Put src as first arg of
190 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
191 * ira-costs.c (record_reg_classes): Modify and move code into
192 record_operands_costs.
193 (find_costs_and_classes): Create prefs for the hard reg of small
195 (process_bb_node_for_hard_reg_moves): Add prefs.
197 2013-10-30 Richard Biener <rguenther@suse.de>
200 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
201 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
203 (pre_and_rev_post_order_compute): ... which now wraps it.
204 * graph.c (draw_cfg_nodes_no_loops): Use
205 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
208 2013-10-30 Christian Bruel <christian.bruel@st.com>
210 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
211 (sh_expand_cmpstr): Handle known align and schedule improvements.
212 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
213 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
215 2013-10-30 Martin Jambor <mjambor@suse.cz>
217 PR rtl-optimization/10474
218 * ira.c (find_moveable_pseudos): Do not calculate dominance info
220 (interesting_dest_for_shprep): New function.
221 (split_live_ranges_for_shrink_wrap): Likewise.
222 (ira): Calculate dominance info and df analysis. Call
223 split_live_ranges_for_shrink_wrap.
225 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
228 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
230 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
232 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
233 * tree.h (POINTER_BOUNDS_P): New.
234 (BOUNDED_TYPE_P): New.
236 (pointer_bounds_type_node): New.
237 * tree.c (build_common_tree_nodes): Initialize
238 pointer_bounds_type_node.
239 * gimple.h (gimple_call_get_nobnd_arg_index): New.
240 (gimple_call_num_nobnd_args): New.
241 (gimple_call_nobnd_arg): New.
242 (gimple_return_retbnd): New.
243 (gimple_return_set_retbnd): New
244 * gimple.c (gimple_build_return): Increase number of ops
245 for return statement.
246 (gimple_call_get_nobnd_arg_index): New.
247 * gimple-pretty-print.c (dump_gimple_return): Print second op.
249 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
251 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
252 with "chkp ctor" and "bnd_legacy" attributes.
253 * gimplify.c (gimplify_init_constructor): Avoid infinite
254 loop during gimplification of bounds initializer.
256 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
258 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
259 (handle_bnd_legacy): New.
260 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
261 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
264 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
266 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
267 (BT_FN_PTR_CONST_PTR): New.
268 (BT_FN_CONST_PTR_CONST_PTR): New.
269 (BT_FN_PTR_CONST_PTR_SIZE): New.
270 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
271 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
272 (BT_FN_VOID_CONST_PTR_SIZE): New.
273 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
274 * chkp-builtins.def: New.
275 * builtins.def: include chkp-builtins.def.
276 (DEF_CHKP_BUILTIN): New.
277 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
278 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
279 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
280 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
281 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
282 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
283 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
284 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
285 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
286 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
287 * common.opt (fcheck-pointer-bounds): New.
288 * toplev.c (process_options): Check Pointer Bounds Checker is supported.
289 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
291 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
293 * target.def (builtin_chkp_function): New.
294 (chkp_bound_type): New.
295 (chkp_bound_mode): New.
296 (fn_abi_va_list_bounds_size): New.
297 (load_bounds_for_arg): New.
298 (store_bounds_for_arg): New.
299 * targhooks.h (default_load_bounds_for_arg): New.
300 (default_store_bounds_for_arg): New.
301 (default_fn_abi_va_list_bounds_size): New.
302 (default_chkp_bound_type): New.
303 (default_chkp_bound_mode): New.
304 (default_builtin_chkp_function): New.
305 * targhooks.c (default_load_bounds_for_arg): New.
306 (default_store_bounds_for_arg): New.
307 (default_fn_abi_va_list_bounds_size): New.
308 (default_chkp_bound_type): New.
309 (default_chkp_bound_mode); New.
310 (default_builtin_chkp_function): New.
311 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
312 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
313 (TARGET_STORE_BOUNDS_FOR_ARG): New.
314 (TARGET_BUILTIN_CHKP_FUNCTION): New.
315 (TARGET_CHKP_BOUND_TYPE): New.
316 (TARGET_CHKP_BOUND_MODE): New.
317 * doc/tm.texi: Regenerated.
318 * langhooks.h (lang_hooks): Add chkp_supported field.
319 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
320 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
322 2013-10-29 Andrew Pinski <apinski@cavium.com>
324 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
325 (ifcombine_ifandif): Handle cases where
326 maybe_fold_and_comparisons fails, combining the branches
328 (tree_ssa_ifcombine): Inverse the order of
329 the basic block walk, increases the number of combinings.
330 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
332 2013-10-29 Mike Stump <mikestump@comcast.net>
334 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
335 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
336 (make_vector_mode): Increase namebuf to 16.
337 (emit_insn_modes_h): When processing BImode, don't
338 also match partial int modes.
339 (emit_class_narrowest_mode): Likewise.
341 * config/bfin/bfin-modes.def: Add precision to PDI.
342 * config/m32c/m32c-modes.def: Add precision to PSI.
343 * config/msp430/msp430-modes.def: Add precision to PSI.
344 * config/rs6000/rs6000-modes.def: Add precision to PTI.
345 * config/sh/sh-modes.def: Add precision to PSI and PDI.
347 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
350 * config/sh/sh.md (*addc): Rename existing variations to ...
351 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
352 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
353 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
354 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
356 2013-10-29 Teresa Johnson <tejohnson@google.com>
359 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
360 insanities when updating probabilities.
362 2013-10-29 David Malcolm <dmalcolm@redhat.com>
364 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
365 prettyprinter for cgraph_node to reflect the conversion of the
366 symtable types to a C++ class hierarchy: it now *is* a
367 symtab_node_base, rather than having one (named "symbol").
369 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
371 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
372 __cilkrts_pop_frame. If matched, then return true for built-in
374 (expand_builtin): Added BUILT_IN_CILK_DETACH and
375 BUILT_IN_CILK_POP_FRAME case.
376 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
377 (lhs_cilk_detect_spawn): Likewise.
378 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
379 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
380 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
381 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
382 (LANG_HOOKS_CILKPLUS): Likewise.
383 * tree.h (CILK_SPAWN_FN): Likewise.
384 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
385 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
386 (OBJS): Added cilk-common.o.
387 (BUILTINS_DEF): Added cilk-builtins.def.
388 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
389 (lhd_cilk_detect_spawn): Likewise.
390 * langhooks.h (lang_hooks_for_cilkplus): New struct.
391 (struct lang_hooks): Added new field called "cilkplus."
392 * cilk-common.c: New file.
394 * cilk-builtins.def: Likewise.
395 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
396 "__cilk" macro and set it to 200.
397 * function.h (struct function::cilk_frame_decl): New field.
398 (struct function::is_cilk_function): Likewise.
399 (struct function::calls_cilk_spawn): Likewise.
400 * gimplify.c (gimplify_call_expr): Added a check if the function call
401 being gimplified is a spawn detach point. If so, then add pop_frame
402 and detach function calls.
403 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
404 for gimplifying _Cilk_spawn and _Cilk_sync statements.
405 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
406 function. If so, added a _Cilk_sync and gimplified it.
407 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
408 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
409 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
411 (can_inline_edge_p): Prevent inling of spawnee function.
412 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
413 for functions that use Cilk keywords.
414 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
415 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
416 CILK_SYNC_STMT cases.
417 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
419 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
420 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
421 * passes.texi (Cilk Keywords): New section that describes the compiler
422 code changes for handling Cilk Keywords.
424 2013-10-29 David Malcolm <dmalcolm@redhat.com>
426 Patch autogenerated by refactor_symtab.py from
427 https://github.com/davidmalcolm/gcc-refactoring-scripts
428 revision 58bb219cc090b2f4516a9297d868c245495ee622
430 * asan.c (asan_finish_file): Update for conversion of symtab types to
431 a true class hierarchy.
432 * cfgexpand.c (estimated_stack_frame_size): Likewise.
433 * cgraph.c (cgraph_get_body): Likewise.
434 (cgraph_get_create_real_symbol_node): Likewise.
435 (verify_cgraph_node): Likewise.
436 (verify_edge_corresponds_to_fndecl): Likewise.
437 (verify_edge_count_and_frequency): Likewise.
438 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
439 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
440 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
441 (cgraph_node_cannot_return): Likewise.
442 (cgraph_set_pure_flag_1): Likewise.
443 (cgraph_set_const_flag_1): Likewise.
444 (cgraph_set_nothrow_flag_1): Likewise.
445 (cgraph_make_node_local_1): Likewise.
446 (cgraph_for_node_and_aliases): Likewise.
447 (cgraph_for_node_thunks_and_aliases): Likewise.
448 (cgraph_node_can_be_local_p): Likewise.
449 (cgraph_node_cannot_be_local_p_1): Likewise.
450 (cgraph_function_body_availability): Likewise.
451 (dump_cgraph_node): Likewise.
452 (cgraph_rtl_info): Likewise.
453 (cgraph_mark_address_taken_node): Likewise.
454 (cgraph_remove_node): Likewise.
455 (cgraph_release_function_body): Likewise.
456 (cgraph_update_edges_for_call_stmt_node): Likewise.
457 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
458 (cgraph_make_edge_direct): Likewise.
459 (cgraph_resolve_speculation): Likewise.
460 (cgraph_speculative_call_info): Likewise.
461 (cgraph_turn_edge_to_speculative): Likewise.
462 (cgraph_create_edge_1): Likewise.
463 (cgraph_set_call_stmt): Likewise.
464 (cgraph_node_for_asm): Likewise.
465 (cgraph_add_thunk): Likewise.
466 (cgraph_same_body_alias): Likewise.
467 (cgraph_create_function_alias): Likewise.
468 (cgraph_create_node): Likewise.
469 (cgraph_create_empty_node): Likewise.
470 (record_function_versions): Likewise.
471 (used_from_object_file_p): Likewise.
472 * cgraph.h (symtab_can_be_discarded): Likewise.
473 (symtab_real_symbol_p): Likewise.
474 (cgraph_mark_force_output_node): Likewise.
475 (cgraph_edge_recursive_p): Likewise.
476 (symtab_alias_target): Likewise.
477 (varpool_all_refs_explicit_p): Likewise.
478 (varpool_can_remove_if_no_refs): Likewise.
479 (cgraph_only_called_directly_or_aliased_p): Likewise.
480 (cgraph_next_function_with_gimple_body): Likewise.
481 (cgraph_first_function_with_gimple_body): Likewise.
482 (cgraph_function_with_gimple_body_p): Likewise.
483 (cgraph_next_function): Likewise.
484 (cgraph_first_function): Likewise.
485 (cgraph_next_defined_function): Likewise.
486 (cgraph_first_defined_function): Likewise.
487 (varpool_next_defined_variable): Likewise.
488 (varpool_first_defined_variable): Likewise.
489 (varpool_next_static_initializer): Likewise.
490 (varpool_first_static_initializer): Likewise.
491 (varpool_next_variable): Likewise.
492 (varpool_first_variable): Likewise.
493 (varpool_node_name): Likewise.
496 (is_a_helper <varpool_node>::test): Likewise.
497 (is_a_helper <cgraph_node>::test): Likewise.
498 (varpool_variable_node): Likewise.
499 (cgraph_function_or_thunk_node): Likewise.
500 (varpool_alias_target): Likewise.
501 (cgraph_alias_target): Likewise.
502 (cgraph_node_name): Likewise.
503 (varpool_node_asm_name): Likewise.
504 (cgraph_node_asm_name): Likewise.
505 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
506 (cgraph_rebuild_references): Likewise.
507 (rebuild_cgraph_edges): Likewise.
508 (record_eh_tables): Likewise.
509 (build_cgraph_edges): Likewise.
510 (mark_store): Likewise.
511 (mark_load): Likewise.
512 (mark_address): Likewise.
513 (record_type_list): Likewise.
514 (record_reference): Likewise.
515 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
516 (cgraph_materialize_clone): Likewise.
517 (cgraph_function_versioning): Likewise.
518 (cgraph_copy_node_for_versioning): Likewise.
519 (update_call_expr): Likewise.
520 (cgraph_find_replacement_node): Likewise.
521 (cgraph_create_virtual_clone): Likewise.
522 (cgraph_clone_node): Likewise.
523 * cgraphunit.c (compile): Likewise.
524 (output_weakrefs): Likewise.
525 (output_in_order): Likewise.
526 (expand_function): Likewise.
527 (assemble_thunks_and_aliases): Likewise.
528 (expand_thunk): Likewise.
529 (mark_functions_to_output): Likewise.
530 (handle_alias_pairs): Likewise.
531 (analyze_functions): Likewise.
532 (walk_polymorphic_call_targets): Likewise.
533 (varpool_finalize_decl): Likewise.
534 (process_function_and_variable_attributes): Likewise.
535 (cgraph_process_same_body_aliases): Likewise.
536 (analyze_function): Likewise.
537 (cgraph_add_new_function): Likewise.
538 (cgraph_finalize_function): Likewise.
539 (referred_to_p): Likewise.
540 (cgraph_reset_node): Likewise.
541 (cgraph_process_new_functions): Likewise.
542 (enqueue_node): Likewise.
543 (decide_is_symbol_needed): Likewise.
544 * coverage.c (coverage_compute_profile_id): Likewise.
545 * dbxout.c (dbxout_expand_expr): Likewise.
546 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
547 (reference_to_unused): Likewise.
548 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
549 * gimplify.c (unvisit_body): Likewise.
550 (unshare_body): Likewise.
551 * ipa-cp.c (ipcp_generate_summary): Likewise.
552 (ipcp_decision_stage): Likewise.
553 (identify_dead_nodes): Likewise.
554 (decide_whether_version_node): Likewise.
555 (decide_about_value): Likewise.
556 (perhaps_add_new_callers): Likewise.
557 (create_specialized_node): Likewise.
558 (update_profiling_info): Likewise.
559 (ipcp_propagate_stage): Likewise.
560 (estimate_local_effects): Likewise.
561 (good_cloning_opportunity_p): Likewise.
562 (devirtualization_time_bonus): Likewise.
563 (propagate_constants_accross_call): Likewise.
564 (initialize_node_lattices): Likewise.
565 (ipcp_cloning_candidate_p): Likewise.
566 (determine_versionability): Likewise.
567 (print_all_lattices): Likewise.
568 (print_lattice): Likewise.
569 (ipcp_discover_new_direct_edges): Likewise.
570 * ipa-devirt.c (ipa_devirt): Likewise.
571 (likely_target_p): Likewise.
572 (update_type_inheritance_graph): Likewise.
573 (possible_polymorphic_call_target_p): Likewise.
574 (dump_possible_polymorphic_call_targets): Likewise.
575 (devirt_variable_node_removal_hook): Likewise.
576 (record_binfo): Likewise.
577 (maybe_record_node): Likewise.
578 (build_type_inheritance_graph): Likewise.
579 * ipa-inline-analysis.c (inline_write_summary): Likewise.
580 (inline_generate_summary): Likewise.
581 (inline_analyze_function): Likewise.
582 (do_estimate_growth): Likewise.
583 (simple_edge_hints): Likewise.
584 (estimate_node_size_and_time): Likewise.
585 (estimate_edge_devirt_benefit): Likewise.
586 (compute_inline_parameters): Likewise.
587 (estimate_function_body_sizes): Likewise.
588 (compute_bb_predicates): Likewise.
589 (initialize_inline_failed): Likewise.
590 (dump_inline_summary): Likewise.
591 (dump_inline_edge_summary): Likewise.
592 * ipa-inline-transform.c (inline_transform): Likewise.
593 (preserve_function_body_p): Likewise.
594 (save_inline_function_body): Likewise.
595 (inline_call): Likewise.
596 (clone_inlined_nodes): Likewise.
597 (can_remove_node_now_p): Likewise.
598 (can_remove_node_now_p_1): Likewise.
599 * ipa-inline.c (early_inliner): Likewise.
600 (early_inline_small_functions): Likewise.
601 (inline_always_inline_functions): Likewise.
602 (ipa_inline): Likewise.
603 (flatten_function): Likewise.
604 (inline_small_functions): Likewise.
605 (speculation_useful_p): Likewise.
606 (recursive_inlining): Likewise.
607 (update_caller_keys): Likewise.
608 (reset_edge_caches): Likewise.
609 (update_edge_key): Likewise.
610 (edge_badness): Likewise.
611 (relative_time_benefit): Likewise.
612 (want_inline_self_recursive_call_p): Likewise.
613 (want_inline_small_function_p): Likewise.
614 (want_early_inline_function_p): Likewise.
615 (num_calls): Likewise.
616 (can_early_inline_edge_p): Likewise.
617 (can_inline_edge_p): Likewise.
618 (report_inline_failed_reason): Likewise.
619 * ipa-profile.c (ipa_profile): Likewise.
620 (ipa_propagate_frequency): Likewise.
621 (ipa_propagate_frequency_1): Likewise.
622 (ipa_profile_generate_summary): Likewise.
623 * ipa-prop.c (ipcp_transform_function): Likewise.
624 (read_replacements_section): Likewise.
625 (ipa_prop_read_section): Likewise.
626 (ipa_modify_call_arguments): Likewise.
627 (ipa_print_node_params): Likewise.
628 (propagate_controlled_uses): Likewise.
629 (update_indirect_edges_after_inlining): Likewise.
630 (remove_described_reference): Likewise.
631 (ipa_make_edge_direct_to_target): Likewise.
632 (ipa_analyze_node): Likewise.
633 (ipa_analyze_params_uses): Likewise.
634 (ipa_compute_jump_functions): Likewise.
635 (ipa_get_callee_param_type): Likewise.
636 (ipa_print_node_jump_functions): Likewise.
637 (ipa_initialize_node_params): Likewise.
638 (ipa_populate_param_decls): Likewise.
639 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
640 (write_agg_replacement_chain): Likewise.
641 (ipa_write_node_info): Likewise.
642 (ipa_edge_duplication_hook): Likewise.
643 (try_decrement_rdesc_refcount): Likewise.
644 * ipa-pure-const.c (propagate_nothrow): Likewise.
645 (propagate_pure_const): Likewise.
646 (pure_const_read_summary): Likewise.
647 (pure_const_write_summary): Likewise.
648 (analyze_function): Likewise.
649 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
650 (ipa_ref_referring_ref_list): Likewise.
651 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
652 (ipa_remove_stmt_references): Likewise.
653 (ipa_find_reference): Likewise.
654 (ipa_dump_referring): Likewise.
655 (ipa_dump_references): Likewise.
656 (ipa_record_reference): Likewise.
657 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
658 (ipa_reference_write_optimization_summary): Likewise.
659 (write_node_summary_p): Likewise.
660 (propagate): Likewise.
661 (read_write_all_from_decl): Likewise.
662 (generate_summary): Likewise.
663 (analyze_function): Likewise.
664 (propagate_bits): Likewise.
665 (ipa_reference_get_not_written_global): Likewise.
666 (ipa_reference_get_not_read_global): Likewise.
667 * ipa-split.c (execute_split_functions): Likewise.
668 (split_function): Likewise.
669 * ipa-utils.c (ipa_merge_profiles): Likewise.
670 (dump_cgraph_node_set): Likewise.
671 (ipa_reverse_postorder): Likewise.
672 (ipa_edge_within_scc): Likewise.
673 (ipa_get_nodes_in_cycle): Likewise.
674 (ipa_free_postorder_info): Likewise.
675 (ipa_reduced_postorder): Likewise.
677 (recursive_call_p): Likewise.
678 * ipa.c (ipa_cdtor_merge): Likewise.
679 (record_cdtor_fn): Likewise.
680 (function_and_variable_visibility): Likewise.
681 (varpool_externally_visible_p): Likewise.
682 (cgraph_externally_visible_p): Likewise.
683 (comdat_can_be_unshared_p): Likewise.
684 (comdat_can_be_unshared_p_1): Likewise.
685 (address_taken_from_non_vtable_p): Likewise.
686 (ipa_discover_readonly_nonaddressable_vars): Likewise.
687 (symtab_remove_unreachable_nodes): Likewise.
688 (walk_polymorphic_call_targets): Likewise.
689 (process_references): Likewise.
690 (enqueue_node): Likewise.
691 (has_addr_references_p): Likewise.
692 (cgraph_non_local_node_p_1): Likewise.
693 * is-a.h (varpool_analyze_node): Likewise.
694 * lto-cgraph.c (input_symtab): Likewise.
695 (merge_profile_summaries): Likewise.
696 (input_cgraph_1): Likewise.
697 (input_edge): Likewise.
698 (input_varpool_node): Likewise.
699 (input_node): Likewise.
700 (input_overwrite_node): Likewise.
701 (compute_ltrans_boundary): Likewise.
702 (output_refs): Likewise.
703 (lto_output_varpool_node): Likewise.
704 (lto_output_node): Likewise.
705 (reachable_from_other_partition_p): Likewise.
706 (referenced_from_other_partition_p): Likewise.
707 (lto_output_edge): Likewise.
708 (output_node_opt_summary): Likewise.
709 (add_node_to): Likewise.
710 (reachable_from_this_partition_p): Likewise.
711 (lto_set_symtab_encoder_in_partition): Likewise.
712 (lto_symtab_encoder_in_partition_p): Likewise.
713 (lto_set_symtab_encoder_encode_initializer): Likewise.
714 (lto_symtab_encoder_encode_initializer_p): Likewise.
715 (lto_set_symtab_encoder_encode_body): Likewise.
716 (lto_symtab_encoder_encode_body_p): Likewise.
717 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
719 * lto-streamer-in.c (lto_read_body): Likewise.
720 (fixup_call_stmt_edges): Likewise.
721 (fixup_call_stmt_edges_1): Likewise.
722 * lto-streamer-out.c (produce_symtab): Likewise.
723 (output_symbol_p): Likewise.
724 (write_symbol): Likewise.
725 (lto_output): Likewise.
726 (copy_function): Likewise.
727 (output_function): Likewise.
728 * passes.c (function_called_by_processed_nodes_p): Likewise.
729 (ipa_write_optimization_summaries): Likewise.
730 (ipa_write_summaries): Likewise.
731 (do_per_function_toporder): Likewise.
732 (do_per_function): Likewise.
733 (dump_passes): Likewise.
734 * symtab.c (symtab_semantically_equivalent_p): Likewise.
735 (symtab_nonoverwritable_alias): Likewise.
736 (symtab_nonoverwritable_alias_1): Likewise.
737 (symtab_for_node_and_aliases): Likewise.
738 (symtab_resolve_alias): Likewise.
739 (fixup_same_cpp_alias_visibility): Likewise.
740 (symtab_alias_ultimate_target): Likewise.
741 (symtab_used_from_object_file_p): Likewise.
742 (verify_symtab_base): Likewise.
743 (dump_symtab_base): Likewise.
744 (symtab_node_name): Likewise.
745 (symtab_node_asm_name): Likewise.
746 (symtab_dissolve_same_comdat_group_list): Likewise.
747 (symtab_add_to_same_comdat_group): Likewise.
748 (symtab_unregister_node): Likewise.
749 (symtab_insert_node_to_hashtable): Likewise.
750 (symtab_register_node): Likewise.
751 (unlink_from_assembler_name_hash): Likewise.
752 (insert_to_assembler_name_hash): Likewise.
753 (eq_assembler_name): Likewise.
754 (hash_node_by_assembler_name): Likewise.
756 (hash_node): Likewise.
757 * toplev.c (wrapup_global_declaration_2): Likewise.
758 * trans-mem.c (ipa_tm_execute): Likewise.
759 (ipa_tm_transform_clone): Likewise.
760 (ipa_tm_transform_transaction): Likewise.
761 (ipa_tm_transform_calls_redirect): Likewise.
762 (ipa_tm_insert_gettmclone_call): Likewise.
763 (ipa_tm_insert_irr_call): Likewise.
764 (ipa_tm_create_version): Likewise.
765 (ipa_tm_create_version_alias): Likewise.
766 (ipa_tm_mark_forced_by_abi_node): Likewise.
767 (ipa_tm_mark_force_output_node): Likewise.
768 (ipa_tm_diagnose_tm_safe): Likewise.
769 (ipa_tm_mayenterirr_function): Likewise.
770 (ipa_tm_scan_irr_function): Likewise.
771 (ipa_tm_note_irrevocable): Likewise.
772 (ipa_tm_scan_calls_clone): Likewise.
773 (get_cg_data): Likewise.
774 * tree-eh.c (tree_could_trap_p): Likewise.
775 * tree-emutls.c (ipa_lower_emutls): Likewise.
776 (create_emultls_var): Likewise.
777 (lower_emutls_function_body): Likewise.
778 (gen_emutls_addr): Likewise.
779 (emutls_decl): Likewise.
780 (new_emutls_decl): Likewise.
781 * tree-inline.c (tree_function_versioning): Likewise.
782 (optimize_inline_calls): Likewise.
783 (expand_call_inline): Likewise.
784 (estimate_num_insns): Likewise.
786 (delete_unreachable_blocks_update_callgraph): Likewise.
787 * tree-nested.c (gimplify_all_functions): Likewise.
788 (create_nesting_tree): Likewise.
789 (check_for_nested_with_variably_modified): Likewise.
790 * tree-pretty-print.c (dump_function_header): Likewise.
791 * tree-profile.c (tree_profiling): Likewise.
792 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
793 (modify_function): Likewise.
794 (convert_callers): Likewise.
795 (convert_callers_for_node): Likewise.
796 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
797 (associate_varinfo_to_alias): Likewise.
798 (create_variable_info_for): Likewise.
799 (get_constraint_for_ssa_var): Likewise.
800 * tree-vectorizer.c (increase_alignment): Likewise.
801 * tree.c (find_decls_types_in_var): Likewise.
802 (find_decls_types_in_node): Likewise.
803 (free_lang_data_in_decl): Likewise.
804 * value-prof.c (gimple_ic_transform): Likewise.
805 (gimple_ic): Likewise.
806 (check_ic_target): Likewise.
807 (init_node_map): Likewise.
808 * varasm.c (decl_binds_to_current_def_p): Likewise.
809 (default_binds_local_p_1): Likewise.
810 (dump_tm_clone_pairs): Likewise.
811 (assemble_alias): Likewise.
812 (find_decl): Likewise.
813 (mark_decl_referenced): Likewise.
814 * varpool.c (varpool_for_node_and_aliases): Likewise.
815 (varpool_extra_name_alias): Likewise.
816 (varpool_create_variable_alias): Likewise.
817 (add_new_static_var): Likewise.
818 (varpool_finalize_named_section_flags): Likewise.
819 (varpool_remove_unreferenced_decls): Likewise.
820 (enqueue_node): Likewise.
821 (varpool_assemble_decl): Likewise.
822 (assemble_aliases): Likewise.
823 (varpool_analyze_node): Likewise.
824 (cgraph_variable_initializer_availability): Likewise.
825 (varpool_add_new_variable): Likewise.
826 (ctor_for_folding): Likewise.
827 (dump_varpool_node): Likewise.
828 (varpool_remove_initializer): Likewise.
829 (varpool_remove_node): Likewise.
830 (varpool_node_for_decl): Likewise.
831 (varpool_create_empty_node): Likewise.
832 * config/i386/i386.c (ix86_generate_version_dispatcher_body):
834 (ix86_get_function_versions_dispatcher): Likewise.
836 2013-10-29 David Malcolm <dmalcolm@redhat.com>
838 * cgraph.h (symtab_node_base): Convert to a class;
839 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
840 chain_next/prev from symtab_node_def.
841 (cgraph_node): Inherit from symtab_node; add GTY option
842 tag ("SYMTAB_FUNCTION").
843 (varpool_node): Inherit from symtab_node; add GTY option
844 tag ("SYMTAB_VARIABLE").
845 (symtab_node_def): Remove.
846 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
847 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
848 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
849 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
851 * ipa-ref.h (symtab_node_def): Drop.
852 (symtab_node): Change underlying type from symtab_node_def to
854 (const_symtab_node): Likwise.
856 * is-a.h: Update examples in comment.
858 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
859 (assembler_name_hash): Likewise.
861 2013-10-29 Martin Liska <marxin.liska@gmail.com>
863 * doc/tree-ssa.texi (gimple_phi_result): Document.
864 (gimple_phi_num_args, gimple_phi_arg): Likewise.
865 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
866 (PHI_RESULT, PHI_NUM_ARGS): Remove.
867 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
869 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
871 * expr.h: Revert change and include tree-core.h.
872 * rtl.h: Revert change and don't include tree-core.h.
874 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
876 * config/darwin.c: Include gimple.h.
877 * config/i386/winnt.c: Likewise.
879 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
881 PR tree-optimization/19831
882 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
884 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
886 * tree-outof-ssa.h: Remove include files.
887 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
889 * tree-ssa-coalesce.c: Likewise.
890 * cfgexpand.c: Likewise.
891 * tree-ssa-ter.c: Likewise.
892 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
893 * lto-streamer.h: Likewise.
894 * cgraphbuild.c: Add gimple.h to include list.
895 * data-streamer-in.c: Likewise.
896 * ipa-cp.c: Likewise.
897 * tree-streamer.c: Likewise.
898 * lto-compress.c: Likewise.
899 * ipa-reference.c: Likewise.
900 * data-streamer-out.c: Likewise.
901 * lto-cgraph.c: Likewise.
902 * cgraphclones.c: Likewise.
903 * ipa-utils.c: Likewise.
904 * data-streamer.c: Likewise.
905 * ipa-split.c: Likewise.
906 * lto-section-in.c: Likewise.
907 * tree-streamer-out.c: Likewise.
908 * ipa-prop.c: Likewise.
909 * tree-streamer-in.c: Likewise.
910 * symtab.c: Likewise.
911 * opts-global.c: Likewise.
912 * lto-opts.c: Likewise.
913 * lto-section-out.c: Likewise.
914 * lto-streamer.c: Likewise.
915 * rtl.h: Add tree-core.h to include list.
916 * expr.h: Remove tree-core.h from include list.
917 * gimple.h: Likewise.
918 * ipa-utils.h: Likewise.
919 * streamer-hooks.h: Likewise.
920 * streamer-hooks.c: Include input.h.
922 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
924 * config/arm/arm.c (cortexa7_extra_costs): New table.
925 (arm_cortex_a7_tune): New.
926 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
928 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
930 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
932 2013-10-29 David Malcolm <dmalcolm@redhat.com>
934 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
935 to use autogenerated markers for a class-hierarchy, every class
936 must have a GTY marker.
937 * gengtype.h (struct type): Add linked list of subclasses to
938 the "s" member of the union.
939 (add_subclass): New decl.
940 * gengtype-state.c (read_state_struct_type): Set up subclass
942 * gengtype.c (get_ultimate_base_class): New.
944 (new_structure): Set up subclass linked list.
945 (set_gc_used_type): Propagate usage information to subclasses.
946 (output_mangled_typename): Use get_ultimate_base_class.
947 (walk_subclasses): Use the subclass linked list, avoiding an
948 O(N^2) when writing out all types.
949 (walk_type): Issue an error if the base class is missing a tag,
950 rather than generating bogus C code. Add a gcc_unreachable
951 default case, in case people omit tags from concrete subclasses,
952 or get the values wrong.
953 (write_func_for_structure): Issue an error for subclasses for
954 which the base doesn't have a "desc", since otherwise the
955 autogenerated routines for the base would silently fail to visit
957 (write_root): Use get_ultimate_base_class, tweaking constness of
958 tp to match that function's signature.
960 2013-10-29 David Malcolm <dmalcolm@redhat.com>
962 * doc/gty.texi (GTY Options): Add note about inheritance to
963 description of desc and tag.
964 (Inheritance and GTY): New.
966 2013-10-29 David Malcolm <dmalcolm@redhat.com>
968 * gengtype-parse.c (opts_have): Drop "static" so that
969 we can use this from gengtype.c.
970 * gengtype.c (set_gc_used_type): Mark any base class as used;
971 update field traversal to visit inherited fields.
972 (output_mangled_typename): Convert references to classes within
973 an inheritance hierarchy to reference the ultimate base class,
974 since only it will have gt_ functions.
975 (get_string_option): New.
976 (walk_subclasses): New.
977 (walk_type): Treat GTY structs that have a "desc" as being the
978 root of an inheritance hierarchy. Generate a switch on it
979 within the marking function which walks all subclasses, adding
980 cases for them via walk_subclasses. For subclasses, visit all
981 fields of the type (including inherited ones).
982 (write_func_for_structure): Don't write fns for subclasses, only
983 for the ultimate base class within an inheritance hierarchy.
984 Subclasses-marking will be handled by the base class marking
986 (write_types): Likewise.
987 (write_local_func_for_structure): Likewise.
988 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
989 a "tag" option (and are thus concrete subclasses).
990 (write_root): Use the marker function for the ultimate base class.
991 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
992 (opts_have): Add declaration.
994 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
996 * lra-spills.c (lra_final_code_change): Remove useless move insns
997 originated from moves of pseudos.
999 2013-10-28 Jeff Law <law@redhat.com>
1001 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1003 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1005 * df-scan.c (df_collection_rec): Adjust.
1006 (copy_defs): New constant.
1007 (copy_uses): Likewise.
1008 (copy_eq_uses): Likewise.
1009 (copy_mw): Likewise.
1010 (copy_all): Likewise.
1011 (df_insn_rescan): Adjust.
1012 (df_notes_rescan): Likewise.
1013 (df_swap_refs): Likewise.
1014 (df_sort_and_compress_refs): Likewise.
1015 (df_sort_and_compress_mws): Likewise.
1016 (df_install_refs): Likewise.
1017 (df_install_mws): Likewise.
1018 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1020 (df_bb_refs_record): Adjust.
1021 (df_record_entry_block_defs): Likewise.
1022 (df_record_exit_block_defs): Likewise.
1023 (df_refs_verify): Likewise.
1024 (df_mws_verify): Likewise.
1025 (df_insn_refs_verify): Likewise.
1026 (df_bb_verify): Likewise.
1027 * ipa-pure-const.c (finish_state): Remove.
1028 (propagate): Adjust.
1029 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1030 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1031 var-tracking.c: Adjust.
1032 * vec.c (stack_vecs): Remove.
1033 (register_stack_vec): Likewise.
1034 (stack_vec_register_index): Likewise.
1035 (unregister_stack_vec): Likewise.
1036 * vec.h (struct va_stack): Remove.
1037 (struct vec<T, A, vl_ptr>): Specialize as
1038 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1039 allocation strategy compatable with the vl_ptr layout.
1040 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1041 specialization anyway.
1042 (class stack_vec): New class.
1043 (vec_stack_alloc): Remove.
1044 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1046 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1047 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1048 Sergey Lega <sergey.s.lega@intel.com>
1049 Anna Tikhonova <anna.tikhonova@intel.com>
1050 Ilya Tocar <ilya.tocar@intel.com>
1051 Andrey Turetskiy <andrey.turetskiy@intel.com>
1052 Ilya Verbin <ilya.verbin@intel.com>
1053 Kirill Yukhin <kirill.yukhin@intel.com>
1054 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1056 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1057 (*prefetch_avx512pf_<mode>): New.
1058 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1059 (avx512f_maskcmp<mode>3): Ditto.
1060 (vashrv16si3): Ditto.
1062 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1063 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1064 Sergey Lega <sergey.s.lega@intel.com>
1065 Anna Tikhonova <anna.tikhonova@intel.com>
1066 Ilya Tocar <ilya.tocar@intel.com>
1067 Andrey Turetskiy <andrey.turetskiy@intel.com>
1068 Ilya Verbin <ilya.verbin@intel.com>
1069 Kirill Yukhin <kirill.yukhin@intel.com>
1070 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1072 * config/i386/i386.md (any_truncate): New.
1073 (trunsuffix): Ditto.
1074 * config/i386/predicates.md (const_8_to_9_operand): New.
1075 (const_10_to_11_operand): Ditto.
1076 (const_12_to_13_operand): Ditto.
1077 (const_14_to_15_operand): Ditto.
1078 (const_16_to_19_operand): Ditto.
1079 (const_20_to_23_operand): Ditto.
1080 (const_24_to_27_operand): Ditto.
1081 (const_28_to_31_operand): Ditto.
1082 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1083 (cvtusi2<ssescalarmodesuffix>32): New.
1084 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1085 (ufloatv16siv16sf2): Ditto.
1086 (avx512f_fix_notruncv16sfv16si): Ditto.
1087 (avx512f_ufix_notruncv16sfv16si): Ditto.
1088 (avx512f_vcvtss2usi): Ditto.
1089 (avx512f_vcvtss2usiq): Ditto.
1090 (avx512f_vcvttss2usi): Ditto.
1091 (avx512f_vcvttss2usiq): Ditto.
1092 (avx512f_vcvtsd2usi): Ditto.
1093 (avx512f_vcvtsd2usiq): Ditto.
1094 (avx512f_vcvttsd2usi): Ditto.
1095 (avx512f_vcvttsd2usiq): Ditto.
1096 (ufloatv8siv8df): Ditto.
1097 (avx512f_cvtdq2pd512_2): Ditto.
1098 (avx512f_cvtpd2dq512): Ditto.
1099 (avx512f_ufix_notruncv8dfv8si): Ditto.
1100 (avx512f_cvtpd2ps512): Ditto.
1101 (vec_unpacks_lo_v16sf): Ditto.
1102 (vec_unpacks_hi_v16sf): Ditto.
1103 (vec_unpacks_float_hi_v16si): Ditto.
1104 (vec_unpacks_float_lo_v16si): Ditto.
1105 (avx512f_unpckhps512): Ditto.
1106 (avx512f_unpcklps512): Ditto.
1107 (avx512f_movshdup512): Ditto.
1108 (avx512f_movsldup512): Ditto.
1109 (vec_extract_lo_v32hi): Ditto.
1110 (vec_extract_hi_v32hi): Ditto.
1111 (vec_extract_lo_v64qi): Ditto.
1112 (vec_extract_hi_v64qi): Ditto.
1113 (avx512f_unpckhpd512): Ditto.
1114 (avx512f_movddup512): Ditto.
1115 (avx512f_unpcklpd512): Ditto.
1116 (*avx512f_unpcklpd512): Ditto.
1117 (avx512f_shufps512_1): Ditto.
1118 (avx512f_shufpd512_1): Ditto.
1119 (avx512f_interleave_highv8di): Ditto.
1120 (avx512f_interleave_lowv8di): Ditto.
1121 (PMOV_DST_MODE): Ditto.
1122 (pmov_src_mode): Ditto.
1123 (pmov_src_lower): Ditto.
1125 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1126 (*avx512f_<code>v8div16qi2): Ditto.
1127 (*avx512f_<code>v8div16qi2_store): Ditto.
1128 (vec_widen_umult_even_v16si): Ditto.
1129 (*vec_widen_umult_even_v16si): Ditto.
1130 (vec_widen_smult_even_v16si): Ditto.
1131 (*vec_widen_smult_even_v16si): Ditto.
1132 (avx512f_interleave_highv16si): Ditto.
1133 (avx512f_interleave_lowv16si): Ditto.
1134 (avx512f_<code>v16qiv16si2): Ditto.
1135 (avx512f_<code>v16hiv16si2): Ditto.
1136 (avx512f_<code>v8qiv8di2): Ditto.
1137 (avx512f_<code>v8hiv8di2): Ditto.
1138 (avx512f_<code>v8siv8di2): Ditto.
1139 (avx512cd_maskb_vec_dupv8di): Ditto.
1140 (avx512cd_maskw_vec_dupv16si): Ditto.
1141 (avx512f_vcvtph2ps512): Ditto.
1142 (avx512f_vcvtps2ph512): Ditto.
1143 (VEC_EXTRACT_MODE): Extened with wider modes.
1144 (VEC_PERM_AVX2): Ditto.
1145 (VEC_PERM_CONST): Ditto.
1148 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1150 * config/arc/arc.c (arc_ccfsm_post_advance):
1151 Add comment about TYPE_RETURN.
1153 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1155 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1156 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1157 (strip_offset): Convert offset to unsigned number.
1159 2013-10-27 Tom de Vries <tom@codesourcery.com>
1161 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1162 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1164 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1166 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1167 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1168 (TARGET_BIG_ENDIAN): New macro.
1169 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1170 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1171 * config/sh/sh.c: Likewise.
1172 * config/sh/sh.h: Likewise.
1174 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1176 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1177 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1179 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1181 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1183 (cmpstr_t, cmpstrsi): Fix formatting.
1185 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1188 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1189 addressing, do not use general_operand for memory operands.
1191 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1194 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1195 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1197 2013-10-26 Jeff Law <law@redhat.com>
1199 * predict.c (PRED_MUDFLAP): Remove.
1200 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1202 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1203 (OBJS): Remove tree-nomudflap.o
1204 (GTFILES): Remove tree-mudflap.c
1205 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1206 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1207 (mfwrap_spec, mflib_spec): Likewise.
1208 (cpp_unique_options, cc1_options, static_specs): Likewise.
1209 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1210 * passes.def: Likewise.
1211 * toplev.c (compile_file, process_options): Likewise.
1212 * tree-inline.c (copy_tree_r): Likewise.
1213 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
1214 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
1215 (build_constant_desc, output_constant_def_contents): Likewise.
1216 (categorize_decl_for_section): Likewise.
1217 * tree-mudflap.c: Removed.
1218 * tree-mudflap.h: Removed.
1219 * tree-nomudflap.c: Removed.
1220 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
1221 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
1222 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1223 * config/sol2.h (MFLIB_SPEC): Likewise.
1224 * doc/install.texi: Remove mudflap references.
1225 * doc/passes.texi: Similarly.
1226 * doc/sourcebuild.texi: Similarly.
1227 * doc/invoke.texi: Remove mudlfap related options.
1229 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1231 PR rtl-optimization/58759
1232 * lra-constraints.c (lra_constraints): Remove wrong condition to
1233 remove insn setting up an equivalent pseudo.
1235 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1237 * config/rs6000/rs6000-protos.h
1238 (rs6000_secondary_memory_needed_mode): New prototype.
1239 * config/rs6000/rs6000.c: Include ira.h.
1240 (TARGET_LRA_P): Redefine.
1241 (rs6000_legitimate_offset_address_p): Call
1242 legitimate_constant_pool_address_p in strict mode for LRA.
1243 (rs6000_legitimate_address_p): Ditto.
1244 (legitimate_lo_sum_address_p): Add code for LRA.
1245 Use lra_in_progress.
1246 (rs6000_emit_move): Add LRA version of code to generate load/store
1248 (rs6000_secondary_memory_needed_mode): New.
1249 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
1250 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
1251 constants, memory, and FP registers.
1252 (rs6000_lra_p): New.
1253 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New
1255 * config/rs6000/rs6000.opt (mlra): New option.
1256 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1258 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
1260 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
1261 has_single_use () and not do the conversion if has_single_use ()
1262 returns false for the multiplication result.
1264 2013-10-25 David Malcolm <dmalcolm@redhat.com>
1266 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
1267 to "NODE", since this works on a "tree", not an
1269 (CONSTANT_CLASS_P): Likewise.
1272 (INDIRECT_REF_P): Likewise.
1273 (REFERENCE_CLASS_P): Likewise.
1274 (COMPARISON_CLASS_P): Likewise.
1275 (UNARY_CLASS_P): Likewise.
1276 (BINARY_CLASS_P): Likewise.
1277 (STATEMENT_CLASS_P): Likewise.
1278 (VL_EXP_CLASS_P): Likewise.
1279 (EXPRESSION_CLASS_P): Likewise.
1280 (IS_TYPE_OR_DECL_P): Likewise.
1282 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
1284 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
1285 ADDR_EXPR in the defining statement.
1287 2013-10-25 Richard Biener <rguenther@suse.de>
1289 PR tree-optimization/58626
1290 * tree-loop-distribution.c (enum rdg_dep_type): Remove
1291 anti_dd, output_dd and input_dd.
1292 (struct rdg_edge): Remove level and relation members.
1293 (RDGE_LEVEL, RDGE_RELATION): Remove.
1294 (dot_rdg_1): Adjust.
1295 (create_rdg_edge_for_ddr): Remove.
1296 (create_rdg_edges_for_scalar): Adjust.
1297 (create_edge_for_control_dependence): Likewise.
1298 (create_rdg_edges): Split into ...
1299 (create_rdg_flow_edges): ... this
1300 (create_rdg_cd_edges): ... and this.
1302 (build_rdg): Likewise, do not compute data dependences or
1304 (pg_add_dependence_edges): New function.
1306 (distribute_loop): First apply all non-dependence based
1307 partition mergings. Then compute dependences between partitions
1308 and merge and order partitions according to them.
1310 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1312 PR rtl-optimization/58831
1313 * alias.c (init_alias_analysis): At the beginning of each iteration, set
1314 the reg_seen[N] bit if static_reg_base_value[N] is non-null.
1316 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1318 * recog.c (search_ofs): New static variable moved from...
1319 (peep2_find_free_register): ...here.
1320 (peephole2_optimize): Initialize it.
1322 2013-10-25 Tobias Burnus <burnus@net-b.de>
1324 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
1326 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
1328 * config/i386/i386.h (TARGET_MPX): New define.
1329 (TARGET_MPX_P): Ditto.
1331 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1333 * config/i386/constraints.md (B): New.
1336 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
1337 * config/i386/i386-modes.def (BND32): New.
1339 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
1340 * config/i386/i386.c (isa_opts): Add mmpx.
1341 (regclass_map): Add bound registers.
1342 (dbx_register_map): Likewise.
1343 (dbx64_register_map): Likewise.
1344 (svr4_dbx_register_map): Likewise.
1346 (ix86_option_override_internal): Support MPX ISA.
1347 (ix86_conditional_register_usage): Support bound registers.
1348 (print_reg): Likewise.
1349 (ix86_code_end): Add MPX bnd prefix.
1350 (output_set_got): Likewise.
1351 (ix86_output_call_insn): Likewise.
1352 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
1353 (ix86_print_operand_punct_valid_p): Likewise.
1354 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
1356 (ix86_class_likely_spilled_p): Add bound regs support.
1357 (ix86_hard_regno_mode_ok): Likewise.
1358 (x86_order_regs_for_local_alloc): Likewise.
1359 (ix86_bnd_prefixed_insn_p): New.
1360 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
1361 (FIXED_REGISTERS): Add bound registers.
1362 (CALL_USED_REGISTERS): Likewise.
1363 (REG_ALLOC_ORDER): Likewise.
1364 (HARD_REGNO_NREGS): Likewise.
1366 (VALID_BND_REG_MODE): New.
1367 (FIRST_BND_REG): New.
1368 (LAST_BND_REG): New.
1369 (reg_class): Add BND_REGS.
1370 (REG_CLASS_NAMES): Likewise.
1371 (REG_CLASS_CONTENTS): Likewise.
1373 (ANY_BND_REG_P): New.
1375 (HI_REGISTER_NAMES): Add bound registers.
1376 * config/i386/i386.md (UNSPEC_BNDMK): New.
1377 (UNSPEC_BNDMK_ADDR): New.
1378 (UNSPEC_BNDSTX): New.
1379 (UNSPEC_BNDLDX): New.
1380 (UNSPEC_BNDLDX_ADDR): New.
1381 (UNSPEC_BNDCL): New.
1382 (UNSPEC_BNDCU): New.
1383 (UNSPEC_BNDCN): New.
1384 (UNSPEC_MPX_FENCE): New.
1387 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
1388 (length_immediate): Likewise.
1389 (prefix_0f): Likewise.
1391 (prefix_rep): Check for bnd prefix.
1392 (length_nobnd): New.
1393 (length): Use length_nobnd if specified.
1398 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
1401 (simple_return_internal): Likewise.
1402 (simple_return_pop_internal): Likewise.
1403 (*indirect_jump): Add MPX bnd prefix.
1404 (*tablejump_1): Likewise.
1405 (simple_return_internal_long): Likewise.
1406 (simple_return_indirect_internal): Likewise.
1410 (*mov<mode>_internal_mpx): New.
1411 (<mode>_<bndcheck>): New.
1412 (*<mode>_<bndcheck>): New.
1417 * config/i386/predicates.md (lea_address_operand): Rename to...
1418 (address_no_seg_operand): ... this.
1419 (address_mpx_no_base_operand): New.
1420 (address_mpx_no_index_operand): New.
1421 (bnd_mem_operator): New.
1422 * config/i386/i386.opt (mmpx): New.
1423 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
1424 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
1426 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1428 * mode-classes.def (MODE_POINTER_BOUNDS): New.
1429 * tree.def (POINTER_BOUNDS_TYPE): New.
1430 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
1431 (POINTER_BOUNDS_MODE): New.
1432 (make_pointer_bounds_mode): New.
1433 * machmode.h (POINTER_BOUNDS_MODE_P): New.
1434 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
1435 (layout_type): Support POINTER_BOUNDS_TYPE.
1436 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
1437 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
1438 (type_contains_placeholder_1): Likewise.
1439 * tree.h (POINTER_BOUNDS_TYPE_P): New.
1440 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
1441 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
1443 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
1445 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
1446 mode of address computation when calling memory_address_addr_space.
1448 2013-08-24 Richard Henderson <rth@twiddle.net>
1451 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
1452 instead of create_convert_operand_to.
1453 (maybe_emit_sync_lock_test_and_set): Likewise.
1454 (expand_atomic_compare_and_swap): Likewise.
1455 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
1457 2013-08-24 Sriraman Tallam <tmsriram@google.com>
1459 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
1460 Change param_is to use the struct and not the pointer to the struct.
1462 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
1464 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
1465 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
1466 Move to targhooks.c.
1467 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
1468 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
1469 Relocate from gimplify.c.
1470 * targhooks.h: Add 2 prototypes.
1471 * tree.h: Delete 2 prototypes.
1473 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
1475 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
1476 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
1478 2013-10-24 Cong Hou <congh@google.com>
1480 * convert.c (convert_to_real): Guard those unsafe math function
1481 convertions with flag_unsafe_math_optimizations. Handle sqrt()
1484 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
1487 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
1489 (cgraph_create_edge): Use the new argument.
1490 (cgraph_create_indirect_edge): Likewise.
1492 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
1494 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
1496 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
1497 changes statep->state.
1499 2013-10-24 Tobias Burnus <burnus@net-b.de>
1502 * tree-cfg.c (replace_loop_annotate): New function.
1503 (execute_build_cfg): Call it.
1504 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
1505 * internal-fn.c (expand_ANNOTATE): New function.
1506 * internal-fn.def (ANNOTATE): Define as new internal function.
1507 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
1508 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
1509 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
1510 * doc/extend.texi (Pragmas): Document #pragma ivdep.
1511 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
1513 2013-10-17 Ian Bolton <ian.bolton@arm.com>
1514 Marcus Shawcroft <marcus.shawcroft@arm.com>
1516 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
1517 Special case reload SP+C into none GENERAL_REGS.
1519 2013-10-24 Michael Matz <matz@suse.de>
1521 * gengtype.c (is_file_equal): Check that files will be same length.
1523 2013-10-25 Christian Bruel <christian.bruel@st.com>
1525 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
1526 * gcc/config/sh/t-sh (sh-mem.o): New rule.
1527 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
1528 (sh_expand_cmpstr): New function.
1529 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
1530 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
1531 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
1532 (rotlhi3_8): Rename.
1534 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1536 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
1537 * Makefile.in (CXXDEPMODE): Assign and change users.
1538 (CCDEPMODE): Delete.
1539 * configure: Regenerate.
1541 2013-10-23 David Malcolm <dmalcolm@redhat.com>
1543 * gengtype-parse.c (require_without_advance): New.
1544 (type): For GTY-marked types that are not GTY((user)), parse any
1545 base classes, requiring them to be single-inheritance, and not
1546 be templates. For non-GTY-marked types and GTY((user)),
1547 continue to skip over any C++ inheritance specification.
1548 * gengtype-state.c (state_writer::write_state_struct_type):
1549 Write base class of type (if any).
1550 (read_state_struct_type): Read base class of type (if any).
1551 * gengtype.c (new_structure): Add a "base_class" parameter.
1552 (create_optional_field_): Update for new parameter to new_structure.
1553 (adjust_field_rtx_def): Likewise.
1554 (adjust_field_tree_exp): Likewise.
1555 * gengtype.h (struct type): Add "base_class" field to the s
1557 (new_structure): Add "base" parameter.
1559 2013-10-23 Sriraman Tallam <tmsriram@google.com>
1562 * config/i386/i386.c (ix86_option_override_internal):
1563 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
1564 (ix86_valid_target_attribute_tree):
1565 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
1566 Change TARGET_SSE to TARGET_SSE_P (opts->...)
1568 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
1570 * tree-ssa-loop.h: Remove include files.
1571 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
1572 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
1573 * graphite-clast-to-gimple.c: Likewise.
1574 * graphite-scop-detection.c: Likewise.
1575 * graphite-sese-to-poly.c: Likewise.
1576 * ipa-inline-analysis.c: Likewise.
1577 * ipa-pure-const.c: Likewise.
1578 * loop-init.c: Likewise.
1579 * passes.c: Likewise.
1580 * predict.c: Likewise.
1581 * tree-cfg.c: Likewise.
1582 * tree-cfgcleanup.c: Likewise.
1583 * tree-chrec.c: Likewise.
1584 * tree-data-ref.c: Likewise.
1585 * tree-loop-distribution.c: Likewise.
1586 * tree-parloops.c: Likewise.
1587 * tree-predcom.c: Likewise.
1588 * tree-scalar-evolution.c: Likewise.
1589 * tree-ssa-address.c: Likewise.
1590 * tree-ssa.c: Likewise.
1591 * tree-ssa-dce.c: Likewise.
1592 * tree-ssa-loop.c: Likewise.
1593 * tree-ssa-loop-im.c: Likewise.
1594 * tree-ssa-loop-ivcanon.c: Likewise.
1595 * tree-ssa-loop-ivopts.c: Likewise.
1596 * tree-ssa-loop-manip.c: Likewise.
1597 * tree-ssa-loop-niter.c: Likewise.
1598 * tree-ssa-loop-prefetch.c: Likewise.
1599 * tree-ssa-loop-unswitch.c: Likewise.
1600 * tree-ssa-reassoc.c: Likewise.
1601 * tree-vect-data-refs.c: Likewise.
1602 * tree-vect-loop.c: Likewise.
1603 * tree-vect-loop-manip.c: Likewise.
1604 * tree-vectorizer.c: Likewise.
1605 * tree-vect-stmts.c: Likewise.
1606 * tree-vrp.c: Likewise.
1608 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1610 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
1612 2013-10-23 Jakub Jelinek <jakub@redhat.com>
1614 PR tree-optimization/58775
1615 PR tree-optimization/58791
1616 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
1617 (insert_stmt_after): Rewritten, don't move the stmt, but really
1619 (get_stmt_uid_with_default): Remove.
1620 (build_and_add_sum): Use insert_stmt_after and
1621 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
1622 (update_range_test): Set uid on stmts added by
1623 force_gimple_operand_gsi. Don't immediately modify statements
1624 in inter-bb optimization, just update oe->op values.
1625 (optimize_range_tests): Return bool whether any changed have been made.
1626 (update_ops): New function.
1627 (struct inter_bb_range_test_entry): New type.
1628 (maybe_optimize_range_tests): Perform statement changes here.
1629 (not_dominated_by, appears_later_in_bb, get_def_stmt,
1630 ensure_ops_are_available): Remove.
1631 (find_insert_point): Rewritten.
1632 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
1633 return LHS of the (new resp. old) stmt. Don't call
1634 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
1635 instead of last, move new stmt at the right place.
1636 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
1637 (negate_value): Likewise. Set uids.
1638 (break_up_subtract_bb): Initialize uids.
1639 (reassociate_bb): Adjust rewrite_expr_tree caller.
1640 (do_reassoc): Don't call renumber_gimple_stmt_uids.
1642 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
1645 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
1646 TARGET_32BIT final condition.
1647 (mulsi3_internal2 and splitter): Same.
1649 2013-10-23 Jeff Law <law@redhat.com>
1651 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
1652 through joiner blocks with abnormal outgoing edges.
1654 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
1655 Add parameter JOINERS, to allow/disallow threading through joiner
1657 (thread_block): New. Call thread_block_1.
1658 (mark_threaded_blocks): Remove code to filter out certain cases
1659 of threading through joiner blocks.
1660 (thread_through_all_blocks): Document how we can have a dangling
1661 edge AUX field and clear it.
1663 2013-10-23 Ian Lance Taylor <iant@google.com>
1665 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
1666 (C++ Dialect Options): Likewise.
1667 (Optimize Options): Likewise.
1669 2013-10-23 Tom de Vries <tom@codesourcery.com>
1671 PR tree-optimization/58805
1672 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
1674 2013-10-23 Jakub Jelinek <jakub@redhat.com>
1676 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
1677 sequence based on get_range_info returned range.
1679 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
1681 * tree-ssa.h: Remove all #include's
1682 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
1683 * alias.c: Move required includes from tree-ssa.h.
1685 * builtins.c: Likewise.
1686 * calls.c: Likewise.
1687 * cfgexpand.c: Likewise.
1688 * cfghooks.c: Likewise.
1689 * cfgloop.c: Likewise.
1690 * cfgloopmanip.c: Likewise.
1691 * cgraph.c: Likewise.
1692 * cgraphbuild.c: Likewise.
1693 * cgraphclones.c: Likewise.
1694 * cgraphunit.c: Likewise.
1696 * except.c: Likewise.
1698 * final.c: Likewise.
1699 * fold-const.c: Likewise.
1700 * ggc-page.c: Likewise.
1701 * gimple-builder.c: Likewise.
1702 * gimple-fold.c: Likewise.
1703 * gimple-iterator.c: Likewise.
1704 * gimple-low.c: Likewise.
1705 * gimple-pretty-print.c: Likewise.
1706 * gimple-ssa-strength-reduction.c: Likewise.
1707 * gimple-streamer-in.c: Likewise.
1708 * gimple-streamer-out.c: Likewise.
1709 * gimplify.c: Likewise.
1710 * graphite-blocking.c: Likewise.
1711 * graphite-clast-to-gimple.c: Likewise.
1712 * graphite-dependences.c: Likewise.
1713 * graphite-interchange.c: Likewise.
1714 * graphite-optimize-isl.c: Likewise.
1715 * graphite-poly.c: Likewise.
1716 * graphite-scop-detection.c: Likewise.
1717 * graphite-sese-to-poly.c: Likewise.
1718 * graphite.c: Likewise.
1719 * ipa-cp.c: Likewise.
1720 * ipa-inline-analysis.c: Likewise.
1721 * ipa-inline-transform.c: Likewise.
1722 * ipa-inline.c: Likewise.
1723 * ipa-prop.c: Likewise.
1724 * ipa-pure-const.c: Likewise.
1725 * ipa-reference.c: Likewise.
1726 * ipa-split.c: Likewise.
1727 * ipa-utils.c: Likewise.
1728 * loop-init.c: Likewise.
1729 * lto-cgraph.c: Likewise.
1730 * lto-section-in.c: Likewise.
1731 * lto-section-out.c: Likewise.
1732 * lto-streamer-in.c: Likewise.
1733 * lto-streamer-out.c: Likewise.
1734 * lto-streamer.c: Likewise.
1735 * omp-low.c: Likewise.
1736 * passes.c: Likewise.
1737 * predict.c: Likewise.
1738 * print-tree.c: Likewise.
1739 * profile.c: Likewise.
1741 * targhooks.c: Likewise.
1742 * tracer.c: Likewise.
1743 * trans-mem.c: Likewise.
1744 * tree-call-cdce.c: Likewise.
1745 * tree-cfg.c: Likewise.
1746 * tree-cfgcleanup.c: Likewise.
1747 * tree-chrec.c: Likewise.
1748 * tree-complex.c: Likewise.
1749 * tree-data-ref.c: Likewise.
1750 * tree-dfa.c: Likewise.
1751 * tree-eh.c: Likewise.
1752 * tree-emutls.c: Likewise.
1753 * tree-if-conv.c: Likewise.
1754 * tree-inline.c: Likewise.
1755 * tree-into-ssa.c: Likewise.
1756 * tree-loop-distribution.c: Likewise.
1757 * tree-mudflap.c: Likewise.
1758 * tree-nested.c: Likewise.
1759 * tree-nrv.c: Likewise.
1760 * tree-object-size.c: Likewise.
1761 * tree-outof-ssa.c: Likewise.
1762 * tree-parloops.c: Likewise.
1763 * tree-phinodes.c: Likewise.
1764 * tree-predcom.c: Likewise.
1765 * tree-pretty-print.c: Likewise.
1766 * tree-profile.c: Likewise.
1767 * tree-scalar-evolution.c: Likewise.
1768 * tree-sra.c: Likewise.
1769 * tree-ssa-address.c: Likewise.
1770 * tree-ssa-alias.c: Likewise.
1771 * tree-ssa-ccp.c: Likewise.
1772 * tree-ssa-coalesce.c: Likewise.
1773 * tree-ssa-copy.c: Likewise.
1774 * tree-ssa-copyrename.c: Likewise.
1775 * tree-ssa-dce.c: Likewise.
1776 * tree-ssa-dom.c: Likewise.
1777 * tree-ssa-dse.c: Likewise.
1778 * tree-ssa-forwprop.c: Likewise.
1779 * tree-ssa-ifcombine.c: Likewise.
1780 * tree-ssa-live.c: Likewise.
1781 * tree-ssa-loop-ch.c: Likewise.
1782 * tree-ssa-loop-im.c: Likewise.
1783 * tree-ssa-loop-ivcanon.c: Likewise.
1784 * tree-ssa-loop-ivopts.c: Likewise.
1785 * tree-ssa-loop-manip.c: Likewise.
1786 * tree-ssa-loop-niter.c: Likewise.
1787 * tree-ssa-loop-prefetch.c: Likewise.
1788 * tree-ssa-loop-unswitch.c: Likewise.
1789 * tree-ssa-loop.c: Likewise.
1790 * tree-ssa-math-opts.c: Likewise.
1791 * tree-ssa-operands.c: Likewise.
1792 * tree-ssa-phiopt.c: Likewise.
1793 * tree-ssa-phiprop.c: Likewise.
1794 * tree-ssa-pre.c: Likewise.
1795 * tree-ssa-propagate.c: Likewise.
1796 * tree-ssa-reassoc.c: Likewise.
1797 * tree-ssa-sccvn.c: Likewise.
1798 * tree-ssa-sink.c: Likewise.
1799 * tree-ssa-strlen.c: Likewise.
1800 * tree-ssa-structalias.c: Likewise.
1801 * tree-ssa-tail-merge.c: Likewise.
1802 * tree-ssa-ter.c: Likewise.
1803 * tree-ssa-threadedge.c: Likewise.
1804 * tree-ssa-threadupdate.c: Likewise.
1805 * tree-ssa-uncprop.c: Likewise.
1806 * tree-ssa-uninit.c: Likewise.
1807 * tree-ssa.c: Likewise.
1808 * tree-ssanames.c: Likewise.
1809 * tree-stdarg.c: Likewise.
1810 * tree-streamer-in.c: Likewise.
1811 * tree-switch-conversion.c: Likewise.
1812 * tree-tailcall.c: Likewise.
1813 * tree-vect-data-refs.c: Likewise.
1814 * tree-vect-generic.c: Likewise.
1815 * tree-vect-loop-manip.c: Likewise.
1816 * tree-vect-loop.c: Likewise.
1817 * tree-vect-patterns.c: Likewise.
1818 * tree-vect-slp.c: Likewise.
1819 * tree-vect-stmts.c: Likewise.
1820 * tree-vectorizer.c: Likewise.
1821 * tree-vrp.c: Likewise.
1824 * value-prof.c: Likewise.
1825 * var-tracking.c: Likewise.
1826 * varpool.c: Likewise.
1827 * vtable-verify.c: Likewise.
1829 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1831 * config/tilegx/tilegx.c: Include "tree.h".
1833 2013-10-23 Jakub Jelinek <jakub@redhat.com>
1835 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
1836 the info, not after it.
1837 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
1838 dump_ssaname_info call.
1839 (pp_gimple_stmt_1): Adjust caller.
1840 (dump_phi_nodes): Likewise. Don't print "# " here.
1842 2013-10-22 Jan Hubicka <jh@suse.cz>
1844 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
1846 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
1847 * i386.c (expand_small_movmem_or_setmem): New function.
1848 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
1850 (alg_usable_p): Add support for value ranges; cleanup.
1851 (ix86_expand_set_or_movmem): Add support for misaligned moves.
1853 2013-10-22 Sterling Augustine <saugustine@google.com>
1855 * doc/invoke.texi: Document -ggnu-pubnames.
1856 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
1858 * dwarf2out.c: Include gdb/gdb-index.h.
1859 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
1860 debug_generate_pub_sections.
1861 (is_java, output_pubtables, output_pubname): New functions.
1862 (include_pubname_in_output): Handle debug_generate_pub_sections at
1864 (size_of_pubnames): Use new local space_for_flags based on
1865 debug_generate_pub_sections.
1866 (output_pubnames): Unify pubnames and pubtypes output logic.
1867 Genericize comments. Call output_pubname.
1868 (dwarf2out_finish): Move logic to output_pubnames and call it.
1870 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
1873 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
1874 Remove CCCmode handling.
1875 <case LTU>: Return 'c' suffix for CCCmode.
1876 <case GEU>: Return 'nc' suffix for CCCmode.
1877 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
1878 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
1879 (*sub<mode>3_cc_overflow): Ditto.
1880 (*subsi3_zext_cc_overflow): Ditto.
1882 2013-10-22 Steve Ellcey <sellcey@mips.com>
1884 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
1885 (AND (NOT OP1) (NOT OP2)).
1887 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
1889 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
1890 meaning of merge-high and merge-low masks for little endian; avoid
1891 use of vector-pack masks for little endian for mismatched modes.
1893 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1895 * config/tilepro/tilepro.c: Include "tree.h".
1897 2013-10-22 Andreas Schwab <schwab@suse.de>
1899 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
1902 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
1904 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
1905 [Ed Smith-Rowland]): New entries.
1906 ([Stephen M. Webb]): Update.
1908 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
1910 * tree-ssa-ter.h: Remove duplicate copy of file contents.
1912 2013-10-21 Marek Polacek <polacek@redhat.com>
1915 * fold-const.c (fold_range_test): Return 0 if the type is not
1918 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
1920 * system.h: Move hwint.h include further down.
1921 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
1922 gcc_checking_asserts.
1923 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
1925 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
1927 Fix volatile issues in optimize_bit_field_compare.
1928 * fold-const.c (optimize_bit_field_compare): Bail out if
1929 lvolatilep or rvolatilep.
1931 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
1933 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
1934 and get_inner_reference returning different pmode for non-volatile
1935 bit-field members dependent on flag_strict_volatile_bitfields.
1936 * stor-layout.c (layout_decl): Remove special handling of
1937 flag_strict_volatile_bitfields.
1938 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
1939 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
1941 2013-10-21 Paulo Matos <pmatos@broadcom.com>
1943 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
1946 2013-10-21 Jeff Law <law@redhat.com>
1948 * tree-ssa-threadedge.c (thread_through_normal_block): New
1949 argument VISITED. Remove VISISTED as a local variable. When we
1950 have a threadable jump, verify the destination of the jump has not
1952 (thread_across_edge): Allocate VISITED bitmap once at function
1953 scope and use it throughout. Make sure to set appropriate bits in
1954 VISITED for E (start of jump thread path).
1955 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
1956 through a joiner if any edge on the path has a recorded jump thread.
1958 2013-10-21 Ian Lance Taylor <iant@google.com>
1960 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
1961 don't imply that attributes can solve all problems.
1962 (Directory Options): Fix typo.
1964 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1966 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
1967 extend and extend_add.
1969 2013-10-21 Richard Biener <rguenther@suse.de>
1971 PR tree-optimization/58794
1972 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
1973 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
1975 2013-10-21 Richard Biener <rguenther@suse.de>
1978 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
1979 to (T) X for sign-changing conversions (or no conversion).
1981 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
1983 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
1985 2013-10-20 Jan Hubicka <jh@suse.cz>
1987 * config/i386/i386-tune.def: Add comment; organize into categories
1989 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
1991 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
1992 argument. Update function comment.
1993 (expand_set_or_movmem_via_rep): New function combining
1994 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
1995 (expand_movmem_via_rep_mov): Remove.
1996 expand_setmem_via_rep_stos): Remove.
1997 (expand_movmem_epilogue): Update calls correspondingly.
1998 (expand_setmem_epilogue_via_loop): Likewise.
2000 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2001 (expand_set_or_movmem_prologue): New function combining
2002 expand_movmem_prologue and expand_setmem_prologue.
2003 (expand_movmem_prologue): Remove.
2004 (expand_setmem_prologue): Remove.
2005 (expand_set_or_movmem_constant_prologue): New function combining
2006 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2007 (expand_constant_movmem_prologue): Remove.
2008 (expand_constant_setmem_prologue): Remove.
2009 (promote_duplicated_reg): Allow vector-const0 value.
2010 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2011 and ix86_expand_setmem.
2012 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2013 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2015 2013-10-21 Diego Novillo <dnovillo@google.com>
2017 * asan.c: Include tree.h
2018 * bb-reorder.c: Likewise.
2019 * cfgcleanup.c: Likewise.
2020 * cfgloopmanip.c: Likewise.
2021 * data-streamer-in.c: Likewise.
2022 * data-streamer-out.c: Likewise.
2023 * data-streamer.c: Likewise.
2024 * dwarf2cfi.c: Likewise.
2025 * graphite-blocking.c: Likewise.
2026 * graphite-clast-to-gimple.c: Likewise.
2027 * graphite-dependences.c: Likewise.
2028 * graphite-interchange.c: Likewise.
2029 * graphite-optimize-isl.c: Likewise.
2030 * graphite-poly.c: Likewise.
2031 * graphite-scop-detection.c: Likewise.
2032 * graphite-sese-to-poly.c: Likewise.
2033 * graphite.c: Likewise.
2034 * ipa-devirt.c: Likewise.
2035 * ipa-profile.c: Likewise.
2038 * loop-init.c: Likewise.
2039 * loop-unroll.c: Likewise.
2040 * lower-subreg.c: Likewise.
2041 * lto/lto-object.c: Likewise.
2042 * recog.c: Likewise.
2043 * reginfo.c: Likewise.
2044 * tree-loop-distribution.c: Likewise.
2045 * tree-parloops.c: Likewise.
2046 * tree-ssa-strlen.c: Likewise.
2047 * tree-streamer.c: Likewise.
2048 * value-prof.c: Likewise.
2049 * target-globals.c: Likewise.
2050 * expr.h: Include tree-core.h instead of tree.h.
2051 * gimple.h: Likewise.
2052 * ipa-prop.h: Likewise.
2053 * ipa-utils.h: Likewise.
2054 * lto-streamer.h: Likewise.
2055 * streamer-hooks.h: Likewise.
2056 * ipa-reference.h: Include cgraph.h instead of tree.h.
2057 * cgraph.h: Include basic-block.h instead of tree.h.
2058 * tree-streamer.h: Do not include tree.h.
2059 * genattrtab.c (write_header): Generate inclusion of
2061 * genautomata.c (main): Likewise.
2062 * genemit.c: Likewise.
2063 * genopinit.c: Likewise.
2064 * genoutput.c (output_prologue): Likewise.
2065 * genpeep.c: Likewise.
2067 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2069 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2071 (vec_unpacku_hi_v8hi): Likewise.
2072 (vec_unpacku_lo_v16qi): Likewise.
2073 (vec_unpacku_lo_v8hi): Likewise.
2075 2013-10-20 Jan Hubicka <jh@suse.cz>
2077 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2078 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2079 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2081 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2083 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2084 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2086 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2088 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2091 * system.h: Undef m_slot.
2093 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2095 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2096 all elements for both endian flavors.
2098 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2101 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2102 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2103 and SI_REG for 64bit SYSV ABI targets.
2105 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2107 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2108 to multi_reg_return. Clarify that we are skipping USEs of multiple
2109 return registers. Use bool type where appropriate.
2111 2013-10-18 Jan Hubicka <jh@suse.cz>
2113 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2115 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2116 (X86_TUNE_PUSH_MEMORY): Likewise.
2117 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2118 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2119 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2121 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2122 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2124 (ix86_option_override_internal): Update to use tune features instead
2127 2013-10-18 Cong Hou <congh@google.com>
2129 PR tree-optimization/58508
2130 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2131 statement that contains data refs with zero-step.
2133 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2135 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2136 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2137 * gimple-low.c (gimple_check_call_arg,
2138 gimple_check_call_matching_types): Move to cgraph.c.
2139 * gimple-low.h: Remove prototype.
2140 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2141 Relocate from gimple-low.c.
2142 * cgraph.h: Add prototype. Don't include basic-block.h.
2143 * gimplify.c: Add gimple-low to include list.
2144 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2145 * tree-eh.c: Add gimple-low to include list.
2146 * tree-nested.c: Likewise.
2147 * cfgexpand.c: Add tree-ssa-address.h to include list.
2149 * gimple-fold.c: Likewise.
2150 * gimple-ssa-strength-reduction.c: Likewise.
2151 * trans-mem.c: Likewise.
2152 * tree-mudflap.c: Likewise.
2153 * tree-ssa-loop-ivopts.c: Likewise.
2154 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2155 (degenerate_phi_result): Move to tree-phinodes.c.
2156 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2157 * tree-ssa-threadedge.c: Likewise.
2158 * tree-vrp.c: Likewise.
2159 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2160 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2161 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2162 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2163 * tree-ssa-forwprop.c: Likewise.
2164 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2165 pass_data_cleanup_cfg_post_optimizing,
2166 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2167 * tree-optimize.c: Delete File.
2168 * graphite.c: Include tree-cfgcleanup.h.
2169 * passes.c: Likewise.
2170 * tree-cfg.c: Likewise.
2171 * tree-profile.c: Likewise.
2172 * tree-ssa-dse.c: Likewise.
2173 * tree-ssa-loop-ivcanon.c: Likewise.
2174 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2175 * tree-outof-ssa.c: Include sbitmap.h.
2176 * tree-ssa-live.c: Likewise.
2177 * tree-ssa-propagate.c: Likewise.
2178 * tree-ssa-structalias.c: Likewise.
2179 * tree-stdarg.c: Likewise.
2180 * Makefile.in (OBJS): Delete tree-optimize.o.
2181 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2182 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2183 * varasm.c: Include basic-block.h.
2184 * cfgloop.h: Include function.h instead of basic-block.h
2185 (bb_loop_depth): Move to cfgloop.c.
2186 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2188 2013-10-18 Teresa Johnson <tejohnson@google.com>
2190 * predict.c (probably_never_executed): Compare frequency-based
2191 count to number of training runs.
2192 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2194 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2196 * config/arm/arm.c (cortexa9_extra_costs): New table.
2197 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2199 2013-10-18 Jeff Law <law@redhat.com>
2201 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2203 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2204 (register_jump_thread): Add "registered_jump_thread" debug
2206 * dbgcnt.def (registered_jump_thread): New debug counter.
2208 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2210 * config/rs6000/rs6000.c: Include cgraph.h.
2212 2013-10-18 Teresa Johnson <tejohnson@google.com>
2214 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2215 weights during merging.
2217 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2219 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
2220 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
2221 * tree-ssa-operands.h: Remove prototype.
2222 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
2223 * gimple.c (virtual_operand_p): Relocate from gimple.c.
2224 * gimple.h: Add prototype.
2225 * gimple-ssa.h: Include tree-ssa-operands.h.
2226 * tree-dump.c: Add tree-cfg.h to include list.
2227 * tree-ssa-alias.c: Add ipa-reference.h to include list.
2228 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
2229 * config/i386/i386.c: Don't include tree-flow.h.
2230 * config/rs6000/rs6000.c: Likewise.
2232 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2234 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
2236 2013-10-18 Richard Biener <rguenther@suse.de>
2238 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
2239 or TYPE_UNSIGNED of integral types.
2240 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
2241 NULL_TREE for zero-precision integral types.
2243 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
2245 * config/aarch64/arm_neon.h
2246 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
2248 2013-10-17 Sriraman Tallam <tmsriram@google.com>
2251 * opth-gen.awk: Define target_flags_explicit.
2253 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2255 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
2256 fields to the reg_addr array that describes the valid addressing
2257 mode for any register, general purpose registers, floating point
2258 registers, and Altivec registers.
2259 (FIRST_RELOAD_REG_CLASS): Likewise.
2260 (LAST_RELOAD_REG_CLASS): Likewise.
2261 (struct reload_reg_map_type): Likewise.
2262 (reload_reg_map_type): Likewise.
2263 (RELOAD_REG_VALID): Likewise.
2264 (RELOAD_REG_MULTIPLE): Likewise.
2265 (RELOAD_REG_INDEXED): Likewise.
2266 (RELOAD_REG_OFFSET): Likewise.
2267 (RELOAD_REG_PRE_INCDEC): Likewise.
2268 (RELOAD_REG_PRE_MODIFY): Likewise.
2269 (reg_addr): Likewise.
2270 (mode_supports_pre_incdec_p): New helper functions to say whether
2271 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
2272 (mode_supports_pre_modify_p): Likewise.
2273 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
2274 print the valid address mode bits for each mode.
2275 (rs6000_debug_print_mode): Likewise.
2276 (rs6000_debug_reg_global): Likewise.
2277 (rs6000_setup_reg_addr_masks): New function to set up the address
2278 mask bits for each type.
2279 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
2280 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
2281 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
2282 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
2283 PRE_MODIFY are supported.
2284 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
2285 registers, instead of {src,dest}_av_p.
2286 (rs6000_print_options_internal): Tweak the debug output slightly.
2288 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2290 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
2293 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
2295 * tree-flow.h (struct omp_region): Move to omp-low.c.
2296 Remove omp_ prototypes and variables.
2297 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
2298 (copy_var_decl): Relocate prototype from tree-flow.h.
2299 * gimple.c (copy_var_decl): Relocate from omp-low.c.
2300 * tree.h: Move prototype to omp-low.h.
2301 * omp-low.h: New File. Relocate prototypes here.
2302 * omp-low.c (struct omp_region): Make local here.
2303 (root_omp_region): Make static.
2304 (copy_var_decl) Move to gimple.c.
2305 (new_omp_region): Make static.
2306 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
2307 * tree-cfg.c: Include omp-low.h.
2308 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
2309 * gimplify.c: Include omp-low.h.
2310 * tree-parloops.c: Likewise.
2312 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2314 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
2315 operands are in memory, prefer to force non-matched operand 1 to
2318 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2321 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
2322 restrict TImode addresses to single indirect registers if both
2323 -mquad-memory and -mvsx-timode are used.
2324 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
2325 we should emit load/store quad. Remove using %y for quad memory
2328 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
2329 constraints to allow load/store quad on machines where TImode is
2330 not allowed in VSX registers. Use 'n' instead of 'F' constraint
2331 for TImode to load integer constants.
2333 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2335 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
2337 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
2339 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
2340 handling of STACK_REG.
2342 2013-10-17 Richard Biener <rguenther@suse.de>
2344 PR tree-optimization/58143
2345 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
2347 (rewrite_to_defined_overflow): Likewise.
2348 (move_computations_dom_walker::before_dom): Rewrite stmts
2349 with undefined signed overflow that are not always executed
2350 into unsigned arithmetic.
2352 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2355 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
2356 explicit isa flag to be an options variable, instead of using
2357 global_options_set. Remove define from rs6000.h.
2358 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
2360 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2361 Initialize rs6000_isa_flags_explicit.
2362 (rs6000_function_specific_save): Add gcc_options* parameter, so
2363 that the powerpc builds after the 2013-10-15 changes.
2364 (rs6000_function_specific_restore): Likewise.
2366 2013-10-16 DJ Delorie <dj@redhat.com>
2368 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
2369 op is a REG before checking REGNO.
2370 (rl78_alloc_physical_registers): Verify pattern is a SET before
2373 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2375 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
2377 (vec_unpacks_lo_v4sf): Likewise.
2378 (vec_unpacks_float_hi_v4si): Likewise.
2379 (vec_unpacks_float_lo_v4si): Likewise.
2380 (vec_unpacku_float_hi_v4si): Likewise.
2381 (vec_unpacku_float_lo_v4si): Likewise.
2383 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2385 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
2386 (vsx_concat_v2sf): Likewise.
2388 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
2390 * config/aarch64/aarch64.md
2391 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
2393 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
2395 PR tree-optimization/58697
2396 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
2397 estimated_loop_iterations_int.
2398 (max_stmt_executions_int): Call get_max_loop_iterations_int.
2399 (get_max_loop_iterations_int): New. HWINT version of
2400 get_max_loop_iterations.
2401 * cfgloop.h: Add prototypes.
2402 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
2403 * loop-unroll.c (decide_peel_once_rolling): Call
2404 get_estimated_loop_iterations_int.
2405 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
2406 * tree-ssa-loop-niter.h: Tweak prototypes.
2408 2013-10-16 David Malcolm <dmalcolm@redhat.com>
2410 * gengtype-parse.c (struct_field_seq): Ignore access-control
2411 keywords ("public:" etc).
2413 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
2415 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
2416 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
2418 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
2420 * config/arm/arm.opt (mlra): New option.
2421 * config/arm/arm.c (arm_lra_p): New function.
2422 (TARGET_LRA_P): Define.
2424 2013-10-16 Paulo Matos <pmatos@broadcom.com>
2426 * tree-core.h (tree_code_name): Remove.
2427 * tree.h (get_tree_code_name): New prototype.
2428 * tree.c (tree_code_name): Make static.
2429 (get_tree_code_name): New function.
2430 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
2431 tree_class_check_failed, tree_range_check_failed,
2432 tree_not_class_check_failed, omp_clause_check_failed,
2433 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
2434 wrapper get_tree_code_name instead of calling tree_code_name directly.
2435 * tree-vrp.c (dump_asserts_for): Likewise.
2436 * tree-dump.c (dequeue_and_dump): Likewise.
2437 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
2438 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
2439 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
2440 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
2441 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
2442 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
2443 dump_gimple_omp_for): Likewise.
2444 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
2445 * tree-ssa-pre.c (print_pre_expr): Likewise.
2446 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
2447 * print-tree.c (print_node_brief, print_node): Likewise.
2448 * gimple.c (gimple_check_failed): Likewise.
2449 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
2450 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
2451 * config/mep/mep.c (mep_validate_vliw): Likewise.
2452 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
2453 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
2455 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
2457 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
2460 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
2462 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
2463 (MULTILIB_MATCHES): Add multilib for -march=v8.
2465 2013-10-15 Sriraman Tallam <tmsriram@google.com>
2468 * optc-save-gen.awk: Add extra parameter to the save and restore
2470 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
2471 * tree.c (build_optimization_node): New parameter. Add extra parameter
2472 to call to cl_optimization_save.
2473 (build_target_option_node): New parameter. Add extra parameter
2474 to call to cl_target_option_save.
2475 * tree.h (build_optimization_node): New parameter.
2476 (build_target_option_node): New parameter.
2477 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
2478 build_optimization_node and build_target_option_node.
2479 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
2480 (handle_pragma_push_options): Ditto.
2481 * toplev.c (process_options): Ditto.
2482 * opts.c (init_options_struct): Check for opts_set non-null.
2483 * target.def (target_option.save): New parameter.
2484 (target_option.restore): New parameter.
2485 * tm.texi: Generate.
2486 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
2487 (ix86_pragma_target_parse): Ditto.
2488 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
2490 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
2491 to build_optimization_node and build_target_option_node.
2492 (rs6000_valid_attribute_p): Ditto.
2493 (rs6000_pragma_target_parse): Ditto.
2494 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
2496 * config/i386/i386.h:
2497 TARGET_64BIT_P: New Macro
2498 TARGET_MMX_P: New Macro.
2499 TARGET_3DNOW_P: New Macro.
2500 TARGET_3DNOW_A_P: New Macro.
2501 TARGET_SSE_P: New Macro.
2502 TARGET_SSE2_P: New Macro.
2503 TARGET_SSE3_P: New Macro.
2504 TARGET_SSSE3_P: New Macro.
2505 TARGET_SSE4_1_P: New Macro.
2506 TARGET_SSE4_2_P: New Macro.
2507 TARGET_AVX_P: New Macro.
2508 TARGET_AVX2_P: New Macro.
2509 TARGET_AVX512F_P: New Macro.
2510 TARGET_AVX512PF_P: New Macro.
2511 TARGET_AVX512ER_P: New Macro.
2512 TARGET_AVX512CD_P: New Macro.
2513 TARGET_FMA_P: New Macro.
2514 TARGET_SSE4A_P: New Macro.
2515 TARGET_FMA4_P: New Macro.
2516 TARGET_XOP_P: New Macro.
2517 TARGET_LWP_P: New Macro.
2518 TARGET_ABM_P: New Macro.
2519 TARGET_BMI_P: New Macro.
2520 TARGET_BMI2_P: New Macro.
2521 TARGET_LZCNT_P: New Macro.
2522 TARGET_TBM_P: New Macro.
2523 TARGET_POPCNT_P: New Macro.
2524 TARGET_SAHF_P: New Macro.
2525 TARGET_MOVBE_P: New Macro.
2526 TARGET_CRC32_P: New Macro.
2527 TARGET_AES_P: New Macro.
2528 TARGET_PCLMUL_P: New Macro.
2529 TARGET_CMPXCHG16B_P: New Macro.
2530 TARGET_FSGSBASE_P: New Macro.
2531 TARGET_RDRND_P: New Macro.
2532 TARGET_F16C_P: New Macro.
2533 TARGET_RTM_P: New Macro.
2534 TARGET_HLE_P: New Macro.
2535 TARGET_RDSEED_P: New Macro.
2536 TARGET_PRFCHW_P: New Macro.
2537 TARGET_ADX_P: New Macro.
2538 TARGET_FXSR_P: New Macro.
2539 TARGET_XSAVE_P: New Macro.
2540 TARGET_XSAVEOPT_P: New Macro.
2541 TARGET_LP64_P: New Macro.
2542 TARGET_X32_P: New Macro.
2543 TARGET_FPMATH_DEFAULT_P: New Macro.
2544 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
2545 * config/i386/i386.c (ix86_option_override_internal): New parameters.
2547 Change ix86_tune_string to access opts->x_ix86_tune_string.
2548 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
2549 Change ix86_arch_string to access opts->x_ix86_arch_string.
2550 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
2551 Change ix86_pmode to access opts->x_ix86_pmode.
2552 Change ix86_abi to access opts->x_ix86_abi.
2553 Change ix86_cmodel to access opts->x_ix86_cmodel.
2554 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
2555 Change ix86_isa_flags_explicit to access
2556 opts->x_ix86_isa_flags_explicit.
2557 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
2558 Change ix86_regparm to access opts->x_ix86_regparm.
2559 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
2560 Change ix86_preferred_stack_boundary_arg to access
2561 opts->x_ix86_preferred_stack_boundary_arg.
2562 Change ix86_force_align_arg_pointer to access
2563 opts->x_ix86_force_align_arg_pointer.
2564 Change ix86_incoming_stack_boundar_arg to access
2565 opts->x_ix86_incoming_stack_boundar_arg.
2566 Change ix86_fpmath to access opts->x_ix86_fpmath.
2567 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
2568 Change ix86_recip_name to access opts->x_ix86_recip_name.
2569 Change ix86_stack_protector_guard to access
2570 opts->x_ix86_stack_protector_guard.
2571 Change ix86_tune_memcpy_strategy to access
2572 opts->x_ix86_tune_memcpy_strategy.
2573 Change ix86_tune_memset_strategy to access
2574 opts->x_ix86_tune_memset_strategy.
2575 Change global_options to access opts.
2576 Change global_options_set to access opts_set.
2577 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
2578 Change TARGET_MMX to TARGET_MMX_P (opts->...).
2579 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
2580 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
2581 Change TARGET_SSE to TARGET_SSE_P (opts->...).
2582 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
2583 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
2584 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
2585 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
2586 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
2587 Change TARGET_AVX to TARGET_AVX_P (opts->...).
2588 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
2589 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
2590 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
2591 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
2592 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
2593 Change TARGET_FMA to TARGET_FMA_P (opts->...).
2594 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
2595 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
2596 Change TARGET_XOP to TARGET_XOP_P (opts->...).
2597 Change TARGET_LWP to TARGET_LWP_P (opts->...).
2598 Change TARGET_ABM to TARGET_ABM_P (opts->...).
2599 Change TARGET_BMI to TARGET_BMI_P (opts->...).
2600 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
2601 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
2602 Change TARGET_TBM to TARGET_TBM_P (opts->...).
2603 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
2604 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
2605 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
2606 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
2607 Change TARGET_AES to TARGET_AES_P (opts->...).
2608 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
2609 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
2610 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
2611 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
2612 Change TARGET_F16C to TARGET_F16C_P (opts->...).
2613 Change TARGET_RTM to TARGET_RTM_P (opts->...).
2614 Change TARGET_HLE to TARGET_HLE_P (opts->...).
2615 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
2616 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
2617 Change TARGET_ADX to TARGET_ADX_P (opts->...).
2618 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
2619 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
2620 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
2621 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
2622 Change TARGET_X32 to TARGET_X32_P (opts->...).
2623 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
2624 Change TARGET_FLOAT_RETURNS_IN_80387 to
2625 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
2626 (ix86_function_specific_save): New parameter. Use opts-> fields
2627 to replace global fields.
2628 (ix86_function_specific_restore): Ditto.
2629 (ix86_valid_target_attribute_inner_p): New parameters.
2631 Fix call to ix86_handle_option and set_option.
2632 (ix86_valid_target_attribute_tree): New parameters.
2633 Change global_options to access opts.
2634 Change global_options_set to access opts_set.
2635 Fix call to ix86_valid_target_attribute_inner_p.
2636 Change ix86_tune_string to access opts->x_ix86_tune_string.
2637 Change ix86_arch_string to access opts->x_ix86_arch_string.
2638 Change ix86_fpmath to access opts->x_ix86_fpmath
2639 Fix call to ix86_option_override_internal.
2640 Fix call to ix86_add_new_builtins.
2641 Fix calls to build_optimization_node and build_target_option_node.
2642 (ix86_valid_target_attribute_p): Remove access to global_options.
2643 Use new gcc_options structure func_options.
2644 Fix call to ix86_valid_target_attribute_tree.
2645 Fix call to build_optimization_node.
2646 (get_builtin_code_for_version): Fix call to
2647 ix86_valid_target_attribute_tree.
2649 2013-10-15 David Malcolm <dmalcolm@redhat.com>
2651 * Makefile.in (PICFLAG): New.
2652 (enable_host_shared): New.
2653 (INTERNAL_CFLAGS): Use PICFLAG.
2654 (LIBIBERTY): Use pic build of libiberty.a if configured with
2655 --enable-host-shared.
2656 * configure.ac: Add --enable-host-shared, setting up new
2658 * configure: Regenerate.
2659 * doc/install.texi (--enable-shared): Add note contrasting it with ...
2660 (--enable-host-shared): New option.
2662 2013-10-15 Richard Biener <rguenther@suse.de>
2664 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
2665 for built-in functions.
2667 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
2669 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
2670 (optimize_range_tests_1): New function,
2671 extracted from optimize_range_tests.
2672 (optimize_range_tests_xor): Similarly.
2673 (optimize_range_tests_diff): New function.
2674 (optimize_range_tests): Use optimize_range_tests_1.
2676 2013-10-15 Cong Hou <congh@google.com>
2678 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
2679 requirement of the reduction pattern so that one operand of the
2680 reduction operation can come from outside of the loop.
2682 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2684 * config/arm/neon-schedgen.ml: Remove.
2685 * config/arm/cortex-a9-neon.md: Remove comment regarding
2688 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2690 * config/arm/types: Remove old neon types.
2692 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2694 * config/arm/cortex-a7.md
2695 (cortex_a7_neon_type): New.
2696 (cortex_a7_neon_mul): Update for new types.
2697 (cortex_a7_neon_mla): Likewise.
2698 (cortex_a7_neon): Likewise.
2700 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2702 * config/arm/cortex-a15-neon.md
2703 (cortex_a15_neon_type): New,
2705 (cortex_a15_neon_int_1): Remove.
2706 (cortex_a15_neon_int_2): Likewise.
2707 (cortex_a15_neon_int_3): Likewise.
2708 (cortex_a15_neon_int_4): Likewise.
2709 (cortex_a15_neon_int_5): Likewise.
2710 (cortex_a15_neon_vqneg_vqabs): Likewise.
2711 (cortex_a15_neon_vmov): Likewise.
2712 (cortex_a15_neon_vaba): Likewise.
2713 (cortex_a15_neon_vaba_qqq): Likewise.
2714 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2715 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2716 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
2718 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2719 (cortex_a15_neon_mla_qqq_8_16): Likewise.
2720 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
2721 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2722 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2723 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
2724 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2725 (cortex_a15_neon_shift_1): Likewise.
2726 (cortex_a15_neon_shift_2): Likewise.
2727 (cortex_a15_neon_shift_3): Likewise.
2728 (cortex_a15_neon_vshl_ddd): Likewise.
2729 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2730 (cortex_a15_neon_vsra_vrsra): Likewise.
2731 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
2732 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
2733 (cortex_a15_neon_bp_3cycle): Likewise.
2734 (cortex_a15_neon_ldm_2): Likewise.
2735 (cortex_a15_neon_stm_2): Likewise.
2736 (cortex_a15_neon_mcr): Likewise.
2737 (cortex_a15_neon_mrc): Likewise.
2738 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
2739 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
2740 (cortex_a15_neon_fp_vmul_ddd): Likewise.
2741 (cortex_a15_neon_fp_vmul_qqd): Likewise.
2742 (cortex_a15_neon_fp_vmla_ddd): Likewise.
2743 (cortex_a15_neon_fp_vmla_qqq): Likewise.
2744 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
2745 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
2746 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2747 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2748 (cortex_a15_neon_bp_simple): Likewise.
2749 (cortex_a15_neon_bp_2cycle): Likewise.
2750 (cortex_a15_neon_bp_3cycle): Likewise.
2751 (cortex_a15_neon_vld1_1_2_regs): Likewise.
2752 (cortex_a15_neon_vld1_3_4_regs): Likewise.
2753 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2754 (cortex_a15_neon_vld2_4_regs): Likewise.
2755 (cortex_a15_neon_vld3_vld4): Likewise.
2756 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2757 (cortex_a15_neon_vst1_3_4_regs): Likewise.
2758 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
2759 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
2760 (cortex_a15_neon_vst3_vst4): Rename to...
2761 (cortex_a15_neon_vst4): This, update for new attributes.
2762 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
2763 (cortex_a15_neon_vld3_vld4_lane): Likewise.
2764 (cortex_a15_neon_vst1_vst2_lane): Likewise.
2765 (cortex_a15_neon_vst3_vst4_lane): Likewise.
2766 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
2767 (cortex_a15_neon_ldm_2): Likewise.
2768 (cortex_a15_neon_stm_2): Likewise.
2769 (cortex_a15_neon_mcr): Likewise.
2770 (cortex_a15_neon_mcr_2_mcrr): Likewise.
2771 (cortex_a15_neon_mrc): Likewise.
2772 (cortex_a15_neon_mrrc): Likewise.
2774 (cortex_a15_neon_abd): New.
2775 (cortex_a15_neon_abd_q): Likewise.
2776 (cortex_a15_neon_aba): Likewise.
2777 (cortex_a15_neon_aba_q): Likewise.
2778 (cortex_a15_neon_acc): Likewise.
2779 (cortex_a15_neon_acc_q): Likewise.
2780 (cortex_a15_neon_arith_basic): Likewise.
2781 (cortex_a15_neon_arith_complex): Likewise.
2782 (cortex_a15_neon_multiply): Likewise.
2783 (cortex_a15_neon_multiply_q): Likewise.
2784 (cortex_a15_neon_mla): Likewise.
2785 (cortex_a15_neon_mla_q): Likewise.
2786 (cortex_a15_neon_sat_mla_long): Likewise.
2787 (cortex_a15_neon_shift_acc): Likewise.
2788 (cortex_a15_neon_shift_imm_basic): Likewise.
2789 (cortex_a15_neon_shift_imm_complex): Likewise.
2790 (cortex_a15_neon_shift_reg_basic): Likewise.
2791 (cortex_a15_neon_shift_reg_basic_q): Likewise.
2792 (cortex_a15_neon_shift_reg_complex): Likewise.
2793 (cortex_a15_neon_shift_reg_complex_q): Likewise.
2794 (cortex_a15_neon_fp_negabs): Likewise
2795 (cortex_a15_neon_fp_arith): Likewise
2796 (cortex_a15_neon_fp_arith_q): Likewise
2797 (cortex_a15_neon_fp_cvt_int): Likewise
2798 (cortex_a15_neon_fp_cvt_int_q): Likewise
2799 (cortex_a15_neon_fp_cvt_16): Likewise
2800 (cortex_a15_neon_fp_mul): Likewise
2801 (cortex_a15_neon_fp_mul_q): Likewise
2802 (cortex_a15_neon_fp_mla): Likewise
2803 (cortex_a15_neon_fp_mla_q): Likewise
2804 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
2805 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
2806 (cortex_a15_neon_bitops): Likewise.
2807 (cortex_a15_neon_bitops_q): Likewise.
2808 (cortex_a15_neon_from_gp): Likewise.
2809 (cortex_a15_neon_from_gp_q): Likewise.
2810 (cortex_a15_neon_tbl3_tbl4): Likewise.
2811 (cortex_a15_neon_zip_q): Likewise.
2812 (cortex_a15_neon_to_gp): Likewise.
2813 (cortex_a15_neon_load_a): Likewise.
2814 (cortex_a15_neon_load_b): Likewise.
2815 (cortex_a15_neon_load_c): Likewise.
2816 (cortex_a15_neon_load_d): Likewise.
2817 (cortex_a15_neon_load_e): Likewise.
2818 (cortex_a15_neon_load_f): Likewise.
2819 (cortex_a15_neon_store_a): Likewise.
2820 (cortex_a15_neon_store_b): Likewise.
2821 (cortex_a15_neon_store_c): Likewise.
2822 (cortex_a15_neon_store_d): Likewise.
2823 (cortex_a15_neon_store_e): Likewise.
2824 (cortex_a15_neon_store_f): Likewise.
2825 (cortex_a15_neon_store_g): Likewise.
2826 (cortex_a15_neon_store_h): Likewise.
2827 (cortex_a15_vfp_to_from_gp): Likewise.
2829 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2831 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
2833 (cortex_a9_neon_vshl_ddd): Remove.
2834 (cortex_a9_neon_vst3_vst4): Likewise.
2835 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
2837 (cortex_a9_neon_bit_ops_q): New.
2839 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
2840 (cortex_a9_neon_int_2): Likewise.
2841 (cortex_a9_neon_int_3): Likewise.
2842 (cortex_a9_neon_int_4): Likewise.
2843 (cortex_a9_neon_int_5): Likewise.
2844 (cortex_a9_neon_vqneg_vqabs): Likewise.
2845 (cortex_a9_neon_vmov): Likewise.
2846 (cortex_a9_neon_vaba): Likewise.
2847 (cortex_a9_neon_vaba_qqq): Likewise.
2848 (cortex_a9_neon_shift_1): Likewise.
2849 (cortex_a9_neon_shift_2): Likewise.
2850 (cortex_a9_neon_shift_3): Likewise.
2851 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2852 (cortex_a9_neon_vsra_vrsra): Likewise.
2853 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2854 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2855 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
2857 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2858 (cortex_a9_neon_mla_qqq_8_16): Likewise.
2859 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
2861 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2862 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2863 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
2864 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2865 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
2866 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
2867 (cortex_a9_neon_fp_vsum): Likewise.
2868 (cortex_a9_neon_fp_vmul_ddd): Likewise.
2869 (cortex_a9_neon_fp_vmul_qqd): Likewise.
2870 (cortex_a9_neon_fp_vmla_ddd): Likewise.
2871 (cortex_a9_neon_fp_vmla_qqq): Likewise.
2872 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
2873 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
2874 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2875 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2876 (cortex_a9_neon_bp_simple): Likewise.
2877 (cortex_a9_neon_bp_2cycle): Likewise.
2878 (cortex_a9_neon_bp_3cycle): Likewise.
2879 (cortex_a9_neon_ldr): Likewise.
2880 (cortex_a9_neon_str): Likewise.
2881 (cortex_a9_neon_vld1_1_2_regs): Likewise.
2882 (cortex_a9_neon_vld1_3_4_regs): Likewise.
2883 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2884 (cortex_a9_neon_vld2_4_regs): Likewise.
2885 (cortex_a9_neon_vld3_vld4): Likewise.
2886 (cortex_a9_neon_vld1_vld2_lane): Likewise.
2887 (cortex_a9_neon_vld3_vld4_lane): Likewise.
2888 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
2889 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2890 (cortex_a9_neon_vst1_3_4_regs): Likewise.
2891 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
2892 (cortex_a9_neon_vst1_vst2_lane): Likewise.
2893 (cortex_a9_neon_vst3_vst4_lane): Likewise.
2894 (cortex_a9_neon_mcr): Likewise.
2895 (cortex_a9_neon_mcr_2_mcrr): Likewise.
2896 (cortex_a9_neon_mrc): Likewise.
2897 (cortex_a9_neon_mrrc): Likewise.
2899 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2901 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
2903 (cortex_a8_neon_vshl_ddd): Remove.
2904 (cortex_a8_neon_vst3_vst4): Likewise.
2905 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
2907 (cortex_a8_neon_bit_ops_q): New.
2909 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
2910 (cortex_a8_neon_int_2): Likewise..
2911 (cortex_a8_neon_int_3): Likewise.
2912 (cortex_a8_neon_int_5): Likewise.
2913 (cortex_a8_neon_vqneg_vqabs): Likewise.
2914 (cortex_a8_neon_int_4): Likewise.
2915 (cortex_a8_neon_vaba): Likewise.
2916 (cortex_a8_neon_vaba_qqq): Likewise.
2917 (cortex_a8_neon_shift_1): Likewise.
2918 (cortex_a8_neon_shift_2): Likewise.
2919 (cortex_a8_neon_shift_3): Likewise.
2920 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2921 (cortex_a8_neon_vsra_vrsra): Likewise.
2922 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2923 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2924 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
2926 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2927 (cortex_a8_neon_mla_qqq_8_16): Likewise.
2928 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
2930 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2931 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2932 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
2933 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2934 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
2935 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
2936 (cortex_a8_neon_fp_vsum): Likewise.
2937 (cortex_a8_neon_fp_vmul_ddd): Likewise.
2938 (cortex_a8_neon_fp_vmul_qqd): Likewise.
2939 (cortex_a8_neon_fp_vmla_ddd): Likewise.
2940 (cortex_a8_neon_fp_vmla_qqq): Likewise.
2941 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
2942 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
2943 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2944 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2945 (cortex_a8_neon_bp_simple): Likewise.
2946 (cortex_a8_neon_bp_2cycle): Likewise.
2947 (cortex_a8_neon_bp_3cycle): Likewise.
2948 (cortex_a8_neon_ldr): Likewise.
2949 (cortex_a8_neon_str): Likewise.
2950 (cortex_a8_neon_vld1_1_2_regs): Likewise.
2951 (cortex_a8_neon_vld1_3_4_regs): Likewise.
2952 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2953 (cortex_a8_neon_vld2_4_regs): Likewise.
2954 (cortex_a8_neon_vld3_vld4): Likewise.
2955 (cortex_a8_neon_vld1_vld2_lane): Likewise.
2956 (cortex_a8_neon_vld3_vld4_lane): Likewise.
2957 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2958 (cortex_a8_neon_vst1_3_4_regs): Likewise.
2959 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
2960 (cortex_a8_neon_vst1_vst2_lane): Likewise.
2961 (cortex_a8_neon_vst3_vst4_lane): Likewise.
2962 (cortex_a8_neon_mcr): Likewise.
2963 (cortex_a8_neon_mcr_2_mcrr): Likewise.
2964 (cortex_a8_neon_mrc): Likewise.
2965 (cortex_a8_neon_mrrc): Likewise.
2967 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2969 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
2971 * config/aarch64/aarch64-simd.md (neon_type): Remove.
2972 (aarch64_simd_dup<mode>): Add "type" attribute.
2973 (aarch64_dup_lane<mode>): Likewise.
2974 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
2975 (*aarch64_simd_mov<mode>): Likewise.
2976 (aarch64_simd_mov_from_<mode>low): Likewise.
2977 (aarch64_simd_mov_from_<mode>high): Likewise.
2978 (orn<mode>3): Likewise.
2979 (bic<mode>3): Likewise.
2980 (add<mode>3): Likewise.
2981 (sub<mode>3): Likewise.
2982 (mul<mode>3): Likewise.
2983 (*aarch64_mul3_elt<mode>): Likewise.
2984 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2985 (*aarch64_mul3_elt_to_128df): Likewise.
2986 (*aarch64_mul3_elt_to_64v2df): Likewise.
2987 (neg<mode>2): Likewise.
2988 (abs<mode>2): Likewise.
2989 (abd<mode>_3): Likewise.
2990 (aba<mode>_3): Likewise.
2991 (fabd<mode>_3): Likewise.
2992 (*fabd_scalar<mode>3): Likewise.
2993 (and<mode>3): Likewise.
2994 (ior<mode>3): Likewise.
2995 (xor<mode>3): Likewise.
2996 (one_cmpl<mode>2): Likewise.
2997 (aarch64_simd_vec_set<mode>): Likewise.
2998 (aarch64_simd_lshr<mode>): Likewise.
2999 (aarch64_simd_ashr<mode>): Likewise.
3000 (aarch64_simd_imm_shl<mode>): Likewise.
3001 (aarch64_simd_reg_sshl<mode): Likewise.
3002 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3003 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3004 (aarch64_simd_vec_setv2di): Likewise.
3005 (aarch64_simd_vec_set<mode>): Likewise.
3006 (aarch64_mla<mode>): Likewise.
3007 (*aarch64_mla_elt<mode>): Likewise.
3008 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3009 (aarch64_mls<mode>): Likewise.
3010 (*aarch64_mls_elt<mode>): Likewise.
3011 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3012 (<su><maxmin><mode>3): Likewise.
3013 (move_lo_quad_<mode>): Likewise.
3014 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3015 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3016 (vec_pack_trunc_<mode>): Likewise.
3017 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3018 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3019 (*aarch64_<su>mlal_lo<mode>): Likewise.
3020 (*aarch64_<su>mlal_hi<mode>): Likewise.
3021 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3022 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3023 (*aarch64_<su>mlal<mode>): Likewise.
3024 (*aarch64_<su>mlsl<mode>): Likewise.
3025 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3026 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3027 (add<mode>3): Likewise.
3028 (sub<mode>3): Likewise.
3029 (mul<mode>3): Likewise.
3030 (div<mode>3): Likewise.
3031 (neg<mode>2): Likewise.
3032 (abs<mode>2): Likewise.
3033 (fma<mode>4): Likewise.
3034 (*aarch64_fma4_elt<mode>): Likewise.
3035 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3036 (*aarch64_fma4_elt_to_128df): Likewise.
3037 (*aarch64_fma4_elt_to_64v2df): Likewise.
3038 (fnma<mode>4): Likewise.
3039 (*aarch64_fnma4_elt<mode>): Likewise.
3040 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3041 (*aarch64_fnma4_elt_to_128df): Likewise.
3042 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3043 (<frint_pattern><mode>2): Likewise.
3044 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3045 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3046 (vec_unpacks_lo_v4sf): Likewise.
3047 (aarch64_float_extend_lo_v2df): Likewise.
3048 (vec_unpacks_hi_v4sf): Likewise.
3049 (aarch64_float_truncate_lo_v2sf): Likewise.
3050 (aarch64_float_truncate_hi_v4sf): Likewise.
3051 (aarch64_vmls<mode>): Likewise.
3052 (<su><maxmin><mode>3): Likewise.
3053 (<maxmin_uns><mode>3): Likewise.
3054 (reduc_<sur>plus_<mode>): Likewise.
3055 (reduc_<sur>plus_v2di): Likewise.
3056 (reduc_<sur>plus_v2si): Likewise.
3057 (reduc_<sur>plus_<mode>): Likewise.
3058 (aarch64_addpv4sf): Likewise.
3059 (clz<mode>2): Likewise.
3060 (reduc_<maxmin_uns>_<mode>): Likewise.
3061 (reduc_<maxmin_uns>_v2di): Likewise.
3062 (reduc_<maxmin_uns>_v2si): Likewise.
3063 (reduc_<maxmin_uns>_<mode>): Likewise.
3064 (reduc_<maxmin_uns>_v4sf): Likewise.
3065 (aarch64_simd_bsl<mode>_internal): Likewise.
3066 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3067 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3068 (aarch64_get_lane<mode>): Likewise.
3069 (*aarch64_combinez<mode>): Likewise.
3070 (aarch64_combine<mode>): Likewise.
3071 (aarch64_simd_combine<mode>): Likewise.
3072 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3073 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3074 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3075 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3076 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3077 (aarch64_<sur>h<addsub><mode>): Likewise.
3078 (aarch64_<sur><addsub>hn<mode>): Likewise.
3079 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3080 (aarch64_pmul<mode>): Likewise.
3081 (aarch64_<su_optab><optab><mode>): Likewise.
3082 (aarch64_<sur>qadd<mode>): Likewise.
3083 (aarch64_sqmovun<mode>): Likewise.
3084 (aarch64_<sur>qmovn<mode>): Likewise.
3085 (aarch64_s<optab><mode>): Likewise.
3086 (aarch64_sq<r>dmulh<mode>): Likewise.
3087 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3088 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3089 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3090 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3091 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3092 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3093 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3094 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3095 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3096 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3097 (aarch64_sqdmull<mode>): Likewise.
3098 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3099 (aarch64_sqdmull_n<mode>): Likewise.
3100 (aarch64_sqdmull2<mode>_internal): Likewise.
3101 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3102 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3103 (aarch64_<sur>shl<mode>): Likewise.
3104 (aarch64_<sur>q<r>shl<mode>
3105 (aarch64_<sur>shll_n<mode>): Likewise.
3106 (aarch64_<sur>shll2_n<mode>): Likewise.
3107 (aarch64_<sur>shr_n<mode>): Likewise.
3108 (aarch64_<sur>sra_n<mode>): Likewise.
3109 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3110 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3111 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3112 (aarch64_cm<optab><mode>): Likewise.
3113 (aarch64_cm<optab>di): Likewise.
3114 (aarch64_cm<optab><mode>): Likewise.
3115 (aarch64_cm<optab>di): Likewise.
3116 (aarch64_cmtst<mode>): Likewise.
3117 (aarch64_cmtstdi): Likewise.
3118 (aarch64_cm<optab><mode>): Likewise.
3119 (*aarch64_fac<optab><mode>): Likewise.
3120 (aarch64_addp<mode>): Likewise.
3121 (aarch64_addpdi): Likewise.
3122 (sqrt<mode>2): Likewise.
3123 (vec_load_lanesoi<mode>): Likewise.
3124 (vec_store_lanesoi<mode>): Likewise.
3125 (vec_load_lanesci<mode>): Likewise.
3126 (vec_store_lanesci<mode>): Likewise.
3127 (vec_load_lanesxi<mode>): Likewise.
3128 (vec_store_lanesxi<mode>): Likewise.
3129 (*aarch64_mov<mode>): Likewise.
3130 (aarch64_ld2<mode>_dreg): Likewise.
3131 (aarch64_ld2<mode>_dreg): Likewise.
3132 (aarch64_ld3<mode>_dreg): Likewise.
3133 (aarch64_ld3<mode>_dreg): Likewise.
3134 (aarch64_ld4<mode>_dreg): Likewise.
3135 (aarch64_ld4<mode>_dreg): Likewise.
3136 (aarch64_tbl1<mode>): Likewise.
3137 (aarch64_tbl2v16qi): Likewise.
3138 (aarch64_combinev16qi): Likewise.
3139 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3140 (aarch64_st2<mode>_dreg): Likewise.
3141 (aarch64_st2<mode>_dreg): Likewise.
3142 (aarch64_st3<mode>_dreg): Likewise.
3143 (aarch64_st3<mode>_dreg): Likewise.
3144 (aarch64_st4<mode>_dreg): Likewise.
3145 (aarch64_st4<mode>_dreg): Likewise.
3146 (*aarch64_simd_ld1r<mode>): Likewise.
3147 (aarch64_frecpe<mode>): Likewise.
3148 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3149 (aarch64_frecps<mode>): Likewise.
3151 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3153 * config/arm/iterators.md (V_elem_ch): New.
3155 (VQH_type): Likewise.
3156 * config/arm/arm.md (is_neon_type): New.
3157 (conds): Use is_neon_type.
3158 (anddi3_insn): Update type attribute.
3159 (xordi3_insn): Likewise.
3160 (one_cmpldi2): Likewise.
3161 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3162 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3163 (*movmisalign<mode>_neon_store): Likewise.
3164 (*movmisalign<mode>_neon_load): Likewise.
3165 (vec_set<mode>_internal): Likewise.
3166 (vec_set<mode>_internal): Likewise.
3167 (vec_setv2di_internal): Likewise.
3168 (vec_extract<mode>): Likewise.
3169 (vec_extract<mode>): Likewise.
3170 (vec_extractv2di): Likewise.
3171 (*add<mode>3_neon): Likewise.
3172 (adddi3_neon): Likewise.
3173 (*sub<mode>3_neon): Likewise.
3174 (subdi3_neon): Likewise.
3175 (fma<VCVTF:mode>4): Likewise.
3176 (fma<VCVTF:mode>4_intrinsic): Likewise.
3177 (*fmsub<VCVTF:mode>4): Likewise.
3178 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3179 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3180 (ior<mode>3): Likewise.
3181 (and<mode>3): Likewise.
3182 (orn<mode>3_neon): Likewise.
3183 (orndi3_neon): Likewise.
3184 (bic<mode>3_neon): Likewise.
3185 (bicdi3_neon): Likewise.
3186 (xor<mode>3): Likewise.
3187 (one_cmpl<mode>2): Likewise.
3188 (abs<mode>2): Likewise.
3189 (neg<mode>2): Likewise.
3190 (negdi2_neon): Likewise.
3191 (*umin<mode>3_neon): Likewise.
3192 (*umax<mode>3_neon): Likewise.
3193 (*smin<mode>3_neon): Likewise.
3194 (*smax<mode>3_neon): Likewise.
3195 (vashl<mode>3): Likewise.
3196 (vashr<mode>3_imm): Likewise.
3197 (vlshr<mode>3_imm): Likewise.
3198 (ashl<mode>3_signed): Likewise.
3199 (ashl<mode>3_unsigned): Likewise.
3200 (neon_load_count): Likewise.
3201 (ashldi3_neon_noclobber): Likewise.
3202 (ashldi3_neon): Likewise.
3203 (signed_shift_di3_neon): Likewise.
3204 (unsigned_shift_di3_neon): Likewise.
3205 (ashrdi3_neon_imm_noclobber): Likewise.
3206 (lshrdi3_neon_imm_noclobber): Likewise.
3207 (<shift>di3_neon): Likewise.
3208 (widen_ssum<mode>3): Likewise.
3209 (widen_usum<mode>3): Likewise.
3210 (quad_halves_<code>v4si): Likewise.
3211 (quad_halves_<code>v4sf): Likewise.
3212 (quad_halves_<code>v8hi): Likewise.
3213 (quad_halves_<code>v16qi): Likewise.
3214 (reduc_splus_v2di): Likewise.
3215 (neon_vpadd_internal<mode>): Likewise.
3216 (neon_vpsmin<mode>): Likewise.
3217 (neon_vpsmax<mode>): Likewise.
3218 (neon_vpumin<mode>): Likewise.
3219 (neon_vpumax<mode>): Likewise.
3220 (*ss_add<mode>_neon): Likewise.
3221 (*us_add<mode>_neon): Likewise.
3222 (*ss_sub<mode>_neon): Likewise.
3223 (*us_sub<mode>_neon): Likewise.
3224 (neon_vadd<mode>_unspec): Likewise.
3225 (neon_vaddl<mode>): Likewise.
3226 (neon_vaddw<mode>): Likewise.
3227 (neon_vhadd<mode>): Likewise.
3228 (neon_vqadd<mode>): Likewise.
3229 (neon_vaddhn<mode>): Likewise.
3230 (neon_vmul<mode>): Likewise.
3231 (neon_vfms<VCVTF:mode>): Likewise.
3232 (neon_vmlal<mode>): Likewise.
3233 (neon_vmls<mode>): Likewise.
3234 (neon_vmlsl<mode>): Likewise.
3235 (neon_vqdmulh<mode>): Likewise.
3236 (neon_vqdmlal<mode>): Likewise.
3237 (neon_vqdmlsl<mode>): Likewise.
3238 (neon_vmull<mode>): Likewise.
3239 (neon_vqdmull<mode>): Likewise.
3240 (neon_vsub<mode>_unspec): Likewise.
3241 (neon_vsubl<mode>): Likewise.
3242 (neon_vsubw<mode>): Likewise.
3243 (neon_vqsub<mode>): Likewise.
3244 (neon_vhsub<mode>): Likewise.
3245 (neon_vsubhn<mode>): Likewise.
3246 (neon_vceq<mode>): Likewise.
3247 (neon_vcge<mode>): Likewise.
3248 (neon_vcgeu<mode>): Likewise.
3249 (neon_vcgt<mode>): Likewise.
3250 (neon_vcgtu<mode>): Likewise.
3251 (neon_vcle<mode>): Likewise.
3252 (neon_vclt<mode>): Likewise.
3253 (neon_vcage<mode>): Likewise.
3254 (neon_vcagt<mode>): Likewise.
3255 (neon_vtst<mode>): Likewise.
3256 (neon_vabd<mode>): Likewise.
3257 (neon_vabdl<mode>): Likewise.
3258 (neon_vaba<mode>): Likewise.
3259 (neon_vabal<mode>): Likewise.
3260 (neon_vmax<mode>): Likewise.
3261 (neon_vmin<mode>): Likewise.
3262 (neon_vpaddl<mode>): Likewise.
3263 (neon_vpadal<mode>): Likewise.
3264 (neon_vpmax<mode>): Likewise.
3265 (neon_vpmin<mode>): Likewise.
3266 (neon_vrecps<mode>): Likewise.
3267 (neon_vrsqrts<mode>): Likewise.
3268 (neon_vqabs<mode>): Likewise.
3269 (neon_vqneg<mode>): Likewise.
3270 (neon_vcls<mode>): Likewise.
3271 (clz<mode>2): Likewise.
3272 (popcount<mode>2): Likewise.
3273 (neon_vrecpe<mode>): Likewise.
3274 (neon_vrsqrte<mode>): Likewise.
3275 (neon_vget_lane<mode>_sext_internal): Likewise.
3276 (neon_vget_lane<mode>_zext_internal): Likewise.
3277 (neon_vdup_n<mode>): Likewise.
3278 (neon_vdup_n<mode>): Likewise.
3279 (neon_vdup_nv2di): Likewise.
3280 (neon_vdup_lane<mode>_interal): Likewise.
3281 (*neon_vswp<mode>): Likewise.
3282 (neon_vcombine<mode>): Likewise.
3283 (float<mode><V_cvtto>2): Likewise.
3284 (floatuns<mode><V_cvtto>2): Likewise.
3285 (fix_trunc<mode><V_cvtto>2): Likewise.
3286 (fixuns_trunc<mode><V_cvtto>2
3287 (neon_vcvt<mode>): Likewise.
3288 (neon_vcvt<mode>): Likewise.
3289 (neon_vcvtv4sfv4hf): Likewise.
3290 (neon_vcvtv4hfv4sf): Likewise.
3291 (neon_vcvt_n<mode>): Likewise.
3292 (neon_vcvt_n<mode>): Likewise.
3293 (neon_vmovn<mode>): Likewise.
3294 (neon_vqmovn<mode>): Likewise.
3295 (neon_vqmovun<mode>): Likewise.
3296 (neon_vmovl<mode>): Likewise.
3297 (neon_vmul_lane<mode>): Likewise.
3298 (neon_vmul_lane<mode>): Likewise.
3299 (neon_vmull_lane<mode>): Likewise.
3300 (neon_vqdmull_lane<mode>): Likewise.
3301 (neon_vqdmulh_lane<mode>): Likewise.
3302 (neon_vqdmulh_lane<mode>): Likewise.
3303 (neon_vmla_lane<mode>): Likewise.
3304 (neon_vmla_lane<mode>): Likewise.
3305 (neon_vmlal_lane<mode>): Likewise.
3306 (neon_vqdmlal_lane<mode>): Likewise.
3307 (neon_vmls_lane<mode>): Likewise.
3308 (neon_vmls_lane<mode>): Likewise.
3309 (neon_vmlsl_lane<mode>): Likewise.
3310 (neon_vqdmlsl_lane<mode>): Likewise.
3311 (neon_vext<mode>): Likewise.
3312 (neon_vrev64<mode>): Likewise.
3313 (neon_vrev32<mode>): Likewise.
3314 (neon_vrev16<mode>): Likewise.
3315 (neon_vbsl<mode>_internal): Likewise.
3316 (neon_vshl<mode>): Likewise.
3317 (neon_vqshl<mode>): Likewise.
3318 (neon_vshr_n<mode>): Likewise.
3319 (neon_vshrn_n<mode>): Likewise.
3320 (neon_vqshrn_n<mode>): Likewise.
3321 (neon_vqshrun_n<mode>): Likewise.
3322 (neon_vshl_n<mode>): Likewise.
3323 (neon_vqshl_n<mode>): Likewise.
3324 (neon_vqshlu_n<mode>): Likewise.
3325 (neon_vshll_n<mode>): Likewise.
3326 (neon_vsra_n<mode>): Likewise.
3327 (neon_vsri_n<mode>): Likewise.
3328 (neon_vsli_n<mode>): Likewise.
3329 (neon_vtbl1v8qi): Likewise.
3330 (neon_vtbl2v8qi): Likewise.
3331 (neon_vtbl3v8qi): Likewise.
3332 (neon_vtbl4v8qi): Likewise.
3333 (neon_vtbl1v16qi): Likewise.
3334 (neon_vtbl2v16qi): Likewise.
3335 (neon_vcombinev16qi): Likewise.
3336 (neon_vtbx1v8qi): Likewise.
3337 (neon_vtbx2v8qi): Likewise.
3338 (neon_vtbx3v8qi): Likewise.
3339 (neon_vtbx4v8qi): Likewise.
3340 (*neon_vtrn<mode>_insn): Likewise.
3341 (*neon_vzip<mode>_insn): Likewise.
3342 (*neon_vuzp<mode>_insn): Likewise.
3343 (neon_vld1<mode>): Likewise.
3344 (neon_vld1_lane<mode>): Likewise.
3345 (neon_vld1_lane<mode>): Likewise.
3346 (neon_vld1_dup<mode>): Likewise.
3347 (neon_vld1_dup<mode>): Likewise.
3348 (neon_vld1_dupv2di): Likewise.
3349 (neon_vst1<mode>): Likewise.
3350 (neon_vst1_lane<mode>): Likewise.
3351 (neon_vst1_lane<mode>): Likewise.
3352 (neon_vld2<mode>): Likewise.
3353 (neon_vld2<mode>): Likewise.
3354 (neon_vld2_lane<mode>): Likewise.
3355 (neon_vld2_lane<mode>): Likewise.
3356 (neon_vld2_dup<mode>): Likewise.
3357 (neon_vst2<mode>): Likewise.
3358 (neon_vst2<mode>): Likewise.
3359 (neon_vst2_lane<mode>): Likewise.
3360 (neon_vst2_lane<mode>): Likewise.
3361 (neon_vld3<mode>): Likewise.
3362 (neon_vld3qa<mode>): Likewise.
3363 (neon_vld3qb<mode>): Likewise.
3364 (neon_vld3_lane<mode>): Likewise.
3365 (neon_vld3_lane<mode>): Likewise.
3366 (neon_vld3_dup<mode>): Likewise.
3367 (neon_vst3<mode>): Likewise.
3368 (neon_vst3qa<mode>): Likewise.
3369 (neon_vst3qb<mode>): Likewise.
3370 (neon_vst3_lane<mode>): Likewise.
3371 (neon_vst3_lane<mode>): Likewise.
3372 (neon_vld4<mode>): Likewise.
3373 (neon_vld4qa<mode>): Likewise.
3374 (neon_vld4qb<mode>): Likewise.
3375 (neon_vld4_lane<mode>): Likewise.
3376 (neon_vld4_lane<mode>): Likewise.
3377 (neon_vld4_dup<mode>): Likewise.
3378 (neon_vst4<mode>): Likewise.
3379 (neon_vst4qa<mode>): Likewise.
3380 (neon_vst4qb<mode>): Likewise.
3381 (neon_vst4_lane<mode>): Likewise.
3382 (neon_vst4_lane<mode>): Likewise.
3383 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3384 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3385 (neon_vec_<US>mult_lo_<mode>): Likewise.
3386 (neon_vec_<US>mult_hi_<mode>): Likewise.
3387 (neon_vec_<US>shiftl_<mode>): Likewise.
3388 (neon_unpack<US>_<mode>): Likewise.
3389 (neon_vec_<US>mult_<mode>): Likewise.
3390 (vec_pack_trunc_<mode>): Likewise.
3391 (neon_vec_pack_trunc_<mode>): Likewise.
3392 (neon_vabd<mode>_2): Likewise.
3393 (neon_vabd<mode>_3): Likewise.
3395 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3397 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
3398 (load_pair): Update type attribute.
3399 (store_pair): Update type attribute.
3400 * config/aarch64/iterators.md (q): New.
3402 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3404 * config/arm/types.md: Add new types for Neon insns.
3406 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3407 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3408 Sergey Lega <sergey.s.lega@intel.com>
3409 Anna Tikhonova <anna.tikhonova@intel.com>
3410 Ilya Tocar <ilya.tocar@intel.com>
3411 Andrey Turetskiy <andrey.turetskiy@intel.com>
3412 Ilya Verbin <ilya.verbin@intel.com>
3413 Kirill Yukhin <kirill.yukhin@intel.com>
3414 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3416 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
3417 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
3418 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
3420 (srcp14<mode>): Ditto.
3421 (rsqrt14<mode>): Ditto.
3422 (rsqrt14<mode>): Ditto.
3423 (avx512f_vmscalef<mode>): Ditto.
3424 (avx512f_scalef<mode>): Ditto.
3425 (avx512f_getexp<mode>): Ditto.
3426 (avx512f_sgetexp<mode>): Ditto.
3427 (avx512f_fixupimm<mode>): Ditto.
3428 (avx512f_sfixupimm<mode>): Ditto.
3429 (avx512f_rndscale<mode>): Ditto.
3430 (*avx512er_exp2<mode>): Ditto.
3431 (*avx512er_rcp28<mode>): Ditto.
3432 (avx512er_rsqrt28<mode>): Ditto.
3433 (avx512f_getmant<mode>): Ditto.
3434 (avx512f_getmant<mode>): Ditto.
3435 (avx512f_rndscale<mode>): Fix formatting.
3437 2013-10-15 Martin Jambor <mjambor@suse.cz>
3439 * ipa-utils.h (ipa_edge_within_scc): Declare.
3440 * ipa-cp.c (edge_within_scc): Moved...
3441 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
3443 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3444 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3445 Sergey Lega <sergey.s.lega@intel.com>
3446 Anna Tikhonova <anna.tikhonova@intel.com>
3447 Ilya Tocar <ilya.tocar@intel.com>
3448 Andrey Turetskiy <andrey.turetskiy@intel.com>
3449 Ilya Verbin <ilya.verbin@intel.com>
3450 Kirill Yukhin <kirill.yukhin@intel.com>
3451 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3453 * config/i386/predicates.md (const_8_to_15_operand): New.
3454 (const_16_to_31_operand): Ditto.
3455 * config/i386/sse.md (V8FI): New.
3457 (reduc_splus_v8df): Ditto.
3458 (reduc_splus_v16sf): Ditto.
3459 (avx512f_vextract<shuffletype>32x4_1): Ditto.
3460 (vec_extract_hi_<mode>): Ditto.
3461 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
3462 (vec_set_lo_<mode>): Ditto.
3463 (vec_set_hi_<mode>): Ditto.
3464 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
3465 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
3466 (avx512f_pshufd_1): Ditto.
3467 (avx512f_broadcast<mode>): Ditto.
3468 (avx512f_broadcast<mode>): Ditto.
3469 (define_split): Split vec_extract_lo into move.
3470 (ssequartermode): Ditto.
3471 (ssedoublemode): Extened with wider modes.
3472 (vec_extract_lo_<mode>): Ditto.
3474 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3475 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3476 Sergey Lega <sergey.s.lega@intel.com>
3477 Anna Tikhonova <anna.tikhonova@intel.com>
3478 Ilya Tocar <ilya.tocar@intel.com>
3479 Andrey Turetskiy <andrey.turetskiy@intel.com>
3480 Ilya Verbin <ilya.verbin@intel.com>
3481 Kirill Yukhin <kirill.yukhin@intel.com>
3482 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3484 * config/i386/predicates.md (register_or_constm1_operand): New.
3485 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
3486 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
3487 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
3488 UNSPEC_SCATTER_PREFETCH
3490 (avx512f_ucmp<mode>3): Ditto.
3491 (avx512f_vternlog<mode>): Ditto.
3492 (avx512f_align<mode>): Ditto.
3493 (<shift_insn><mode>3): Ditto.
3494 (avx512f_<rotate>v<mode>): Ditto.
3495 (avx512f_<rotate><mode>): Ditto.
3496 (avx512f_eq<mode>3): Ditto.
3497 (avx512f_eq<mode>3_1): Ditto.
3498 (avx512f_gt<mode>3): Ditto.
3499 (avx512f_testm<mode>3): Ditto.
3500 (avx512f_testnm<mode>3): Ditto.
3501 (avx512pf_gatherpf<mode>): Ditto.
3502 (*avx512pf_gatherpf<mode>_mask): Ditto.
3503 (*avx512pf_gatherpf<mode>): Ditto.
3504 (avx512pf_scatterpf<mode>): Ditto.
3505 (*avx512pf_scatterpf<mode>_mask): Ditto.
3506 (*avx512pf_scatterpf<mode>): Ditto.
3507 (avx512f_vec_dup_gpr<mode>): Ditto.
3508 (clz<mode>2): Ditto.
3509 (conflict<mode>): Ditto.
3510 (REDUC_SMINMAX_MODE): Extened with wider modes.
3511 (reduc_<code>_<mode>): Ditto.
3512 (vlshr<mode>3): Ditto.
3513 (vashl<mode>3): Ditto.
3515 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3516 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3517 Sergey Lega <sergey.s.lega@intel.com>
3518 Anna Tikhonova <anna.tikhonova@intel.com>
3519 Ilya Tocar <ilya.tocar@intel.com>
3520 Andrey Turetskiy <andrey.turetskiy@intel.com>
3521 Ilya Verbin <ilya.verbin@intel.com>
3522 Kirill Yukhin <kirill.yukhin@intel.com>
3523 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3525 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
3528 (avx512fmaskmode): Ditto.
3529 (bcstscalarsuff): Ditto.
3530 (avx512f_blendm<mode>): Ditto.
3531 (cmp_imm_predicate): Ditto.
3532 (avx512f_cmp<mode>3): Ditto.
3533 (avx512f_vec_dup<mode>): Ditto.
3534 (avx512f_vec_dup_mem<mode>): Ditto.
3535 (avx512f_vpermi2var<mode>3): Ditto.
3536 (avx512f_vpermt2var<mode>3): Ditto.
3537 (vec_init<mode>): Ditto.
3538 (avx512f_gathersi<mode>): Ditto.
3539 (*avx512f_gathersi<mode>): Ditto.
3540 (*avx512f_gathersi<mode>_2): Ditto.
3541 (avx512f_gatherdi<mode>): Ditto.
3542 (*avx512f_gatherdi<mode>): Ditto.
3543 (*avx512f_gatherdi<mode>_2): Ditto.
3544 (avx512f_scattersi<mode>): Ditto.
3545 (*avx512f_scattersi<mode>): Ditto.
3546 (avx512f_scatterdi<mode>): Ditto.
3547 (*avx512f_scatterdi<mode>): Ditto.
3548 (sseintprefix): Extened with wider modes.
3549 (VEC_GATHER_IDXSI): Ditto.
3550 (VEC_GATHER_IDXDI): Ditto.
3551 (VEC_GATHER_SRCDI): Ditto.
3553 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
3554 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3556 * config/arm/t-aprofile: New file.
3557 * config.gcc: Handle --with-multilib-list option.
3559 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
3561 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
3564 2013-10-15 Richard Biener <rguenther@suse.de>
3566 * tree-loop-distribution.c (build_empty_rdg): Inline into
3568 (rdg_flag_vertex): Inline into single user.
3569 (rdg_flag_vertex_and_dependent): Likewise.
3570 (build_rdg_partition_for_vertex): Remove processed bitmap.
3571 (rdg_build_partitions): Simplify.
3573 2013-10-15 Richard Biener <rguenther@suse.de>
3575 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
3576 Restructure forwarding through conversions and copies to
3577 avoid performing copy-propagation the wrong way. Adjust
3578 recursion invocations.
3579 (forward_propagate_addr_expr): Add argument stating if we
3580 are recursing from a single-use.
3581 (ssa_forward_propagate_and_combine): Adjust.
3583 2013-10-14 David Malcolm <dmalcolm@redhat.com>
3585 * dumpfile.h (gcc::dump_manager): New class, to hold state
3586 relating to dumpfile management.
3587 (get_dump_file_name): Remove in favor of method of dump_manager.
3588 (dump_initialized_p): Likewise.
3589 (dump_start): Likewise.
3590 (dump_finish): Likewise.
3591 (dump_switch_p): Likewise.
3592 (dump_register): Likewise.
3593 (get_dump_file_info): Likewise.
3594 * context.c (gcc::context::context): Construct the dump_manager
3596 * context.h (gcc::context::get_dumps): New.
3597 (gcc::context::m_dumps): New.
3598 * coverage.c (coverage_init): Port to dump_manager API.
3599 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
3600 (extra_dump_files_in_use): Likewise.
3601 (extra_dump_files_alloced): Likewise.
3602 (gcc::dump_manager::dump_manager): New.
3603 (dump_register): Convert to...
3604 (gcc::dump_manager::dump_register): ...method, replacing
3605 function-static next_dump with m_next_dump field.
3606 (get_dump_file_info): Convert to...
3607 (gcc::dump_manager::get_dump_file_info): ...method.
3608 (get_dump_file_name): Convert to...
3609 (gcc::dump_manager::get_dump_file_name): ...method.
3610 (dump_start): Convert to...
3611 (gcc::dump_manager::dump_start): ...method.
3612 (dump_finish): Convert to...
3613 (gcc::dump_manager::dump_finish): ...method.
3614 (dump_begin): Replace body with...
3615 (gcc::dump_manager::dump_begin): ...new method.
3616 (dump_phase_enabled_p): Convert to...
3617 (gcc::dump_manager::dump_phase_enabled_p): ...method.
3618 (dump_phase_enabled_p): Convert to...
3619 (gcc::dump_manager::dump_phase_enabled_p): ...method.
3620 (dump_initialized_p): Convert to...
3621 (gcc::dump_manager::dump_initialized_p): ...method.
3622 (dump_flag_name): Replace body with...
3623 (gcc::dump_manager::dump_flag_name): ...new method.
3624 (dump_enable_all): Convert to...
3625 (gcc::dump_manager::dump_enable_all): ...new method.
3626 (opt_info_enable_passes): Convert to...
3627 (gcc::dump_manager::opt_info_enable_passes): ...new method.
3628 (dump_switch_p_1): Convert to...
3629 (gcc::dump_manager::dump_switch_p_1): ...new method.
3630 (dump_switch_p): Convert to...
3631 (gcc::dump_manager::dump_switch_p): ...new method.
3632 (opt_info_switch_p): Port to dump_manager API.
3633 (enable_rtl_dump_file): Likewise.
3634 * opts-global.c (handle_common_deferred_options): Port to new
3636 * passes.c (pass_manager::finish_optimization_passes): Likewise.
3637 (pass_manager::register_one_dump_file): Likewise.
3638 (pass_manager::register_pass): Likewise.
3639 (pass_init_dump_file): Likewise.
3640 (pass_fini_dump_file): Likewise.
3641 * statistics.c (statistics_early_init): Likewise.
3643 2013-10-14 Richard Biener <rguenther@suse.de>
3645 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
3646 iterative_hash_canonical_type, gimple_canonical_type_hash,
3647 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
3648 gimple_register_canonical_type, print_gimple_types_stats,
3649 free_gimple_type_tables): Move to lto/lto.c
3650 (gt-gimple.h): Do not include.
3651 * gimple.h (gimple_register_canonical_type,
3652 print_gimple_types_stats, free_gimple_type_tables): Remove.
3653 * Makefile.in (GTFILES): Remove gimple.c.
3655 2013-10-14 Travis Snoozy <quandary@remstate.com>
3658 * config/msp430/msp430.c (msp430_option_override): Correct thinko
3659 scanning for msp430x targets.
3661 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
3664 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
3665 (registers_ok_for_ldd_peep): Move around.
3666 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
3667 * config/sparc/sparc.md (widening peepholes): Use it.
3669 2013-10-14 Richard Biener <rguenther@suse.de>
3673 * gimple.c (iterative_hash_canonical_type): Make sure to
3674 record the hash into the correct hashtable slot.
3676 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
3678 PR rtl-optimization/58662
3679 * combine.c (try_combine): Take into account death nodes on I2 when
3680 splitting a PARALLEL of two independent SETs. Fix dump message.
3682 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
3685 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
3686 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
3687 * config/sh/t-sh (sh_treg_combine.o): New entry.
3688 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
3689 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
3690 (register_sh_passes): New function. Register sh_treg_combine pass.
3691 (sh_option_override): Invoke it.
3692 (sh_canonicalize_comparison): Handle op0_preserve_value.
3693 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
3694 opportunities. Canonicalize branch condition.
3695 (nott): Allow only if pseudos can be created for non-SH2A.
3697 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
3700 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
3701 (ix86_expand_movmem): Replace copy_addr_to_reg with
3702 ix86_copy_addr_to_reg.
3703 (ix86_expand_setmem): Likewise.
3705 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
3707 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
3708 provided by ix86_fp_compare_mode instead of CCFPUmode.
3710 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
3712 * config/aarch64/arm_neon.h
3713 (vtbx<1,3>_<psu>8): Fix register constriants.
3715 2013-10-11 Jeff Law <law@redhat.com>
3717 PR tree-optimization/58640
3718 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
3719 threading paths that cross over two loop entry points.
3721 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3723 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
3724 handle vector float as well.
3725 (*vsx_le_perm_load_v4si): Likewise.
3726 (*vsx_le_perm_store_v2di): Likewise.
3727 (*vsx_le_perm_store_v4si): Likewise.
3729 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3731 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
3732 directly to circumvent subtract from splat{31} workaround.
3733 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
3735 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
3736 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
3737 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
3738 (altivec_vperm_<mode>): Convert to define_insn_and_split to
3739 separate big and little endian logic.
3740 (*altivec_vperm_<mode>_internal): New define_insn.
3741 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
3742 separate big and little endian logic.
3743 (*altivec_vperm_<mode>_uns_internal): New define_insn.
3744 (vec_permv16qi): Add little endian logic.
3746 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
3748 * doc/extend.texi (returns_nonnull): Remove arguments.
3750 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3751 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3752 Sergey Lega <sergey.s.lega@intel.com>
3753 Anna Tikhonova <anna.tikhonova@intel.com>
3754 Ilya Tocar <ilya.tocar@intel.com>
3755 Andrey Turetskiy <andrey.turetskiy@intel.com>
3756 Ilya Verbin <ilya.verbin@intel.com>
3757 Kirill Yukhin <kirill.yukhin@intel.com>
3758 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3760 * config/i386/sse.md (VI48F_256_512): New.
3761 (avx2_permvar<mode>): Change to ...
3762 (<avx2_avx512f>_permvar<mode>): This.
3764 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3765 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3766 Sergey Lega <sergey.s.lega@intel.com>
3767 Anna Tikhonova <anna.tikhonova@intel.com>
3768 Ilya Tocar <ilya.tocar@intel.com>
3769 Andrey Turetskiy <andrey.turetskiy@intel.com>
3770 Ilya Verbin <ilya.verbin@intel.com>
3771 Kirill Yukhin <kirill.yukhin@intel.com>
3772 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3774 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
3775 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
3776 * config/i386/sse.md (VI4_AVX): New.
3777 (sf2simodelower): Ditto.
3778 (sse2_cvtps2dq): Change to ...
3779 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
3781 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3782 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3783 Sergey Lega <sergey.s.lega@intel.com>
3784 Anna Tikhonova <anna.tikhonova@intel.com>
3785 Ilya Tocar <ilya.tocar@intel.com>
3786 Andrey Turetskiy <andrey.turetskiy@intel.com>
3787 Ilya Verbin <ilya.verbin@intel.com>
3788 Kirill Yukhin <kirill.yukhin@intel.com>
3789 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3791 * config/i386/sse.md (V_512): New.
3793 (vcond<V_512:mode><VF_512:mode>): Ditto.
3794 (vcond<V_512:mode><VI_512:mode>): Ditto.
3795 (vcondu<V_512:mode><VI_512:mode>): Ditto.
3797 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3798 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3799 Sergey Lega <sergey.s.lega@intel.com>
3800 Anna Tikhonova <anna.tikhonova@intel.com>
3801 Ilya Tocar <ilya.tocar@intel.com>
3802 Andrey Turetskiy <andrey.turetskiy@intel.com>
3803 Ilya Verbin <ilya.verbin@intel.com>
3804 Kirill Yukhin <kirill.yukhin@intel.com>
3805 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3807 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
3808 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
3810 (fma<mode>4): Removed condition.
3811 (fms<mode>4): Ditto.
3812 (fnma<mode>4): Ditto.
3813 (fnms<mode>4): Ditto.
3814 (fma4i_fmadd_<mode>): Ditto.
3815 (*fma_fmadd_<mode>): Ditto.
3816 (*fma_fmsub_<mode>): Ditto.
3817 (*fma_fnmadd_<mode>): Ditto.
3818 (*fma_fnmsub_<mode>): Ditto.
3819 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
3820 (*fma_fmaddsub_<mode>): Ditto.
3821 (*fma_fmsubadd_<mode>): Ditto.
3822 (*fmai_fmadd_<mode>): Ditto.
3823 (*fmai_fmsub_<mode>): Ditto.
3824 (*fmai_fnmadd_<mode>): Ditto.
3825 (*fmai_fnmsub_<mode>): Ditto.
3827 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3828 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3829 Sergey Lega <sergey.s.lega@intel.com>
3830 Anna Tikhonova <anna.tikhonova@intel.com>
3831 Ilya Tocar <ilya.tocar@intel.com>
3832 Andrey Turetskiy <andrey.turetskiy@intel.com>
3833 Ilya Verbin <ilya.verbin@intel.com>
3834 Kirill Yukhin <kirill.yukhin@intel.com>
3835 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3837 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
3838 (VI124_256): Changed to ...
3839 (VI124_256_48_512): This.
3840 (ssepackmode): Extended with wider modes.
3841 (<code><mode>3): Changed iterator.
3842 (*avx2_<code><mode>3): Ditto.
3843 (vec_pack_trunc_<mode>): Ditto.
3845 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3846 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3847 Sergey Lega <sergey.s.lega@intel.com>
3848 Anna Tikhonova <anna.tikhonova@intel.com>
3849 Ilya Tocar <ilya.tocar@intel.com>
3850 Andrey Turetskiy <andrey.turetskiy@intel.com>
3851 Ilya Verbin <ilya.verbin@intel.com>
3852 Kirill Yukhin <kirill.yukhin@intel.com>
3853 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3855 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
3856 (VI8F_256_512): Ditto.
3857 (abs<mode>2): Changed iterator.
3858 (avx2_perm<mode>): Changed to ...
3859 (<avx2_avx512f>_perm<mode>): This.
3860 (avx2_perm<mode>_1): Changed to ...
3861 (<avx2_avx512f>_perm<mode>_1): This.
3863 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3864 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3865 Sergey Lega <sergey.s.lega@intel.com>
3866 Anna Tikhonova <anna.tikhonova@intel.com>
3867 Ilya Tocar <ilya.tocar@intel.com>
3868 Andrey Turetskiy <andrey.turetskiy@intel.com>
3869 Ilya Verbin <ilya.verbin@intel.com>
3870 Kirill Yukhin <kirill.yukhin@intel.com>
3871 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3873 * config/i386/sse.md (VI48_AVX512F): New.
3874 (VI48_AVX2): Changed to ...
3875 (VI48_AVX2_48_AVX512F): This.
3876 (avx2_ashrv<mode>): Changed to ...
3877 (<avx2_avx512f>_ashrv<mode>): This.
3878 (avx2_<shift_insn>v<mode>): Changed to ...
3879 (<avx2_avx512f>_<shift_insn>v<mode>): This.
3881 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3882 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3883 Sergey Lega <sergey.s.lega@intel.com>
3884 Anna Tikhonova <anna.tikhonova@intel.com>
3885 Ilya Tocar <ilya.tocar@intel.com>
3886 Andrey Turetskiy <andrey.turetskiy@intel.com>
3887 Ilya Verbin <ilya.verbin@intel.com>
3888 Kirill Yukhin <kirill.yukhin@intel.com>
3889 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3891 * config/i386/sse.md (VI4_AVX512F): New.
3892 (VI8_AVX2_AVX512F): Ditto.
3893 (mul<mode>3): Extended with wider modes.
3894 (*<sse4_1_avx2>_mul<mode>3): Ditto.
3895 (mul<mode>3): Ditto.
3896 (vec_widen_<s>mult_odd_<mode>): Ditto.
3898 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3899 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3900 Sergey Lega <sergey.s.lega@intel.com>
3901 Anna Tikhonova <anna.tikhonova@intel.com>
3902 Ilya Tocar <ilya.tocar@intel.com>
3903 Andrey Turetskiy <andrey.turetskiy@intel.com>
3904 Ilya Verbin <ilya.verbin@intel.com>
3905 Kirill Yukhin <kirill.yukhin@intel.com>
3906 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3908 * config/i386/sse.md (VI2_AVX512F): New.
3909 (VI124_AVX512F): Ditto.
3910 (sseunpackmode): Extended with wider modes.
3911 (sseunpackfltmode): Ditto.
3912 (vec_unpacks_float_hi_<mode>): Ditto.
3913 (vec_unpacks_float_lo_<mode>): Ditto.
3914 (vec_unpacku_float_hi_<mode>): Ditto.
3915 (vec_unpacku_float_lo_<mode>): Ditto.
3916 (vec_unpacks_lo_<mode>): Ditto.
3917 (vec_unpacks_hi_<mode>): Ditto.
3918 (vec_unpacku_lo_<mode>): Ditto.
3919 (vec_unpacku_hi_<mode>): Ditto.
3921 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3922 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3923 Sergey Lega <sergey.s.lega@intel.com>
3924 Anna Tikhonova <anna.tikhonova@intel.com>
3925 Ilya Tocar <ilya.tocar@intel.com>
3926 Andrey Turetskiy <andrey.turetskiy@intel.com>
3927 Ilya Verbin <ilya.verbin@intel.com>
3928 Kirill Yukhin <kirill.yukhin@intel.com>
3929 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3931 * config/i386/i386.md (multdiv): New.
3932 (multdiv_mnemonic): Ditto.
3933 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
3934 (<sse>_vm<multdiv_mnemonic><mode>3): This.
3935 (<sse>_vmdiv<mode>3): Removed.
3937 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3938 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3939 Sergey Lega <sergey.s.lega@intel.com>
3940 Anna Tikhonova <anna.tikhonova@intel.com>
3941 Ilya Tocar <ilya.tocar@intel.com>
3942 Andrey Turetskiy <andrey.turetskiy@intel.com>
3943 Ilya Verbin <ilya.verbin@intel.com>
3944 Kirill Yukhin <kirill.yukhin@intel.com>
3945 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3947 * config/i386/sse.md (V): Extended with wider modes.
3949 (ssehalfvecmode): Ditto.
3951 (ssepackfltmode): Ditto.
3952 (avx_vec_concat<mode>): Ditto.
3953 (V_256_512): New iterator.
3954 (VF2_512_256): Ditto.
3955 (si2dfmode): New attribute.
3956 (si2dfmodelower): Ditto.
3958 (concat_tg_mode): Ditto.
3959 (floatv4siv4df2): Changed to ...
3960 (float<si2dfmodelower><mode>2): This.
3961 (avx_cvtps2pd256): Changed to ...
3962 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
3963 (vec_pack_trunc_v4df): Changed to ...
3964 (vec_pack_trunc_<mode>): This.
3965 (avx_vpermil<mode>): Changed to ...
3966 (<sse2_avx_avx512f>_vpermil<mode>): This.
3967 (<fixsuffix>fix_truncv8dfv8si2): New.
3968 (vec_pack_sfix_trunc_v8df): Ditto.
3969 (avx512f_rndscale<mode>): Ditto.
3970 (avx512f_roundpd512): Ditto.
3971 (vec_pack_ufix_trunc_<mode>): Updated iterator.
3973 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3974 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3975 Sergey Lega <sergey.s.lega@intel.com>
3976 Anna Tikhonova <anna.tikhonova@intel.com>
3977 Ilya Tocar <ilya.tocar@intel.com>
3978 Andrey Turetskiy <andrey.turetskiy@intel.com>
3979 Ilya Verbin <ilya.verbin@intel.com>
3980 Kirill Yukhin <kirill.yukhin@intel.com>
3981 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3983 * config/i386/i386.md (any_fix): New iterator.
3984 (fixsuffix): New attribute.
3985 * config/i386/sse.md (VF1): Extened with wider modes.
3989 (sseintvecmodelower): Ditto.
3990 (ssescalarmode): Ditto.
3991 (ssescalarnum): Ditto.
3993 (ssexmmmode): Ditto.
3994 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
3995 (<sse>_rcp<mode>2): Change iterator.
3996 (rsqrt<mode>2): Ditto.
3997 (<sse>_rsqrt<mode>2): Ditto.
3998 (avx2_vec_dup<mode>): Ditto.
3999 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4000 (round<mode>2_sfix): Ditto.
4001 (avx2_pbroadcast<mode>): Ditto.
4002 (*andnot<mode>3): Handle XI mode.
4003 (*<code><mode>3): Ditto.
4004 (AVXTOSSEMODE): Removed.
4005 (avx_vpermil<mode>): Changed to ...
4006 (<sse2_avx_avx512f>_vpermil<mode>): This.
4008 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4009 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4010 Sergey Lega <sergey.s.lega@intel.com>
4011 Anna Tikhonova <anna.tikhonova@intel.com>
4012 Ilya Tocar <ilya.tocar@intel.com>
4013 Andrey Turetskiy <andrey.turetskiy@intel.com>
4014 Ilya Verbin <ilya.verbin@intel.com>
4015 Kirill Yukhin <kirill.yukhin@intel.com>
4016 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4018 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4019 (<sse>_comi): Ditto.
4020 (<sse>_ucomi): Ditto.
4021 (sse_cvtss2siq_2): Ditto.
4022 (sse2_cvtsd2si): Ditto.
4023 (sse2_cvtsd2siq): Ditto.
4024 (sse2_cvttsd2si): Ditto.
4025 (sse2_cvttsd2siq): Ditto.
4026 (<shift_insn><mode>3): Ditto.
4027 (sse2_cvtsi2sdq): Update constraint and prefix.
4028 (sse_cvtsi2ss): Update prefix.
4029 (sse_cvtsi2ssq): Ditto.
4031 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4033 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4034 ignore internal calls.
4036 2013-10-11 Richard Biener <rguenther@suse.de>
4038 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4039 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4040 TREE_OVERFLOW is set.
4042 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4044 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4046 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4047 * gimple.h: Likewise.
4049 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4051 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4054 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4056 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4057 * configure: Regenerate.
4059 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4061 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4062 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4063 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4064 similarly pp_right_paren.
4065 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4066 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4067 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4068 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4070 (omp_declare_simd_clauses_equal,
4071 omp_remove_redundant_declare_simd_attrs): New functions.
4072 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4073 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4074 * tree.h (OMP_LOOP_CHECK): Define.
4075 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4076 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4077 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4078 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4079 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4080 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4081 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4082 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4083 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4084 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4085 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4086 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4087 (omp_remove_redundant_declare_simd_attrs): New prototype.
4088 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4089 GIMPLE_OMP_TEAMS): New codes.
4090 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4091 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4093 (target_nesting_level): New variable.
4094 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4095 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4096 for collapse > 1 static schedule unless ordered.
4097 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4098 (determine_parallel_type): Adjust get_ws_args_for caller.
4099 (install_var_field): Handle mask & 4 for double indirection.
4100 (scan_sharing_clauses): Ignore shared clause on teams construct.
4101 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4102 (create_omp_child_function): If inside target or declare target
4103 constructs, set "omp declare target" attribute on the child function.
4104 (find_combined_for): New function.
4105 (scan_omp_parallel): Handle combined loops.
4106 (scan_omp_target, scan_omp_teams): New functions.
4107 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4108 restrictions and set ctx->cancellable for cancellable constructs.
4109 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4110 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4111 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4112 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4113 (omp_clause_aligned_alignment): New function.
4114 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4115 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4116 on teams constructs. Handle user defined reductions and new
4118 (lower_reduction_clauses): Don't set placeholder to address of ref
4119 if it has already the right type.
4120 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4121 (expand_parallel_call): Use the new non-_start suffixed builtins,
4122 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4123 and GOMP_parallel_end after the call.
4124 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4125 (expand_omp_for_init_counts): Handle combined loops.
4126 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4128 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4129 end of cancellable loops.
4130 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4131 Likewise. Handle collapse > 1 loops.
4132 (expand_omp_simd): Handle combined loops.
4133 (expand_omp_for): Add inner_stmt argument, adjust callers of
4134 expand_omp_for* functions, use expand_omp_for_static*chunk even
4135 for collapse > 1 unless ordered.
4136 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4137 of cancellable sections.
4138 (expand_omp_single): Remove need_barrier variable, just rely on
4139 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4140 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4141 (expand_omp_atomic_load, expand_omp_atomic_store,
4142 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4143 (expand_omp_target): New function.
4144 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4145 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4146 (build_omp_regions_1): Immediately close region for
4147 GF_OMP_TARGET_KIND_UPDATE.
4148 (maybe_add_implicit_barrier_cancel): New function.
4149 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4151 (lower_omp_single): Likewise. Add clobber after the barrier.
4152 (lower_omp_taskgroup): New function.
4153 (lower_omp_for): Handle combined loops. Adjust
4154 lower_rec_input_clauses caller. Handle cancellation.
4155 (lower_depend_clauses): New function.
4156 (lower_omp_taskreg): Lower depend clauses. Adjust
4157 lower_rec_input_clauses caller. Add clobber after the call. Handle
4159 (lower_omp_target, lower_omp_teams): New functions.
4160 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4161 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4162 and GOMP_cancellation_point calls.
4163 (lower_omp): Fold stmts inside of target region.
4164 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4165 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4166 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4167 (BT_FN_VOID_OMPFN_PTR_UINT,
4168 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4169 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4170 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4171 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4172 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4173 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4174 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4175 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4176 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4177 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4178 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4179 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4180 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4181 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4182 BUILT_IN_GOMP_PARALLEL_END.
4183 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4184 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4185 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4186 GF_OMP_FOR_KIND_DISTRIBUTE.
4187 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4188 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4189 (dump_gimple_omp_return): Print lhs if it has any.
4190 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4191 gimple_omp_atomic_seq_cst_p.
4192 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4193 and GIMPLE_OMP_TEAMS.
4194 * langhooks.c (lhd_omp_mappable_type): New function.
4195 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4196 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4198 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4199 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4200 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4201 (struct gimplify_omp_ctx): Add combined_loop field.
4202 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4203 on stmts inside of target region.
4204 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4205 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4206 ORT_TARGET and ORT_TARGET_DATA.
4207 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4209 (omp_notice_threadprivate_variable): Complain about threadprivate
4210 variables in target region.
4211 (omp_notice_variable): Complain about vars with non-mappable type
4212 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4213 (omp_check_private): Ignore ORT_TARGET* regions.
4214 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4215 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
4216 (find_combined_omp_for): New function.
4217 (gimplify_omp_for): Handle gimplification of combined loops.
4218 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
4220 (gimplify_omp_target_update): New function.
4221 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
4222 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
4223 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
4224 (gimplify_body): If fndecl has "omp declare target" attribute, add
4225 implicit ORT_TARGET context around it.
4226 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
4227 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
4228 * tree-nested.c (convert_nonlocal_reference_stmt,
4229 convert_local_reference_stmt, convert_gimple_call): Handle
4230 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4231 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
4232 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
4233 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
4234 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
4235 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
4236 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
4237 BUILT_IN_GOMP_LOOP_END_CANCEL,
4238 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
4239 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
4240 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
4241 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
4242 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
4243 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
4244 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
4245 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
4246 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
4247 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
4248 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
4249 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
4250 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
4251 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
4252 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
4253 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4254 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
4255 gimple_build_omp_teams): New functions.
4256 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
4257 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
4258 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
4259 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4260 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
4261 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
4262 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
4263 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
4264 GF_OMP_ATOMIC_SEQ_CST): New.
4265 (gimple_build_omp_taskgroup, gimple_build_omp_target,
4266 gimple_build_omp_teams): New prototypes.
4267 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
4268 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4269 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
4270 GIMPLE_OMP_SINGLE as end of range.
4271 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
4272 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
4273 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
4274 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
4275 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
4276 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
4277 gimple_omp_target_kind, gimple_omp_target_set_kind,
4278 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
4279 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
4280 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
4281 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
4282 gimple_omp_teams_set_clauses): New inlines.
4283 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
4284 and GIMPLE_OMP_TASKGROUP.
4285 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
4286 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
4287 enum omp_clause_proc_bind_kind): New.
4288 (union omp_clause_subcode): Add depend_kind, map_kind and
4289 proc_bind_kind fields.
4290 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
4291 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4292 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
4293 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
4294 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
4296 2013-10-10 Teresa Johnson <tejohnson@google.com>
4298 * predict.c (tree_estimate_probability): Add new parameter
4299 for estimate_bb_frequencies.
4300 (estimate_bb_frequencies): Add new parameter to force estimation.
4301 (rebuild_frequencies): When max frequency in function is small,
4302 recompute counts from frequencies.
4303 * predict.h (estimate_bb_frequencies): New parameter.
4305 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4307 * ipa-inline.c (ipa_inline): Fix leak of "order" when
4308 optimizations are disabled.
4310 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4312 * coverage.c (coverage_finish): Fix leak of da_file_name.
4314 2013-10-10 Jan Hubicka <jh@suse.cz>
4316 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
4317 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
4320 2013-10-10 Jakub Jelinek <jakub@redhat.com>
4323 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
4324 if any labels are in FALLTHRU_BB, use a special label emitted
4325 immediately after the asm goto insn rather than label_rtx
4327 (expand_asm_stmt): Adjust caller.
4328 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
4329 edge if the last insn in predecessor is a jump with single successor,
4330 but it isn't simplejump_p.
4332 2013-10-10 Richard Biener <rguenther@suse.de>
4334 PR tree-optimization/58656
4335 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
4337 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4339 * gimplify.c: Include expr.h and tm_p.h for targets with special
4340 va-arg padding requirements.
4342 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4344 * tree-flow.h: Move some prototypes to gimple.h.
4345 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
4346 * gimple.h: Relocate some prototypes from tree-flow.h
4347 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
4349 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
4350 (build_va_arg_indirect_ref): Relocate and make static.
4351 (std_gimplify_va_arg_expr): Relocate here.
4352 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
4353 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
4355 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4357 * doc/md.texi: Document the mnemonic attribute.
4359 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4362 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
4363 (set_attr_alternative x ...) when searching for user defined
4366 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4368 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
4370 2013-10-09 Easwaran Raman <eraman@google.com>
4372 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
4373 * cfgexpand.c (defer_stack_allocation): ...use here
4374 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
4376 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
4378 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
4379 (operand_equal_for_value_replacement): New function, extracted from
4380 value_replacement and enhanced to catch more cases.
4381 (value_replacement): Use operand_equal_for_value_replacement.
4383 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4385 * loop-doloop.c (doloop_modify, doloop_optimize): Use
4386 get_max_loop_iterations.
4388 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4390 * config/arm/aarch-common.c (arm_early_load_addr_dep):
4391 Place comment above function.
4393 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4395 * tree-flow.h: Remove all remaining prototypes, enums and structs that
4396 are not related to tree-cfg.c.
4397 * tree-ssa-address.h: New file. Relocate prototypes.
4398 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
4399 (addr_for_mem_ref): New. Combine call to get_address_description and
4400 return addr_for_mem_ref.
4401 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
4402 * tree-ssa-live.h: Adjust prototypes.
4403 * passes.c: Include tree-ssa-live.h.
4404 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
4405 * graphite.c (graphite_transform_loops): Make static.
4406 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
4407 make_pass_graphite, pass_data_graphite_transforms,
4408 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
4409 * ipa-pure-const.c (warn_function_noreturn): Make static.
4410 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4411 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4412 Relocate from tree-cfg.c
4413 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
4415 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4416 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4417 Move to ipa-pure-const.c.
4418 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
4419 Relocate from tree-optimize.c.
4420 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
4421 make_pass_fixup_cfg): Move to tree-cfg.c.
4422 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
4423 Relocate some prototypes.
4424 * tree-data-ref.h (tree_check_data_deps) Add prototype.
4425 * tree-dump.c (dump_function_to_file): Remove prototype.
4426 Add tree-flow.h to the include file.
4427 * tree-dump.h: Remove prototype.
4428 * tree-parloops.h: New File. Add prototypes.
4429 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
4430 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
4431 from tree-ssa-loop.c.
4432 * tree-predcom.c (run_tree_predictive_commoning,
4433 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
4434 Relocate here from tree-ssa-loop.c.
4435 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
4436 ssa_name_values.release ().
4437 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
4438 (ssa_name_values): Relocate from tree-flow.h.
4439 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
4440 * tree-ssa-loop.c (run_tree_predictive_commoning,
4441 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
4442 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
4443 make_pass_graphite, pass_data_graphite_transforms,
4444 make_pass_graphite_transforms, gate_tree_parallelize_loops,
4445 tree_parallelize_loops, pass_data_parallelize_loops,
4446 make_pass_parallelize_loops): Move to other files.
4447 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
4449 * tree.h: Remove prototypes from tree-address.c.
4451 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4453 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
4454 (struct int_tree_map): Move to tree-hasher.h
4455 (SCALE, LABEL, PERCENT): Move to gimple.h
4456 * tree-flow-inline.h: Delete. Move functions to other files.
4457 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
4458 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
4459 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
4460 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
4461 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
4462 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
4463 * tree-hasher.h: Don't include tree-flow.h.
4464 (struct int_tree_map): Relocate from tree-flow.h.
4465 * tree-sra.c (contains_view_convert_expr_p): Relocate from
4466 tree-flow-inline.h and make static.
4467 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
4469 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
4470 tree-flow-inline.h and make static.
4471 * tree.h (is_global_var, may_be_aliased): Relocate from
4473 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
4474 * value-prof.c: No longer include tree-flow-inline.h.
4475 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
4477 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4479 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
4480 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
4481 (enum move_pos): Move to tree-ssa-loop-im.h
4482 * cfgloop.h: Move some prototypes.
4483 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
4484 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
4485 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
4486 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
4487 (loop_containing_stmt): Relocate from tree-flow-inline.h.
4488 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
4489 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
4490 (enum move_pos): Relocate here.
4491 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
4493 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
4494 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
4495 make_pass_lim): Relocate here from tree-ssa-loop.c.
4496 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
4498 (loop_edge_to_cancel, unloop_loops): Make static.
4499 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
4500 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
4501 (tree_complete_unroll, gate_tree_complete_unroll,
4502 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
4503 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
4504 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
4505 * tree-ssa-loop-ivopts.c: Remove local prototypes.
4506 (stmt_invariant_in_loop_p): Remove unused function.
4507 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
4508 * tree-ssa-loop-manip.h: New file. Add prototypes.
4509 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
4510 (gcov_type_to_double_int): Move to cfgloop.h.
4511 (double_int_cmp, bound_index,
4512 estimate_numbers_of_iterations_loop): Make static.
4513 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
4514 (max_loop_iterations): Factor out get_max_loop_iterations.
4515 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
4517 * tree-ssa-loop-niter.h: New file. Add prototypes.
4518 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
4519 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
4520 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
4521 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
4522 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
4523 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
4524 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
4525 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
4526 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
4527 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
4528 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
4529 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
4530 pass_data_complete_unroll, make_pass_complete_unroll,
4531 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
4532 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
4533 tree-ssa-loop-ivcanon.c.
4534 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
4535 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
4536 tree-ssa-loop-prefetch.c.
4537 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
4539 (get_lsm_tmp_name): Relocate and add suffix parameter.
4540 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
4541 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
4542 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
4543 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
4544 (get_estimated_loop_iterations): Factor out accessor from
4545 estimated_loop_iterations in tree-ssa-loop-niter.c.
4546 (get_max_loop_iterations): Factor out accessor from
4547 _max_loop_iterations in tree-ssa-niter.c.
4548 * loop-unroll.c (decide_unroll_constant_iterations,
4549 decide_unroll_runtime_iterations, decide_peel_simple,
4550 decide_unroll_stupid): Use new get_* accessors.
4552 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
4554 PR tree-optimization/20318
4555 * doc/extend.texi (returns_nonnull): New function attribute.
4556 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
4558 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
4559 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
4561 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
4564 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
4565 false if both components are bitfields.
4567 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4569 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
4570 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
4571 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
4572 * config/aarch64/aarch64.h
4573 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
4574 * config/aarch64/aarch64-simd-builtins.def
4575 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
4577 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4579 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
4581 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4583 * config/aarch64/arm_neon.h (vdiv_f64): Added.
4585 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4587 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
4588 (vneg_f64): New intrinsic.
4589 (vneg_s8): Asm replaced with C.
4590 (vneg_s16): Likewise.
4591 (vneg_s32): Likewise.
4592 (vneg_s64): New intrinsic.
4593 (vnegq_f32): Asm replaced with C.
4594 (vnegq_f64): Likewise.
4595 (vnegq_s8): Likewise.
4596 (vnegq_s16): Likewise.
4597 (vnegq_s32): Likewise.
4598 (vnegq_s64): Likewise.
4600 2013-10-09 Renlin Li <Renlin.Li@arm.com>
4602 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
4604 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4606 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
4607 packed stack special handling.
4608 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
4609 back to fixed stack slots for FPRs saved due to stdarg.
4611 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4613 * config/s390/s390.c (s390_frame_info): Restructure function.
4615 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4617 * config/s390/s390.c (struct s390_frame_layout): New field
4619 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
4620 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
4621 regs_ever_clobbered to char*.
4622 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
4623 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
4624 (s390_register_info): Enable FPR save slots. Move/Copy some
4625 functionality into ...
4626 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
4627 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
4629 (s390_frame_info): Do gpr slot allocation here now. stdarg does
4630 not imply a stack frame.
4631 (s390_init_frame_layout): Remove variable clobbered_regs.
4632 (s390_update_register_info): Remove function.
4633 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
4634 cannot be used for register renaming.
4635 (s390_hard_regno_scratch_ok): New function.
4636 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
4637 (s390_initial_elimination_offset): Change offset calculation of
4638 the return address pointer.
4639 (save_gprs): Deal with only r6 being saved from the call-saved regs.
4640 (restore_gprs): Set frame related flag.
4641 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
4642 (s390_emit_prologue): Call s390_register_info instead of
4643 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
4644 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
4645 (s390_optimize_prologue): Call s390_optimize_register_info.
4646 Try to remove also FPR slot save/restore INSNs. Remove frame
4647 related flags from restore INSNs.
4649 2013-10-08 DJ Delorie <dj@redhat.com>
4651 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
4654 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
4655 avoid conflict with the MI use of %c.
4656 * config/rl78/rl78-real.md: change %c to %C throughout.
4657 * config/rl78/rl78-virt.md: Likewise.
4659 2013-10-08 Jan Hubicka <jh@suse.cz>
4661 * config/i386/i386.c (ix86_option_override_internal): Switch
4662 to SSE math for -ffast-math when target ISA supports SSE2.
4664 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
4666 * tree-flow.h: Remove some prototypes.
4667 * tree.h: Remove some protypes, add a couple.
4668 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
4669 using_eh_for_cleanups_p): Add interface routines for front ends.
4670 * tree-eh.h: New file. Add protoptyes.
4671 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
4672 (add_stmt_to_eh_lp_fn): Make static.
4673 (lower_try_finally): Use new using_eh_for_cleanups_p.
4674 * emit-rtl.c: Include tree-eh.h.
4675 * gimple.h: Include tree-eh.h.
4677 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
4679 PR tree-optimization/58480
4680 * tree-vrp.c (infer_nonnull_range): New function.
4681 (infer_value_range): Call infer_nonnull_range.
4683 2013-10-08 Dehao Chen <dehao@google.com>
4685 PR tree-optimization/58619
4686 * tree-inline.c (copy_phis_for_bb): Combine location data
4689 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4692 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
4693 RTX_FRAME_RELATED_P on INSN.
4695 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4697 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
4698 (altivec_expand_vec_perm_const): Call it.
4700 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4702 * config/rs6000/vector.md (mov<mode>): Emit permuted move
4703 sequences for LE VSX loads and stores at expand time.
4704 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
4706 * config/rs6000/rs6000.c (rs6000_const_vec): New.
4707 (rs6000_gen_le_vsx_permute): New.
4708 (rs6000_gen_le_vsx_load): New.
4709 (rs6000_gen_le_vsx_store): New.
4710 (rs6000_gen_le_vsx_move): New.
4711 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
4712 (*vsx_le_perm_load_v4si): New.
4713 (*vsx_le_perm_load_v8hi): New.
4714 (*vsx_le_perm_load_v16qi): New.
4715 (*vsx_le_perm_store_v2di): New.
4716 (*vsx_le_perm_store_v4si): New.
4717 (*vsx_le_perm_store_v8hi): New.
4718 (*vsx_le_perm_store_v16qi): New.
4719 (*vsx_xxpermdi2_le_<mode>): New.
4720 (*vsx_xxpermdi4_le_<mode>): New.
4721 (*vsx_xxpermdi8_le_V8HI): New.
4722 (*vsx_xxpermdi16_le_V16QI): New.
4723 (*vsx_lxvd2x2_le_<mode>): New.
4724 (*vsx_lxvd2x4_le_<mode>): New.
4725 (*vsx_lxvd2x8_le_V8HI): New.
4726 (*vsx_lxvd2x16_le_V16QI): New.
4727 (*vsx_stxvd2x2_le_<mode>): New.
4728 (*vsx_stxvd2x4_le_<mode>): New.
4729 (*vsx_stxvd2x8_le_V8HI): New.
4730 (*vsx_stxvd2x16_le_V16QI): New.
4732 2013-10-07 Renlin Li <Renlin.Li@arm.com>
4734 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
4736 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4738 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
4739 loop to work also for 31bit ABI.
4740 Save the stack pointer for frame_size > 0.
4742 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4744 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
4745 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
4746 constraint letters from expanders.
4747 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
4748 retry count to general_operand.
4749 ("tabort"): Give operand 0 a mode.
4750 ("tabort_1"): Add mode and constraint letter for operand 0.
4751 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
4753 2013-10-04 Jeff Law <law@redhat.com>
4755 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
4757 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
4759 (thread_across_edge): Here. Call it.
4761 2013-10-04 Cary Coutant <ccoutant@google.com>
4763 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
4764 discarding a location list expression (or a piece of one).
4766 2013-10-03 Jan Hubicka <jh@suse.cz>
4768 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
4769 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
4770 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
4772 2013-10-03 Jan Hubicka <jh@suse.cz>
4774 * config/i386/i386.c (ix86_option_override_internal): Do not enable
4775 accumulate-outgoing-args when producing unwind info.
4777 2013-10-03 Wei Mi <wmi@google.com>
4779 * lra-constraints.c (insert_move_for_subreg): New function
4780 extracted from simplify_operand_subreg.
4781 (simplify_operand_subreg): Add reload for paradoxical subreg.
4783 2013-10-03 Rong Xu <xur@google.com>
4785 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
4786 the candidate of builtin_expect such that we should fix the
4787 size/time estimation.
4788 (estimate_function_body_sizes): Do the acutally size/time fix-up
4791 2013-10-03 Rong Xu <xur@google.com>
4793 * predict.c (tree_predict_by_opcode): Get the probability
4794 for builtin_expect from param builtin_expect_probability.
4795 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
4796 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
4797 * doc/invoke.texi: Add documentation for builtin-expect-probability.
4799 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
4802 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
4803 * calls.c (alloca_call_p): Use get_callee_fndecl.
4804 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
4805 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
4807 (vrp_visit_stmt): Remove duplicated code.
4809 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
4811 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
4812 ceildf2, btruncdf2, instead of vsx_* name.
4814 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
4815 iterators to only do V2DF and V4SF here. Move the DF code to
4816 rs6000.md where it is combined with SF mode. Replace <VSv> with
4817 just 'v' since only vector operations are handled with these insns
4818 after moving the DF support to rs6000.md.
4819 (vsx_sub<mode>3): Likewise.
4820 (vsx_mul<mode>3): Likewise.
4821 (vsx_div<mode>3): Likewise.
4822 (vsx_fre<mode>2): Likewise.
4823 (vsx_neg<mode>2): Likewise.
4824 (vsx_abs<mode>2): Likewise.
4825 (vsx_nabs<mode>2): Likewise.
4826 (vsx_smax<mode>3): Likewise.
4827 (vsx_smin<mode>3): Likewise.
4828 (vsx_sqrt<mode>2): Likewise.
4829 (vsx_rsqrte<mode>2): Likewise.
4830 (vsx_fms<mode>4): Likewise.
4831 (vsx_nfma<mode>4): Likewise.
4832 (vsx_copysign<mode>3): Likewise.
4833 (vsx_btrunc<mode>2): Likewise.
4834 (vsx_floor<mode>2): Likewise.
4835 (vsx_ceil<mode>2): Likewise.
4836 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
4837 (vsx_sminsf3): Likewise.
4838 (vsx_fmadf4): Likewise.
4839 (vsx_fmsdf4): Likewise.
4840 (vsx_nfmadf4): Likewise.
4841 (vsx_nfmsdf4): Likewise.
4842 (vsx_cmpdf_internal1): Likewise.
4844 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
4845 simpler to select whether a target has SPE or traditional floating
4846 point support in iterators.
4847 (TARGET_DF_SPE): Likewise.
4848 (TARGET_SF_FPR): Likewise.
4849 (TARGET_DF_FPR): Likewise.
4850 (TARGET_SF_INSN): Macros to say whether floating point support
4851 exists for a given operation for expanders.
4852 (TARGET_DF_INSN): Likewise.
4854 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
4855 combining of SF/DF mode operations, using both traditional and VSX
4863 (abs<mode>2): Combine SF/DF modes using traditional floating point
4864 instructions. Add support for using the upper DF registers with
4865 VSX support, and SF registers with power8-vector support. Update
4866 expanders for operations supported by both the SPE and traditional
4867 floating point units.
4868 (abs<mode>2_fpr): Likewise.
4869 (nabs<mode>2): Likewise.
4870 (nabs<mode>2_fpr): Likewise.
4871 (neg<mode>2): Likewise.
4872 (neg<mode>2_fpr): Likewise.
4873 (add<mode>3): Likewise.
4874 (add<mode>3_fpr): Likewise.
4875 (sub<mode>3): Likewise.
4876 (sub<mode>3_fpr): Likewise.
4877 (mul<mode>3): Likewise.
4878 (mul<mode>3_fpr): Likewise.
4879 (div<mode>3): Likewise.
4880 (div<mode>3_fpr): Likewise.
4881 (sqrt<mode>3): Likewise.
4882 (sqrt<mode>3_fpr): Likewise.
4883 (fre<Fs>): Likewise.
4884 (rsqrt<mode>2): Likewise.
4885 (cmp<mode>_fpr): Likewise.
4886 (smax<mode>3): Likewise.
4887 (smin<mode>3): Likewise.
4888 (smax<mode>3_vsx): Likewise.
4889 (smin<mode>3_vsx): Likewise.
4890 (negsf2): Delete SF operations that are merged with DF.
4897 (fmasf4_fpr): Likewise.
4898 (fmssf4_fpr): Likewise.
4899 (nfmasf4_fpr): Likewise.
4900 (nfmssf4_fpr): Likewise.
4901 (sqrtsf2): Likewise.
4902 (rsqrtsf_internal1): Likewise.
4903 (smaxsf3): Likewise.
4904 (sminsf3): Likewise.
4905 (cmpsf_internal1): Likewise.
4906 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
4907 (negdf2): Delete DF operations that are merged with SF.
4909 (nabsdf2): Likewise.
4915 (rsqrtdf_internal1): Likewise.
4916 (fmadf4_fpr): Likewise.
4917 (fmsdf4_fpr): Likewise.
4918 (nfmadf4_fpr): Likewise.
4919 (nfmsdf4_fpr): Likewise.
4920 (sqrtdf2): Likewise.
4921 (smaxdf3): Likewise.
4922 (smindf3): Likewise.
4923 (cmpdf_internal1): Likewise.
4924 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
4925 (btrunc<mode>2): Delete separate expander, and combine with the
4926 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
4927 (btrunc<mode>2_fpr): Likewise.
4928 (ceil<mode>2): Likewise.
4929 (ceil<mode>2_fpr): Likewise.
4930 (floor<mode>2): Likewise.
4931 (floor<mode>2_fpr): Likewise.
4932 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
4933 Add support for using the upper registers with VSX and
4934 power8-vector. Move insns to be closer to the define_expands. On
4935 VSX systems, prefer the traditional form of FMA over the VSX
4936 version, since the traditional form allows the target not to
4937 overlap with the inputs.
4938 (fms<mode>4_fpr): Likewise.
4939 (nfma<mode>4_fpr): Likewise.
4940 (nfms<mode>4_fpr): Likewise.
4942 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4943 Richard Earnshaw <richard.earnshaw@arm.com>
4945 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
4946 (struct mult_cost_table): Likewise.
4947 (struct mem_cost_table): Likewise.
4948 (struct fp_cost_table): Likewise.
4949 (struct vector_cost_table): Likewise.
4950 (cpu_cost_table): Likewise.
4951 * config/arm/arm.opt (mold-rts-costs): New option.
4952 (mnew-generic-costs): Likewise.
4953 * config/arm/arm.c (generic_extra_costs): New table.
4954 (cortexa15_extra_costs): Likewise.
4955 (arm_slowmul_tune): Use NULL as new costs.
4956 (arm_fastmul_tune): Likewise.
4957 (arm_strongarm_tune): Likewise.
4958 (arm_xscale_tune): Likewise.
4959 (arm_9e_tune): Likewise.
4960 (arm_v6t2_tune): Likewise.
4961 (arm_cortex_a5_tune): Likewise.
4962 (arm_cortex_a9_tune): Likewise.
4963 (arm_v6m_tune): Likewise.
4964 (arm_fa726te_tune): Likewise.
4965 (arm_cortex_a15_tune): Use cortex15_extra_costs.
4966 (arm_cortex_tune): Use generict_extra_costs.
4967 (shifter_op_p): New function.
4968 (arm_unspec_cost): Likewise.
4969 (LIBCALL_COST): Define.
4970 (arm_new_rtx_costs): New function.
4971 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
4972 table is available. Use old costs otherwise unless mnew-generic-costs
4974 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
4975 (cpu_cost_table): Declare.
4977 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
4980 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
4981 (*subs_mul_imm_<mode>)
4982 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
4983 (*sub_<shift>_<mode>)
4984 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
4985 Remove k constraint.
4987 2013-10-03 Ian Bolton <ian.bolton@arm.com>
4989 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
4991 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
4993 2013-10-02 Teresa Johnson <tejohnson@google.com>
4995 * predict.c (probably_never_executed): New function.
4996 (probably_never_executed_bb_p): Invoke probably_never_executed.
4997 (probably_never_executed_edge_p): Ditto.
4998 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
4999 Treat profile insanities conservatively.
5001 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5003 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5005 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5007 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5008 use it. Use smaller increase for scratch. Don't increase reject
5009 for early clobber scratch.
5010 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5011 setting eliminated regs except setting fp from hfp.
5012 (lra_eliminate): Check lra_insn_recog_data on NULL.
5014 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5017 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5018 setting -mvsx-timode by default until the underlying problem is fixed.
5019 (RS6000_CPU, power7 defaults): Likewise.
5021 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5023 * config/x-linux (host-linux.o): Remove header dependencies.
5024 Use $(COMPILE) and $(POSTCOMPILE).
5025 * config/t-linux-android (linux-android.o): Ditto.
5027 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5029 * Makefile.in (expmed.o-warn): Remove.
5031 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5033 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5034 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5036 2013-10-02 Teresa Johnson <tejohnson@google.com>
5038 * dojump.c (do_jump_1): Divide probability between
5039 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5041 2013-10-02 Tom Tromey <tromey@redhat.com>
5043 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5045 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5047 * tree-flow.h: Remove some prototypes.
5048 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5049 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5050 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5051 mark_virtual_phi_result_for_renaming): Relocate here.
5052 * tree-into-ssa.h: Add prototypes.
5053 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5054 single_pred_before_succ_order.
5055 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5056 (nonfreeing_call_p) Move to gimple.c.
5057 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5059 * basic-block.h (single_pred_before_succ_order): Add prototype.
5060 * gimple.c (nonfreeing_call_p): Relocate here.
5061 * gimple.h: Add prototype.
5062 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5063 * tree-ssa-dom.h: New file. Relocate prototypes here.
5064 * tree-ssa.h: Include tree-ssa-dom.h.
5066 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5068 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5069 Use $(COMPILE) and $(POSTCOMPILE).
5071 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5073 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5075 * tree-flow.h: Remove some prototypes.
5076 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5077 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5078 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5079 propagate_tree_value*): Move from here to...
5080 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5081 propagate_tree_value*): Relocate here.
5082 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5083 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5084 * gimple-fold.c: Remove gimple-fold.h from include list.
5085 * tree-vrp.c: Remove gimple-fold.h from include list.
5086 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5087 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5088 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5089 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5090 * sese.c: Add tree-ssa-propagate.h to include list.
5092 2013-10-02 Richard Biener <rguenther@suse.de>
5094 * tree-loop-distribution.c: Include tree-vectorizer.h for
5096 (enum partition_kind): Remove PKIND_REDUCTION.
5097 (struct partition_s): Remove has_writes member, add reduction_p member.
5098 (partition_alloc): Adjust.
5099 (partition_builtin_p): Likewise.
5100 (partition_has_writes): Remove.
5101 (partition_reduction_p): New function.
5102 (partition_merge_into): Likewise.
5103 (generate_code_for_partition): Commonize builtin partition
5105 (rdg_cannot_recompute_vertex_p): Remove.
5106 (already_processed_vertex_p): Likewise.
5107 (rdg_flag_vertex): Do not set has_writes.
5108 (classify_partition): Adjust.
5109 (rdg_build_partitions): Do not set has_writes, treat all
5110 partitions as useful.
5111 (distribute_loop): Record number of library calls generated. Adjust.
5112 (tree_loop_distribution): Report number of loops and library
5113 calls generated as opt-info.
5115 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5117 * tree-flow.h: Include new .h files. Move prototypes.
5118 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5119 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5120 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5121 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5122 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5123 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5124 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5125 * tree.h (get_ref_base_and_extent): Move prototype out.
5126 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5128 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5129 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5130 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5131 * tree-scalar-evolution.c: Include tree.h.
5132 * sese.c: Include tree.h.
5133 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5134 * dwarf2out.c: Include tree-dfa.h.
5135 * tree-chrec.c: Include tree.h.
5136 * tree-data-ref.c: Include tree.h.
5138 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5140 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5143 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5145 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5146 Use $(COMPILE) and $(POSTCOMPILE).
5147 (sol2-cxx.o): Likewise.
5148 (sol2-stubs.o): Likewise.
5150 * config/x-solaris (host-solaris.o): Likewise.
5152 * config/sparc/t-sparc (sparc.o): Remove.
5153 (sparc-c.o): Remove header dependencies.
5154 Use $(COMPILE) and $(POSTCOMPILE).
5155 * config/sparc/x-sparc: Likewise.
5157 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5159 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5160 * config/arc/arc700.md: Likewise.
5161 * config/arc/arc-modes.def: Likewise.
5162 * config/arc/arc-simd.h: Likewise.
5163 * config/arc/t-arc-uClibc: Likewise.
5164 * config/arc/t-arc-newlib: Likewise.
5166 2013-10-02 Renlin Li <renlin.li@arm.com>
5168 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5170 (aarch64_expand_epilogue): Likewise.
5172 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5173 Yufeng Zhang <yufeng.zhang@arm.com>
5175 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5177 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5178 'base_in' represent a conversion and legal_cast_p_1 holds; set
5179 'base_in' with the returned value from get_unwidened.
5181 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5183 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5184 plus_constant is not used.
5186 2013-10-01 Wei Mi <wmi@google.com>
5188 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5189 * config/i386/i386.md: Add define_peephole2 to
5190 break partial reg stall for cvtss2sd/cvtsd2ss.
5192 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5194 * config/arc/arc.c (pass_arc_ifcvt::clone):
5195 Update for ctxt_ -> m_ctxt change.
5197 2013-10-01 Jeff Law <law@redhat.com>
5199 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5200 outgoing_edge and intermediate_edge fields. Instead store the path.
5201 (redirection_data::hash): Hash on the last edge's destination index.
5202 (redirection_data::equal): Check the entire thread path.
5203 (lookup_redirectio_data): Corresponding changes.
5204 (create_edge_and_update_destination_phis): Likewise.
5205 (thread_single_edge): Likewise.
5207 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5208 Diego Novillo <dnovillo@google.com>
5210 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5211 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5212 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5213 versions of these patterns.
5215 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
5216 (__builtin_arc_aligned): Likewise.
5218 * config/arc/arc.md: Expand adc_0 comment stating the intended
5219 purpose and why it isn't ready.
5220 Replace commented out call_value_via_label_mixed with a
5221 plain comment about bl_s.
5223 * config/arc/arc.c (stdio.h): Don't include directly.
5224 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
5225 Remove if (1) condition.
5226 (arc_encode_section_info): Fix comment.
5228 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5230 * config/arc/arc.c (arc_conditional_register_usage):
5231 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
5232 Also set reg_alloc_order for DMA config regs.
5234 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5235 Jeremy Bennett <jeremy.bennett@embecosm.com>
5237 * doc/install.texi (--with-cpu): Mention ARC.
5238 (arc-*-elf32): New paragraph.
5239 (arc-linux-uclibc): Likewise.
5240 * doc/md.texi (Machine Constraints): Add ARC part.
5241 * doc/invoke.texi: (menu): Add ARC Options.
5242 (Machine Dependent Options) <ARC Options>: Add synopsis.
5243 (node ARC Options): Add.
5244 * doc/extend.texi (long_call / short_call attribute): Add ARC.
5245 (ARC Built-in Functions): New section defining
5246 generic ARC built-in functions.
5247 (ARC SIMD Built-in Functions): New section defining SIMD specific
5249 (Declaring Attributes of Functions): Extended
5250 description of short_call and long_call attributes for ARC and
5251 added index entries.
5253 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
5254 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
5255 Joern Rennecke <joern.rennecke@embecosm.com>
5256 Muhammad Khurram Riaz <khurram.riaz@arc.com>
5257 Brendan Kehoe <brendan@zen.org>
5258 Michael Eager <eager@eagercon.com>
5259 Simon Cook <simon.cook@embecosm.com>
5260 Jeremy Bennett <jeremy.bennett@embecosm.com>
5262 * config/arc, common/config/arc: New directories.
5264 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5265 Brendan Kehoe <brendan@zen.org>
5266 Simon Cook <simon.cook@embecosm.com>
5268 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
5270 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5272 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
5273 * tree-ssa-coalesce.h: New. Move prototype to here.
5274 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
5275 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
5276 (gimple_can_coalesce_p): Move to...
5277 * gimple.c (gimple_can_coalesce_p): Here.
5279 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5281 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
5282 (dump_decl_set): Move to gimple.c.
5283 * gimple.h: Don't include tree-ssa-operands.h.
5284 (dump_decl_set): Add prototype.
5285 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5286 Move to gimple-ssa.h.
5287 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
5288 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
5289 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
5290 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
5291 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
5293 (dump_decl_set): Relocate here.
5294 * gimple-ssa.h: New file.
5295 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5296 Relocate from gimple.h.
5297 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
5298 * tree-ssa-operands.c (swap_ssa_operands): Rename from
5299 swap_tree_operands and remove non-ssa path.
5300 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
5301 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
5303 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
5304 vect_is_simple_reduction_1): Use swap_ssa_operands.
5305 * tree-flow.h: Move various prototypes to tree-phinodes.h.
5306 (enum need_phi_state): Move to tree-into-ssa.c.
5307 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5308 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
5309 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
5310 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
5311 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5312 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5313 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5314 num_imm_uses): Move to ssa-iterators.h.
5315 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
5316 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
5318 (op_iter_done, op_iter_next_def, op_iter_next_tree,
5319 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
5320 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
5321 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
5322 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
5323 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
5324 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
5325 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
5326 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
5327 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
5328 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
5329 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
5330 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
5331 (set_phi_nodes): Move to tree-phinodes.h.
5332 * tree-ssa-operands.h (enum ssa_op_iter_type,
5333 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
5334 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
5335 (dump_decl_set): Remove prototype.
5336 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
5337 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
5338 (set_phi_nodes): Relocate from tree-flow-inline.h.
5339 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
5341 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
5342 include list. Temporarily add gimple.h to include list.
5343 * ssa-iterators.h: New file.
5344 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5345 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
5346 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
5347 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
5348 Relocate from tree-ssa-operands.h.
5349 (delink_imm_use, link_imm_use_to_list, link_imm_use,
5350 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5351 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5352 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5353 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
5354 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
5355 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
5356 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
5357 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
5358 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
5359 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
5360 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
5361 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
5362 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
5363 Relocate from tree-flow-inline.h.
5364 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
5366 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
5369 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
5370 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
5371 Insert '\t' to output template.
5372 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
5373 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
5374 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
5375 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
5377 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
5379 * doc/install.texi (Host/target specific installation notes for GCC):
5380 Put @anchor before @heading.
5381 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
5382 Use @email for email addresses.
5384 2013-10-01 Jeff Law <law@redhat.com>
5386 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
5387 a vec. Only delete the path if we create one without successfully
5388 registering a jump thread.
5389 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
5391 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
5392 (paths): New vector of jump threading paths.
5393 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
5394 (THREAD_PATH): New accessor macro for the entire thread path.
5395 (lookup_redirection_data): Get intermediate and final outgoing edge
5396 from the thread path.
5397 (create_edge_and_update_destination_phis): Copy the threading path.
5398 (ssa_fix_duplicate_block_edges): Get edges and block types from the
5399 jump threading path.
5400 (ssa_redirect_edges): Get edges and block types from the jump threading
5401 path. Free the path vector.
5402 (thread_block): Get edges from the jump threading path. Look at the
5403 entire path to see if we thread to a loop exit. If we cancel a jump
5404 thread request, then free the path vector.
5405 (thread_single_edge): Get edges and block types from the jump threading
5406 path. Free the path vector.
5407 (thread_through_loop_header): Get edges and block types from the jump
5408 threading path. Free the path vector.
5409 (mark_threaded_blocks): Iterate over the vector of paths and store
5410 the path on the appropriate edge. Get edges and block types from the
5411 jump threading path.
5412 (mark_threaded_blocks): Get edges and block types from the jump
5413 threading path. Free the path vector.
5414 (thread_through_all_blocks): Use the vector of paths rather than
5415 a vector of 3-edge sets.
5416 (register_jump_thread): Accept pointer to a path vector rather
5417 than the path vector itself. Store the path vector for later use.
5420 2013-10-01 Jakub Jelinek <jakub@redhat.com>
5421 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5424 * config/s390/s390.c (s390_split_branches): Modify check for table
5426 (s390_chunkify_start): Rearrange table jump insn check in order to
5427 deal with compare and branch insns correctly.
5429 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
5433 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
5434 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
5435 define_insn_and_split.
5436 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
5437 (shiftsi3_compare): New pattern.
5439 * config/arm/unspecs.md (UNSPEC_RRX): New.
5441 2013-10-01 Alan Modra <amodra@gmail.com>
5443 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
5444 casing inout operands.
5446 2013-10-01 Richard Biener <rguenther@suse.de>
5448 PR tree-optimization/58553
5449 * tree-loop-distribution.c (struct partition_s): Add niter member.
5450 (classify_partition): Populate niter member for the partition
5451 and properly identify whether the relevant store happens before
5452 or after the loop exit.
5453 (generate_memset_builtin): Use niter member from the partition.
5454 (generate_memcpy_builtin): Likewise.
5456 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
5458 * vec.h (vec_prefix, vec): Prefix member names with "m_".
5459 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
5461 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
5463 * basic-block.h (edge_list): Prefix member names with "m_".
5464 * context.h (context): Likewise.
5465 * domwalk.h (dom_walker): Likewise.
5466 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
5467 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
5468 * hash-table.h (hash_table): Likewise.
5469 * machmode.h (bit_field_mode_iterator): Likewise.
5470 * pass_manager.h (pass_list): Likewise.
5471 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
5472 * tree-pass.h (pass_data): Likewise.
5473 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
5474 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
5475 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
5476 * asan.c (pass_data_asan): Update accordingly.
5477 * cfganal.c (control_dependences::find_control_dependence): Likewise.
5478 (control_dependences::control_dependences): Likewise.
5479 (control_dependences::~control_dependences): Likewise.
5480 (control_dependences::~control_dependences): Likewise.
5481 (control_dependences::get_edges_dependent_on): Likewise.
5482 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
5483 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
5484 * context.c (gcc::context::context): Likewise.
5485 * cprop.c (pass_rtl_cprop::clone): Likewise.
5486 * domwalk.c (dom_walker::walk): Likewise.
5487 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
5488 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
5489 * mode-switching.c (pass_mode_switching::clone): Likewise.
5490 * passes.c (opt_pass::opt_pass): Likewise.
5491 (pass_manager::pass_manager): Likewise.
5492 * predict.c (pass_strip_predict_hints::clone): Likewise.
5493 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
5494 (pass_split_all_insns::clone): Likewise.
5495 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
5497 (bit_field_mode_iterator::next_mode): Likewise.
5498 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
5499 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
5500 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
5501 * tree-complex.c (pass_lower_complex::clone): Likewise.
5502 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
5503 * tree-object-size.c (pass_object_sizes::clone): Likewise.
5504 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
5505 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
5506 (pass_fold_builtins::clone): Likewise.
5507 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
5508 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
5509 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
5510 (pass_cd_dce::clone): Likewise.
5511 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
5512 (pass_phi_only_cprop::clone): Likewise.
5513 * tree-ssa-dse.c (pass_dse::clone): Likewise.
5514 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
5515 * tree-ssa-loop.c (pass_lim::clone): Likewise.
5516 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
5517 * tree-ssa-pre.c (pass_fre::clone): Likewise.
5518 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
5519 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
5520 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
5521 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
5522 * tree-vrp.c (pass_vrp::clone): Likewise.
5523 * tsan.c (pass_tsan::clone): Likewise.
5525 2013-09-30 Jakub Jelinek <jakub@redhat.com>
5528 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
5529 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
5532 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
5533 optimization, punt if sign_bit_p looked through any zero extension.
5535 2013-09-30 Teresa Johnson <tejohnson@google.com>
5537 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
5538 Update redirected out edge count in joiner case.
5539 (ssa_redirect_edges): Common the joiner and non-joiner cases
5540 so that joiner case gets profile updates.
5542 2013-09-30 Richard Biener <rguenther@suse.de>
5544 PR tree-optimization/58554
5545 * tree-loop-distribution.c (classify_partition): Require
5546 unconditionally executed stores for memcpy and memset recognition.
5547 (tree_loop_distribution): Calculate dominance info.
5549 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5551 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
5552 (NO_PROFILE_COUNTERS): Likewise.
5553 (PROFILE_HOOK): Likewise.
5554 (FUNCTION_PROFILER): Likewise.
5555 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
5557 2013-09-30 Iain Sandoe <iain@codesourcery.com>
5559 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
5560 calls and defines in TARGET_MACHO conditional.
5561 (load_macho_picbase_di): Likewise.
5562 (reload_macho_picbase): Likewise.
5563 (reload_macho_picbase_si): Likewise.
5564 (reload_macho_picbase_di): Likewise.
5565 (nonlocal_goto_receiver): Likewise.
5567 2013-09-30 Nick Clifton <nickc@redhat.com>
5569 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
5570 that use the MSP430X ISA.
5571 (msp430_option_override): Scan -mmcu command line option for any
5572 MCU name that supports the MSP430X ISA.
5573 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
5574 -mmcu options which enable the MSP430X ISA.
5576 2013-09-30 Richard Biener <rguenther@suse.de>
5579 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
5580 before looking for setjmp-like calls.
5582 2013-09-29 Iain Sandoe <iain@codesourcery.com>
5585 * config/darwin-protos.h (machopic_get_function_picbase): New.
5586 * config/darwin.c (machopic_get_function_picbase): New.
5587 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
5588 label for a new func. (load_macho_picbase_di): Likewise.
5589 (reload_macho_picbase): New expand.
5590 (reload_macho_picbase_si): New insn.
5591 (reload_macho_picbase_di): New insn.
5592 (nonlocal_goto_receiver): New define and split.
5593 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
5594 (unspecv enum): Add UNSPECV_NLGR.
5596 2013-09-29 Iain Sandoe <iain@codesourcery.com>
5598 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
5599 that altivec registers are correctly sized on Darwin.
5601 2013-09-29 Iain Sandoe <iain@codesourcery.com>
5603 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
5604 darwin-driver.o): Use COMPILE and POSTCOMPILE.
5605 * config/x-darwin (host-darwin.o): Likewise.
5606 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
5607 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
5608 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
5610 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
5612 * doc/invoke.texi: Fix usage of @tie{} command.
5614 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
5616 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
5618 2013-09-28 Jan Hubicka <jh@suse.cz>
5620 * config/i386/x86-tune.def: Add documentation for each of the options;
5623 2013-09-28 Jan Hubicka <jh@suse.cz>
5625 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
5627 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
5628 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
5629 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
5630 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
5631 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
5633 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
5635 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
5636 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
5637 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
5638 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
5639 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
5640 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
5641 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
5642 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
5643 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
5644 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
5645 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
5646 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
5647 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
5648 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
5649 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
5650 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
5651 gimple.h, godump.c, graphite-clast-to-gimple.c,
5652 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
5653 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
5654 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
5655 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5656 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
5657 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
5658 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
5659 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
5660 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
5661 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
5662 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
5663 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
5664 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
5665 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
5666 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
5667 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
5668 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
5669 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
5670 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
5671 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
5672 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
5673 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
5674 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
5675 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
5676 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5677 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5678 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
5679 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
5680 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
5681 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5682 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5683 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
5684 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
5685 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5686 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5687 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
5688 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
5689 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
5690 whitespace before "(".
5692 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
5694 * expr.h (extract_bit_field): Remove packedp parameter.
5695 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
5696 from forward declaration.
5697 (store_split_bit_field): Remove packedp arg from calls to
5698 extract_fixed_bit_field.
5699 (extract_bit_field_1): Remove packedp parameter and packedp
5700 argument from recursive calls and calls to extract_fixed_bit_field.
5701 (extract_bit_field): Remove packedp parameter and corresponding
5702 arg to extract_bit_field_1.
5703 (extract_fixed_bit_field): Remove packedp parameter. Remove code
5705 (extract_split_bit_field): Remove packedp arg from call to
5706 extract_fixed_bit_field.
5707 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
5708 (copy_blkmode_from_reg): Likewise.
5709 (copy_blkmode_to_reg): Likewise.
5710 (read_complex_part): Likewise.
5711 (store_field): Likewise.
5712 (expand_expr_real_1): Likewise.
5713 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
5714 to extract_bit_field.
5715 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
5716 call to extract_bit_field.
5717 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
5718 call to extract_bit_field.
5719 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
5720 and special packedp behavior from -fstrict-volatile-bitfields
5723 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5725 * lra-eliminations.c (init_elim_table): Guard value_p.
5727 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5729 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
5730 DFmode, DImode, and SFmode in the upper VSX registers based on the
5731 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
5732 if -mpower8-vector. Combine -mvsx-timode handling with the rest
5733 of the VSX register handling.
5735 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
5737 (zero_extendsidi2_lfiwzx): Add support for loading into the
5738 Altivec registers with -mpower8-vector. Use wu/wv constraints to
5739 only do VSX memory options on Altivec registers.
5740 (extendsidi2_lfiwax): Likewise.
5741 (extendsfdf2_fpr): Likewise.
5742 (mov<mode>_hardfloat, SF/SD modes): Likewise.
5743 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
5744 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
5745 (movdi_internal64): Likewise.
5747 2013-09-27 Xinliang David Li <davidxl@google.com>
5749 * opts.c (finish_options): Adjust parameters
5750 according to vect cost model.
5751 (common_handle_option): Set dynamic vect cost
5753 targhooks.c (default_add_stmt_cost): Compute stmt cost
5755 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
5756 Use helper function.
5757 * tree-vectorizer.h (unlimited_cost_model): New function.
5758 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
5759 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
5761 (vect_enhance_data_refs_alignment): Ditto.
5762 * flag-types.h: New enum.
5763 * common/config/i386/i386-common.c (ix86_option_init_struct):
5764 No need to initialize vect_cost_model flag.
5765 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
5768 2013-09-27 Diego Novillo <dnovillo@google.com>
5770 * gimple.h (enum ssa_mode): Remove.
5772 2013-09-27 Paulo Matos <pmatos@broadcom.com>
5774 * cfgloop.h (number_of_loops): Fix typo in check for null.
5776 2013-09-27 Jakub Jelinek <jakub@redhat.com>
5779 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
5780 are children of outermost saved_cfun's loop, and set it up to
5781 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
5782 if loop != loop0 and SESE region contains bbs that belong to loop0.
5784 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
5786 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
5787 (binary_scale_code_p, get_base_term, get_index_term): New functions.
5788 (set_address_segment, set_address_base, set_address_index)
5789 (set_address_disp): Accept the argument unconditionally.
5790 (baseness): Remove must_be_base_p and must_be_index_p checks.
5791 (decompose_normal_address): Classify as much as possible in the
5794 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
5796 * cse.c (count_reg_usage): Handle INT_LIST.
5797 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
5798 * reginfo.c (reg_scan_mark_refs): Likewise.
5799 * reload1.c (eliminate_regs_1): Likewise.
5801 2013-09-27 Iain Sandoe <iain@codesourcery.com>
5804 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
5807 2013-09-27 Richard Biener <rguenther@suse.de>
5809 PR tree-optimization/58459
5810 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
5811 restriction not propagating into loops.
5813 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
5815 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
5816 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
5817 * doc/tree-ssa.texi (Walking use-def chains): Delete.
5819 2013-09-26 Richard Biener <rguenther@suse.de>
5821 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
5823 2013-09-26 Richard Biener <rguenther@suse.de>
5825 * alias.h (component_uses_parent_alias_set): Rename to ...
5826 (component_uses_parent_alias_set_from): ... this.
5827 * alias.c (component_uses_parent_alias_set): Rename to ...
5828 (component_uses_parent_alias_set_from): ... this and return
5830 (reference_alias_ptr_type_1): Use the result from
5831 component_uses_parent_alias_set_from instead of stripping
5832 components one at a time.
5833 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
5835 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
5837 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
5838 Move prototypes to...
5839 * tree-ssa-ter.h: New File. Move prototypes here.
5840 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
5841 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
5842 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
5843 from is_replaceable_p.
5844 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
5845 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
5846 (process_replaceable): Use ter_is_replaceable_p.
5847 (find_replaceable_in_bb): Use ter_is_replaceable_p.
5848 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
5849 newly refactored ssa_is_replaceable_p.
5850 * cfgexpand.c: Include tree-outof-ssa.h.
5851 * ssaexpand.h: Delete.
5853 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
5855 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
5856 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
5858 (create_gimple_tmp): Delete.
5859 (get_expr_type, build_assign, build_type_cast): Move to...
5860 * gimple-builder.c: New File.
5861 (get_expr_type): Relocate from gimple.c.
5862 (build_assign, build_type_cast): Change to only create ssanames.
5863 * gimple.h: Move prototypes to...
5864 * gimple-builder.h: New File. Here.
5865 * tree-ssa.h: And here.
5866 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
5867 count_uses_and_derefs): Relocate from gimple.c.
5868 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
5869 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
5870 * tree-ssa-math-opts (execute_cse_reciprocals): Use
5871 gimple_replace_ssa_lhs.
5872 * asan.c: Include gimple-builder.h.
5873 * Makefile.in: Add gimple-builder.o.
5875 2013-09-26 Richard Biener <rguenther@suse.de>
5877 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
5878 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
5879 (loe_visit_block): Use gcc_checking_assert.
5880 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
5881 gimple_assign_ssa_name_copy_p.
5882 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
5885 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
5887 * config/rs6000/t-rs6000 (rs6000.o): Remove.
5888 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
5890 2013-09-26 Richard Biener <rguenther@suse.de>
5892 PR tree-optimization/58539
5893 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
5894 the fact that debug statements are not taking part in loop-closed
5897 2013-09-26 Nick Clifton <nickc@redhat.com>
5899 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
5900 time warning message.
5901 (msp430_print_operand_raw): Delete unused letter parameter.
5902 (TARGET_PRINT_OPERAND_ADDRESS): Define.
5903 (msp430_print_operand_address): New function.
5904 (msp430_print_operand): Move address printing code from here to
5906 * config/msp430/msp430.md (movsipsi2): Add comment in generated
5908 (zero_extendpsisi2): Likewise.
5909 (extendpsisi2): New pattern.
5910 (andneghi3): New pattern.
5912 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
5914 * config/aarch64/aarch64.opt (mlra): New option.
5915 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
5916 (TARGET_LRA_P): Define.
5918 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
5920 * expr.c (expand_assignment): Remove obsolete comment.
5922 2013-09-25 Jeff Law <law@redhat.com>
5924 * tree-flow.h (thread_through_all_blocks): Prototype moved into
5925 tree-ssa-threadupdate.h.
5926 (register_jump_thread): Similarly.
5927 * tree-ssa-threadupdate.h: New header file.
5928 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
5929 * tree-vrp.c: Likewise.
5930 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
5931 (thread_around_empty_blocks): Change type of path vector argument to
5932 an edge,type pair from just an edge. Initialize both elements when
5933 appending to a jump threading path. Tweak references to elements
5935 (thread_across_edge): Similarly. Release memory for the elements
5937 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
5938 (dump_jump_thread_path): New function broken out from
5939 register_jump_thread.
5940 (register_jump_thread): Use dump_jump_thread_path. Change type of
5941 path vector entries. Search the path for NULL edges and dump
5942 the path if one is found. Tweak the conversion of path to 3-edge
5943 form to use the block copy type information embedded in the path.
5945 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
5947 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
5949 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
5950 Vladimir Makarov <vmakarov@redhat.com>
5952 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
5953 from the least significant bit.
5954 (strip_address_mutations): Add bitfield operations handling.
5955 (must_be_index_p): Add shifting and rotate operations handling.
5956 (set_address_base): Use must_be_base_p predicate.
5957 (set_address_index): Use must_be_index_p predicate.
5959 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
5960 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5961 Sergey Lega <sergey.s.lega@intel.com>
5962 Anna Tikhonova <anna.tikhonova@intel.com>
5963 Ilya Tocar <ilya.tocar@intel.com>
5964 Andrey Turetskiy <andrey.turetskiy@intel.com>
5965 Ilya Verbin <ilya.verbin@intel.com>
5966 Kirill Yukhin <kirill.yukhin@intel.com>
5967 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5969 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
5971 (ix86_expand_vector_move_misalign): Support new unaligned load and
5972 stores and use new names.
5973 (CODE_FOR_sse2_storedqu): Rename to ...
5974 (CODE_FOR_sse2_storedquv16qi): ... this.
5975 (CODE_FOR_sse2_loaddqu): Rename to ...
5976 (CODE_FOR_sse2_loaddquv16qi): ... this.
5977 (CODE_FOR_avx_loaddqu256): Rename to ...
5978 (CODE_FOR_avx_loaddquv32qi): ... this.
5979 (CODE_FOR_avx_storedqu256): Rename to ...
5980 (CODE_FOR_avx_storedquv32qi): ... this.
5981 * config/i386/i386.md (fpint_logic): New.
5982 * config/i386/sse.md (VMOVE): Extend for AVX512.
5986 (VI_UNALIGNED_LOADSTORE): Ditto.
5987 (sse2_avx_avx512f): Ditto.
5988 (sse2_avx2): Extend for AVX512.
5989 (sse4_1_avx2): Ditto.
5990 (avx2_avx512f): New.
5991 (sse): Extend for AVX512.
5994 (avxsizesuffix): Ditto.
5995 (sseintvecmode): Ditto.
5997 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
5998 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
5999 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6000 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6001 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6002 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6003 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6004 (STORENT_MODE): Extend for AVX512.
6005 (*absneg<mode>2): Replace constraint "x" with "v".
6006 (*mul<mode>3): Ditto.
6007 (*ieee_smin<mode>3): Ditto.
6008 (*ieee_smax<mode>3): Ditto.
6009 (avx_cmp<mode>3): Replace VF with VF_128_256.
6010 (*<sse>_maskcmp<mode>3_comm): Ditto.
6011 (<sse>_maskcmp<mode>3): Ditto.
6012 (<sse>_andnot<mode>3): Extend for AVX512.
6013 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6014 (<code><mode>3, fpint_logic): New.
6015 (*<code><mode>3): Extend for AVX512.
6016 (avx512flogicsuff): New.
6017 (avx512f_<logic><mode>): Ditto.
6018 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6020 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6021 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6022 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6023 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6024 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6025 (xop_vpermil2<mode>3): Ditto.
6026 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6027 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6028 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6029 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6031 2013-09-25 Tom Tromey <tromey@redhat.com>
6033 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6034 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6035 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6036 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6037 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6038 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6039 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6040 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6041 (GRAPHITE_HTAB_H): Remove.
6043 2013-09-25 Tom Tromey <tromey@redhat.com>
6045 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6047 2013-09-25 Tom Tromey <tromey@redhat.com>
6049 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6051 2013-09-25 Tom Tromey <tromey@redhat.com>
6053 * config/i386/t-i386 (i386.o): Remove.
6054 (i386-c.o): Use COMPILE and POSTCOMPILE.
6056 2013-09-25 Tom Tromey <tromey@redhat.com>
6058 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6060 2013-09-25 Tom Tromey <tromey@redhat.com>
6062 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6063 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6064 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6065 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6066 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6067 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6068 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6069 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6070 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6071 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6072 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6073 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6074 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6075 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6076 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6077 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6078 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6079 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6080 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6081 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6082 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6083 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6084 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6085 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6086 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6087 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6088 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6089 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6090 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6091 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6092 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6093 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6094 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6095 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6096 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6097 (graphite-clast-to-gimple.o, graphite-dependences.o)
6098 (graphite-interchange.o, graphite-poly.o)
6099 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6100 (graphite-optimize-isl.o, tree-vect-loop.o)
6101 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6102 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6103 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6104 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6105 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6106 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6107 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6108 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6109 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6110 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6111 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6112 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6113 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6114 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6115 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6116 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6117 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6118 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6119 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6120 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6121 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6122 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6123 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6124 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6125 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6126 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6127 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6128 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6129 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6130 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6131 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6132 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6133 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6134 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6135 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6136 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6137 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6138 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6139 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6140 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6141 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6142 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6143 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6144 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6145 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6146 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6147 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6148 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6149 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6150 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6151 (gcov-dump.o): Remove.
6152 (default-c.o): Use COMPILE and POSTCOMPILE.
6153 (CFLAGS-gcc.o): New variable.
6154 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6156 2013-09-25 Tom Tromey <tromey@redhat.com>
6158 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6159 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6160 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6161 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6162 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6163 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6164 (c-family/array-notation-common.o, c-family/stub-objc.o)
6165 (c-family/c-ubsan.o): Remove.
6167 2013-09-25 Tom Tromey <tromey@redhat.com>
6169 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6171 2013-09-25 Tom Tromey <tromey@redhat.com>
6173 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6174 to add -DENABLE_SHARED_LIBGCC.
6175 (gcc.o): Don't use subshell.
6177 2013-09-25 Tom Tromey <tromey@redhat.com>
6179 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6180 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6181 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6182 * configure, config.in: Rebuild.
6184 2013-09-25 Tom Tromey <tromey@redhat.com>
6186 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6187 (COMPILE, POSTCOMPILE): New variables.
6188 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6189 (DEPFILES): New variable.
6190 Include ".Po" files.
6191 * configure.ac: Add checks for dependency checking.
6192 * configure, aclocal.m4: Regenerate.
6194 2013-09-25 Tom Tromey <tromey@redhat.com>
6196 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6197 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6199 2013-09-25 Tom Tromey <tromey@redhat.com>
6201 * Makefile.in (generated_files): Add options.h,
6202 target-hooks-def.h, insn-opinit.h,
6203 common/common-target-hooks-def.h, pass-instances.def,
6204 c-family/c-target-hooks-def.h.
6206 2013-09-25 Jeff Law <law@redhat.com>
6208 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6209 than foo[foo.length () - 1] to access last member in a vec.
6210 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6212 2013-09-25 Richard Biener <rguenther@suse.de>
6215 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
6217 2013-09-25 Jan Hubicka <jh@suse.cz>
6219 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
6222 2013-09-25 Marek Polacek <polacek@redhat.com>
6225 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
6226 when determining the type name.
6228 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
6230 * config/sh/sh.md: Fix formatting.
6232 2013-09-24 Xinliang David Li <davidxl@google.com>
6234 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
6235 max peel iterations parameter.
6236 * param.def: New parameter.
6237 * doc/invoke.texi: Document New parameter.
6239 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
6241 * gimple-pretty-print.c: Various whitespace tweaks.
6242 * tree-core.h: Likewise.
6243 * tree-pretty-print.c: Likewise.
6244 * tree-ssa-alias.c: Likewise.
6245 * tree-ssa-copy.c: Likewise.
6246 * tree-ssanames.c: Likewise.
6247 * tree-ssanames.h: Likewise.
6248 * tree-vrp.c: Likewise.
6250 2013-09-24 Alan Modra <amodra@gmail.com>
6254 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
6255 for output operands that disallow regs. Don't use EXPAND_WRITE on
6258 2013-09-24 Richard Biener <rguenther@suse.de>
6261 * tree.c (reference_alias_ptr_type): Move ...
6262 * alias.c (reference_alias_ptr_type): ... here and implement
6263 in terms of the new reference_alias_ptr_type_1.
6264 (ref_all_alias_ptr_type_p): New helper.
6265 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
6266 use ref_all_alias_ptr_type_p.
6267 (get_deref_alias_set): Add flag_strict_aliasing check here.
6268 (reference_alias_ptr_type_1): New function, split out from ...
6269 (get_alias_set): ... here.
6270 (alias_ptr_types_compatible_p): New function.
6271 * alias.h (reference_alias_ptr_type): Declare.
6272 (alias_ptr_types_compatible_p): Likewise.
6273 * tree.h (reference_alias_ptr_type): Remove.
6274 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
6275 to compare MEM_REF alias types.
6277 2013-09-24 Richard Biener <rguenther@suse.de>
6279 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
6281 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6283 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
6284 reload helper function arrays into a single array reg_addr.
6285 (reload_fpr_gpr): Likewise.
6286 (reload_gpr_vsx): Likewise.
6287 (reload_vsx_gpr): Likewise.
6288 (struct rs6000_reg_addr): Likewise.
6289 (reg_addr): Likewise.
6290 (rs6000_debug_reg_global): Change rs6000_vector_reload,
6291 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
6292 (rs6000_init_hard_regno_mode_ok): Likewise.
6293 (rs6000_secondary_reload_direct_move): Likewise.
6294 (rs6000_secondary_reload): Likewise.
6296 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
6297 constraints: wu, ww, and wy. Repurpose wv constraint added during
6298 power8 changes. Put wg constraint in alphabetical order.
6300 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
6301 for future work to add ISA 2.07 VSX single precision support.
6302 (-mvsx-scalar-double): Change default from -1 to 1, update
6303 documentation comment.
6304 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
6305 (-mupper-regs-df): New debug switch to control whether DF values
6306 can go in the traditional Altivec registers.
6307 (-mupper-regs-sf): New debug switch to control whether SF values
6308 can go in the traditional Altivec registers.
6310 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
6312 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
6313 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
6314 Add new constraints, wu/ww/wy. Repurpose wv constraint.
6315 (rs6000_debug_legitimate_address_p): Print if we are running
6316 before, during, or after reload.
6317 (rs6000_secondary_reload): Add a comment.
6318 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
6320 * config/rs6000/constraints.md (wa constraint): Sort w<x>
6321 constraints. Update documentation string.
6322 (wd constraint): Likewise.
6323 (wf constraint): Likewise.
6324 (wg constraint): Likewise.
6325 (wn constraint): Likewise.
6326 (ws constraint): Likewise.
6327 (wt constraint): Likewise.
6328 (wx constraint): Likewise.
6329 (wz constraint): Likewise.
6330 (wu constraint): New constraint for ISA 2.07 SFmode scalar
6332 (ww constraint): Likewise.
6333 (wy constraint): Likewise.
6334 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
6335 in the previous submissions.
6336 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
6338 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
6340 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
6341 (REG_BR_PROB): Say that the probability is stored in an int_list.
6342 * reg-notes.def: Update commentary to mention INT_LIST.
6343 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
6344 (INT_LIST): New rtx.
6345 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
6346 * rtlanal.c (int_reg_note_p): New function.
6347 (alloc_reg_note): Assert that the note does not have an int argument.
6348 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
6349 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
6350 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
6351 rather than an INSN_LIST. Handle INT_LIST.
6352 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
6353 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
6354 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
6355 Manipulate them as ints rather than rtxes.
6356 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
6357 * regmove.c (copy_src_to_dest): Likewise.
6358 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
6360 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
6361 into the cases that need it.
6362 * config/arm/arm.c (arm_unwind_emit): Likewise.
6364 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
6365 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
6366 * loop-doloop.c (add_test, doloop_modify): Likewise.
6367 * loop-unswitch.c (compare_and_jump_seq): Likewise.
6368 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
6369 * predict.c (combine_predictions_for_insn): Likewise.
6370 * print-rtl.c (print_rtx): Handle INT_LIST.
6371 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
6372 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
6373 * config/arm/arm.c (emit_unlikely_jump): Likewise.
6374 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
6375 (ix86_split_fp_branch, predict_jump): Likewise.
6376 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
6377 * config/sh/sh.c (expand_cbranchsi4): Likewise.
6378 * config/spu/spu.c (ea_load_store_inline): Likewise.
6380 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
6381 value of a REG_BR_PROB note.
6382 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
6383 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
6384 * emit-rtl.c (try_split): Likewise.
6385 * predict.c (br_prob_note_reliable_p): Likewise.
6386 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
6387 * reorg.c (mostly_true_jump): Likewise.
6388 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
6389 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
6390 * config/i386/i386.c (ix86_print_operand): Likewise.
6391 * config/ia64/ia64.c (ia64_print_operand): Likewise.
6392 * config/mmix/mmix.c (mmix_print_operand): Likewise.
6393 * config/rs6000/rs6000.c (output_cbranch): Likewise.
6394 * config/s390/s390.c (s390_expand_tbegin): Likewise.
6395 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
6396 * config/sparc/sparc.c (output_cbranch): Likewise.
6397 * config/spu/spu.c (get_branch_target): Likewise.
6398 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
6399 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
6401 2013-09-23 Jan Hubicka <jh@suse.cz>
6403 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
6405 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
6406 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
6408 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
6409 (gimple_fold_call): Dump inconsistent devirtualizations; add
6410 sanity check for type based devirtualizations.
6411 * ipa-prop.c: Include ipa-utils.h
6412 (ipa_intraprocedural_devirtualization): Add sanity check.
6413 (try_make_edge_direct_virtual_call): Likewise.
6415 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
6417 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
6418 assignment statements.
6420 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
6422 * gimple-pretty-print.c (dump_ssaname_info): New function.
6423 (dump_gimple_phi): Call it.
6424 (pp_gimple_stmt_1): Likewise.
6425 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
6426 (range_info_def): Declare.
6427 * tree-pretty-print.c (pp_double_int): New function.
6428 (dump_generic_node): Call it.
6429 * tree-pretty-print.h (pp_double_int): Declare.
6430 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
6431 * tree-ssanames.h (range_info_def): New structure.
6432 (value_range_type): Move definition here.
6433 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
6435 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
6437 (set_range_info): New function.
6438 (get_range_info): Likewise.
6439 (duplicate_ssa_name_range_info): Likewise.
6440 (duplicate_ssa_name_fn): Check pointer type and call
6441 duplicate_ssa_name_range_info.
6442 * tree-ssa-copy.c (fini_copy_prop): Likewise.
6443 * tree-vrp.c (value_range_type): Remove definition, now in
6445 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
6446 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
6447 (SSA_NAME_RANGE_INFO): New macro.
6449 2013-09-23 Richard Biener <rguenther@suse.de>
6451 PR tree-optimization/58464
6452 * tree-ssa-pre.c (phi_trans_lookup): Remove.
6453 (phi_trans_add): Change to add conditionally on being not present.
6454 (phi_translate_1): Remove recursion detection here.
6455 (phi_translate): Pre-seed the cache with NULL to catch
6456 recursion here in a more generic way.
6457 (bitmap_find_leader): Adjust comment.
6458 (get_representative_for): Dump value-numbers.
6459 (create_expression_by_pieces): Likewise.
6460 (insert_into_preds_of_block): Likewise.
6462 2013-09-23 Christian Bruel <christian.bruel@st.com>
6465 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
6466 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
6468 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
6471 2013-09-20 Renlin Li <renlin.li@arm.com>
6473 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6475 (aarch64_expand_epilogue): Likewise.
6476 (aarch64_legitimize_reload_address): Likewise.
6478 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
6480 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
6481 NULL_TREE before pushing them onto the vector. Likewise for labels.
6483 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
6485 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
6486 comparison operators when -fno-trapping-math is in effect.
6487 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
6488 comparison operators in TFmode and assert that unsupported operators
6490 (ia64_print_operand): Likewise.
6492 2013-09-21 Jan Hubicka <jh@suse.cz>
6494 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
6495 (sse_typeless_stores): Enable for core
6496 (sse_load0_by_pxor): Likewise.
6497 (four_jump_limit): Disable for core.
6498 (pad_returns): Likewise.
6499 (avoid_vector_decode): Likewise.
6500 (fuse_cmp_and_branch): Enable for cores.
6501 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
6503 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
6506 * config/pa/pa.c (pa_option_override): Disable auto increment and
6507 decrement instructions until reload is completed.
6509 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
6510 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
6511 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
6513 2013-09-20 DJ Delorie <dj@redhat.com>
6514 Nick Clifton <nickc@redhat.com>
6516 * config/rl78/rl78.c: Various whitespace and comment tweaks.
6517 (need_to_save): Save bank 0 on interrupts.
6518 (characterize_address): Strip far address wrappers.
6519 (rl78_as_legitimate_address): Likewise.
6520 (transcode_memory_rtx): Likewise.
6521 (rl78_peep_movhi_p): Disable this peephole after devirt.
6522 (rl78_propogate_register_origins): Forget all origins when a
6524 * config/rl78/rl78-virt.md: Various whitespace tweaks.
6525 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
6527 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
6528 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
6529 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
6531 * config/rl78/constraints.md: Various whitespace and paren tweaks.
6533 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
6535 * config/pa/pa.md: In "scc" insn patterns, change output template to
6536 handle const0_rtx in reg_or_0_operand operands.
6538 2013-09-20 Martin Husemann <martin@NetBSD.org>
6541 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
6542 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
6544 2013-09-20 Richard Biener <rguenther@suse.de>
6547 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
6548 remembering SSA name version and block index.
6549 (new_scev_info_str): Adjust.
6550 (hash_scev_info): Likewise. Also hash the block index.
6551 (eq_scev_info): Adjust.
6552 (find_var_scev_info): Likewise.
6553 (struct instantiate_cache_entry): Remove.
6554 (struct instantiate_cache_type): Use a htab to map name, block
6556 (instantiate_cache_type::~instantiate_cache_type): Adjust.
6557 (get_instantiated_value_entry): Likewise.
6558 (hash_idx_scev_info, eq_idx_scev_info): New functions.
6559 (instantiate_scev_name): Adjust.
6561 2013-09-20 Jeff Law <law@redhat.com>
6563 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
6565 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
6567 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
6568 Call aarch64_simd_expand_args to update op[argc].
6570 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
6572 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
6575 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
6577 * gengtype.c (file_rules): Added rule for *.cc files.
6578 (get_output_file_with_visibility): Give fatal message when no
6581 2013-09-20 Renlin Li <renlin.li@arm.com>
6583 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
6584 (aarch64_expand_epilogue): Likewise.
6585 (aarch64_legitimize_reload_address): Likewise.
6587 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
6590 * expr.c (expand_assignment): Remove misalignp code path.
6592 2013-09-20 Marek Polacek <polacek@redhat.com>
6595 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
6596 TYPE_PRECISION. Add asserts.
6598 2013-09-20 Richard Biener <rguenther@suse.de>
6600 PR tree-optimization/58453
6601 * tree-loop-distribution.c (distribute_loop): Apply the cost
6602 model for -ftree-loop-distribute-patterns, too.
6604 2013-09-20 Richard Biener <rguenther@suse.de>
6607 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
6608 make type comparison less strict.
6610 2013-09-20 Alan Modra <amodra@gmail.com>
6612 * configure: Regenerate.
6613 * aclocal.m4: Regenerate.
6615 2013-09-20 Marek Polacek <polacek@redhat.com>
6618 * doc/extend.texi: Document that attribute used is meant to be used
6619 on variables with static storage duration.
6621 2013-09-19 Jakub Jelinek <jakub@redhat.com>
6623 PR tree-optimization/58472
6624 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
6625 simd_lane_access set inv_p = false.
6626 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
6627 the simduid magic VAR_DECL.
6629 2013-09-19 Jan Hubicka <jh@suse.cz>
6631 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
6633 2013-09-17 Jeff Law <law@redhat.com>
6635 * tree-ssa-dom.c (record_temporary_equivalences): New function
6636 split out of dom_opt_dom_walker::after_dom_children.
6637 (dom_opt_dom_walker::thread_across_edge): Move common code
6638 in here from dom_opt_dom_walker::after_dom_children.
6639 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
6641 2013-09-19 Jan Hubicka <jh@suse.cz>
6643 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
6644 (TARGET_GENERIC): Use PROCESOR_GENERIC
6645 (enum processor_type): Unify generic32 and 64.
6646 * i386.md (cpu): Likewise.
6647 * x86-tune.def (use_leave): Enable for generic32.
6648 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
6649 * athlon.md: Change generic64 to generic in all occurences.
6650 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
6651 (ix86_target_macros_internal): Likewise.
6652 * driver-i386.c (host_detect_local_cpu): Likewise.
6653 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
6655 (generic_memcpy, generic_memset, generic_cost): This one.
6656 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
6657 (m_GENERIC32, m_GENERIC64): Remove.
6658 (m_GENERIC): Turn into one flag.
6659 (processor_target): Unify generic tunnings.
6660 (ix86_option_override_internal): Replace generic32/64 by generic.
6661 (ix86_issue_rate): Likewise.
6662 (ix86_adjust_cost): Likewise.
6664 2013-09-19 Jan Hubicka <jh@suse.cz>
6666 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
6667 of speculative flag.
6669 2013-09-19 Jakub Jelinek <jakub@redhat.com>
6671 * omp-low.c (expand_omp_sections): Always pass len - 1 to
6672 GOMP_sections_start, even if !exit_reachable.
6674 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
6676 * lra-constraints.c (need_for_all_save_p): Use macro
6677 HARD_REGNO_CALL_PART_CLOBBERED.
6678 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
6679 set up pseudo conflict hard regs.
6681 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6684 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
6687 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
6689 PR rtl-optimization/58438
6690 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
6691 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
6692 from previous subpasses.
6694 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
6696 * arm.c (arm_get_frame_offsets): Validate architecture supports
6697 LDRD/STRD before accepting the tuning preference.
6698 (arm_expand_prologue): Likewise.
6699 (arm_expand_epilogue): Likewise.
6701 2013-09-18 Richard Biener <rguenther@suse.de>
6703 PR tree-optimization/58417
6704 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
6705 have chrecs with symbols defined in the loop as operands.
6706 (chrec_fold_multiply): Likewise.
6707 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
6708 parameters before folding binary operations.
6709 (struct instantiate_cache_entry_hasher): Remove.
6710 (struct instantiate_cache_type): Use a pointer-map.
6711 (instantiate_cache_type::instantiate_cache_type): New function.
6712 (instantiate_cache_type::get): Likewise.
6713 (instantiate_cache_type::set): Likewise.
6714 (instantiate_cache_type::~instantiate_cache_type): Adjust.
6715 (get_instantiated_value_entry): Likewise.
6716 (global_cache): New global.
6717 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
6718 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
6719 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
6720 (instantiate_scev_name): Adjust.
6721 (instantiate_scev): Construct global instead of local cache.
6722 (resolve_mixers): Likewise.
6724 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
6725 Paolo Carlini <paolo.carlini@oracle.com>
6728 * doc/implement-cxx.texi: Fix references to the C++ standards.
6730 2013-09-18 Jakub Jelinek <jakub@redhat.com>
6732 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
6733 * tree-vect-data-refs.c (vect_analyze_data_refs): For
6734 simd_lane_access drs, update also DR_ALIGNED_TO.
6736 2013-09-18 Marek Polacek <polacek@redhat.com>
6739 * doc/extend.texi: Document no_sanitize_undefined attribute.
6740 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
6741 no_sanitize_undefined attribute.
6743 2013-09-18 Nick Clifton <nickc@redhat.com>
6745 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
6746 (ASM_DECLARE_FUNCTION_NAME): Define.
6748 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
6750 * compare-elim.c (find_comparison_dom_walker): New class
6751 (find_comparisons_in_bb): Rename to
6752 find_comparison_dom_walker::before_dom_children
6753 (find_comparisons): Adjust
6754 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
6756 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
6757 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
6758 (init_walk_dominator_tree): Remove declaration.
6759 (fini_walk_dominator_tree): Remove declaration.
6760 * fwprop.c (single_def_use_dom_walker): New class
6761 (single_def_use_enter_block): Convert to
6762 single_def_use_dom_walker::before_dom_children.
6763 (single_def_use_leave_block): Convert to
6764 single_def_use_dom_walker::after_dom_children.
6765 (build_single_def_use_links): Adjust.
6766 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
6768 (find_candidates_in_block): Convert to
6769 find_candidates_dom_walker::before_dom_children.
6770 (execute_strength_reduction): Adjust.
6771 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
6772 (sese_dom_walker): New class.
6773 (sese_dom_walker::sese_dom_walker): New constructor.
6774 (sese_dom_walker::~sese_dom_walker): New destructor.
6775 (build_sese_conditions_before): Convert to
6776 sese_dom_walker::before_dom_children.
6777 (build_sese_conditions_after): Convert to
6778 sese_dom_walker::after_dom_children.
6779 (build_poly_scop): Adjust
6780 * tree-into-ssa.c (rewrite_dom_walker): New class
6781 (rewrite_enter_block): Convert to
6782 rewrite_dom_walker::before_dom_children.
6783 (rewrite_leave_block): Convert to
6784 rewrite_dom_walker::after_dom_children.
6785 (rewrite_update_dom_walker): New class.
6786 (rewrite_update_enter_block): Convert to
6787 rewrite_update_dom_walker::before_dom_children.
6788 (rewrite_update_leave_block): Convert to
6789 rewrite_update_dom_walker::after_dom_children.
6790 (rewrite_blocks, rewrite_into_ssa): Adjust.
6791 (mark_def_dom_walker): New class.
6792 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
6793 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
6794 (mark_def_sites_blocks): Convert to
6795 mark_def_dom_walker::before_dom_children.
6796 (mark_def_site_blocks): Remove.
6797 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
6798 (tree_ssa_dominator_optimize): Adjust.
6799 (dom_thread_across_edge): Convert to method
6800 dom_opt_dom_walker::thread_across_edge.
6801 (dom_opt_enter_block): Convert to member function
6802 dom_opt_dom_walker::before_dom_children.
6803 (dom_opt_leave_block): Convert to member function
6804 dom_opt_dom_walker::after_dom_children.
6805 * tree-ssa-dse.c (dse_dom_walker): New class.
6806 (dse_enter_block): Convert to member function
6807 dse_dom_walker::before_dom_children.
6808 (tree_ssa_dse): Adjust.
6809 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
6810 (determine_invariantness_stmt): Convert to method
6811 invariantness_dom_walker::before_dom_children.
6812 (determine_invariantness): Remove
6813 (move_computations_dom_walker): New class.
6814 (move_computations_stmt): Convert to method
6815 move_computations_dom_walker::before_dom_children.
6816 (move_computations, tree_ssa_lim): Adjust.
6817 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
6818 (nt_init_block): Convert to method
6819 notrappping_dom_walker::before_dom_children.
6820 (nt_fini_block): Convert to method
6821 method nontrapping_dom_walker::after_dom_children.
6822 (get_non_trapping): Adjust.
6823 * tree-ssa-pre.c (eliminate_dom_walker): New class.
6824 (eliminate_bb): Convert to method
6825 eliminate_dom_walker::before_dom_children.
6826 (eliminate_leave_block): Convert to method
6827 eliminate_dom_walker::after_dom_children.
6828 (eliminate): Adjust.
6829 * tree-ssa-strlen.c (strlen_dom_walker): New class.
6830 (strlen_enter_block): Convert to method
6831 strlen_dom_walker::before_dom_children.
6832 (strlen_leave_block): Convert to method
6833 method strlen_dom_walker::after_dom_children.
6834 (tree_ssa_strlen): Adjust.
6835 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
6836 (tree_ssa_uncprop): Adjust.
6837 (uncprop_leave_block): Convert to method
6838 uncprop_dom_walker::after_dom_children.
6839 (uncprop_leave_block): Convert to method
6840 uncprop_dom_walker::before_dom_children.
6842 2013-09-18 Bin Cheng <bin.cheng@arm.com>
6844 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
6845 branch in same basic block. Check both src and dest of the move insn.
6847 2013-09-17 Nick Clifton <nickc@redhat.com>
6849 * config/rl78/rl78-real.md (bf): New pattern.
6851 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
6852 (rl78_print_operand): Do not put a # before a %B.
6853 * config/rl78/rl78.opt: Tweak doc strings.
6855 2013-09-17 DJ Delorie <dj@redhat.com>
6857 * config/rl78/constraints.md (Wcv): Allow up to $r31.
6858 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
6859 (rl78_option_override): Likewise, if -mallregs.
6860 (is_virtual_register): Likewise.
6861 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
6862 (REGNO_OK_FOR_BASE_P): Likewise.
6863 * config/rl78/rl78.opt (-mallregs): New.
6865 2013-09-17 Nick Clifton <nickc@redhat.com>
6867 * config/rl78/rl78.c (need_to_save): Change return type to bool.
6868 For interrupt functions: save all call clobbered registers if the
6869 interrupt handler is not a leaf function.
6870 (rl78_expand_prologue): Always recompute the frame information.
6871 For interrupt functions: only select bank 0 if one of the bank 0
6872 registers is going to be psuhed.
6874 2013-09-17 DJ Delorie <dj@redhat.com>
6876 * config/rl78/constraints.md: For each W* constraint, rename to C*
6877 and create a W* constraint that checks for an optional ES: prefix
6879 * config/rl78/rl78.md (UNS_ES_ADDR): New.
6880 (es_addr): New. Used to wrap far addresses.
6881 * config/rl78/rl78-protos.h (rl78_es_addr): New.
6882 (rl78_es_base): New.
6883 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
6884 wrapped far addresses.
6885 (rl78_print_operand_1): Unwrap far addresses before processing.
6886 (rl78_lo16): Wrap far addresses in unspecs.
6887 (rl78_es_addr): New.
6888 (rl78_es_base): New.
6889 (insn_ok_now): Check for not-yet-wrapped far addresses.
6890 (transcode_memory_rtx): Properly re-wrap far addresses.
6892 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
6894 * config/sparc/t-rtems: Add leon3 multilibs.
6896 2013-09-17 Cong Hou <congh@google.com>
6898 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
6899 when checking the dot production pattern. The type of rhs operand
6900 of multiply is now checked correctly.
6902 2013-09-17 Jeff Law <law@redhat.com>
6904 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6905 edge implied equivalences into successor phis.
6906 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
6907 here from tree-ssa-threadedge.c.
6908 (mark_threaded_blocks): When threading through a joiner, if both
6909 successors of the joiner's clone reach the same block, verify the
6910 PHI arguments are equal. If not, cancel the jump threading request.
6911 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
6912 tree-ssa-threadupdate.c
6913 (thread_across_edge): Don't check PHI argument equality when
6914 threading through joiner block here.
6916 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
6918 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
6919 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
6920 (warn_uninit, warn_uninitialized_vars,
6921 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
6922 Move to tree-ssa-uninit.c.
6923 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
6924 (has_undefined_value_p): New. Pass dependant parts of
6925 ssa_undefined_value_p.
6926 (uninit_undefined_value_p): Use has_undefined_value_p.
6927 (warn_uninit, warn_uninitialized_vars,
6928 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
6929 Move from tree-ssa.c.
6930 * tree-ssa.h: Adjust prototypes.
6932 2013-09-17 Jan Hubicka <jh@suse.cz>
6935 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
6936 * ipa-inline.c (can_inline_edge_p): Do not downgrade
6937 FUNCTION_NOT_OPTIMIZED.
6938 * ipa-inline-analysis.c (compute_inline_parameters): Function
6939 not optimized is not inlinable unless it is alwaysinline.
6940 (inline_analyze_function): Force calls in not optimized
6941 function not inlinable.
6943 2013-09-17 Jan Hubicka <jh@suse.cz>
6946 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
6948 * ipa.c (function_and_variable_visibility): Likewise.
6949 * ipa-profile.c (ipa_profile): Likewise.
6951 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
6954 * cgraph.c (cgraph_function_body_availability): Check for ifunc
6955 attribute, and don't inline the resolver in this case.
6957 2013-09-17 Teresa Johnson <tejohnson@google.com>
6959 * coverage.c (get_coverage_counts): Add missing newline.
6961 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6963 PR tree-optimization/58088
6964 * fold-const.c (mask_with_trailing_zeros): New function.
6965 (fold_binary_loc): Make sure we don't recurse infinitely
6966 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
6967 Use mask_with_trailing_zeros where appropriate.
6969 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
6971 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
6972 of 'prev' var to get better distance estimation.
6974 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
6976 * tree-inline.h (struct copy_body_data): Add transform_parameter.
6977 * tree-inline.c (is_parameter_of): New predicate.
6978 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
6979 a parameter has been remapped.
6980 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
6981 (optimize_inline_calls): Initialize transform_parameter.
6982 (copy_gimple_seq_and_replace_locals): Likewise.
6983 (tree_function_versioning): Likewise.
6984 (maybe_inline_call_in_expr): Likewise.
6986 2013-09-17 Nick Clifton <nickc@redhat.com>
6988 * config/msp430/msp430-protos.h: Add prototypes for new functions.
6989 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
6991 (is_attr_func): New function.
6992 (msp430_is_interrupt_func): New function.
6993 (is_naked_func): New function.
6994 (is_reentrant_func): New function.
6995 (is_critical_func): New function.
6996 (msp430_start_function): Add annotations for function attributes.
6997 (msp430_attr): New function.
6998 (msp430_attribute_table): New.
6999 (msp430_function_section): New function.
7000 (TARGET_ASM_FUNCTION_SECTION): Define.
7001 (msp430_builtin): New enum.
7002 (msp430_init_builtins): New function.
7003 (msp430_builtin_devl): New function.
7004 (msp430_expand_builtin): New function.
7005 (TARGET_INIT_BUILTINS): Define.
7006 (TARGET_EXPAND_BUILTINS): Define.
7007 (TARGET_BUILTIN_DECL): Define.
7008 (msp430_expand_prologue): Add support for naked, interrupt,
7009 critical and reentrant functions.
7010 (msp430_expand_epilogue): Likewise.
7011 (msp430_print_operand): Handle 'O' character.
7012 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7014 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7015 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7016 (pushm): Use a 'n' rather than an 'i' constraint.
7017 (msp_return): Add generation of the interrupt return instruction.
7018 (disable_interrupts): New pattern.
7019 (enable_interrupts): New pattern.
7020 (push_intr_state): New pattern.
7021 (pop_intr_state): New pattern.
7022 (bic_SR): New pattern.
7023 (bis_SR): New pattern.
7024 * doc/extend.texi: Document MSP430 function attributes and builtin
7027 2013-09-17 Richard Biener <rguenther@suse.de>
7029 PR tree-optimization/58432
7030 * tree-loop-distribution.c (tree_loop_distribution): Also
7031 scan PHIs for outside loop uses and seed a partition from them.
7033 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7035 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7036 (restructure_reference): Call backtrace_base_for_ref.
7038 2013-09-17 Alan Modra <amodra@gmail.com>
7041 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7044 2013-09-16 DJ Delorie <dj@redhat.com>
7046 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7047 vregs location for RL78/G10.
7048 (rl78_expand_prologue): Avoid SEL on G10.
7049 (rl78_expand_epilogue): Likewise.
7050 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7051 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7052 __RL78_G10__ when appropriate.
7053 (ASM_SPEC): Pass -mg10 along to the assembler.
7054 * config/rl78/rl78.md (sel_rb): Disable for G10.
7055 * config/rl78/rl78.opt: Add -mg10 option.
7056 * config/rl78/t-rl78: Add -mg10 multilib.
7058 2013-09-16 Xinliang David Li <davidxl@google.com>
7060 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7061 * omp-low.c (omp_max_vf): Ditto.
7062 (expand_omp_simd): Ditto.
7063 * tree-vectorizer.c (vectorize_loops): Ditto.
7064 (gate_vect_slp): Ditto.
7065 (gate_increase_alignment): Ditto.
7066 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7067 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7068 (gate_tree_vectorize): Name change.
7069 (tree_vectorize): Ditto.
7070 (pass_vectorize::gate): Call new function.
7071 (pass_vectorize::execute): Ditto.
7072 * opts.c: O3 default setting change.
7073 (finish_options): Check new flag.
7074 * doc/invoke.texi: Document new flags.
7075 * common.opt: New flags.
7077 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7079 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7080 (Condition Code): Don't reference it.
7081 * doc/tm.texi: Regenerate.
7083 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7086 * lra-constraints.c (undo_optional_reloads): Consider all optional
7087 reload even if it did not get a hard reg.
7089 2013-09-16 Teresa Johnson <tejohnson@google.com>
7091 * dumpfile.c (dump_loc): Remove newline emission.
7092 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7093 emission to dump_printf_loc calls where missing.
7094 (vect_mark_for_runtime_alias_test): Ditto.
7095 (vect_analyze_data_ref_dependence): Ditto.
7096 (vect_analyze_data_ref_dependences): Ditto.
7097 (vect_slp_analyze_data_ref_dependence): Ditto.
7098 (vect_slp_analyze_data_ref_dependences): Ditto.
7099 (vect_compute_data_ref_alignment): Ditto.
7100 (vect_update_misalignment_for_peel): Ditto.
7101 (vect_verify_datarefs_alignment): Ditto.
7102 (vector_alignment_reachable_p): Ditto.
7103 (vect_get_data_access_cost): Ditto.
7104 (vect_enhance_data_refs_alignment): Ditto.
7105 (vect_find_same_alignment_drs): Ditto.
7106 (vect_analyze_data_refs_alignment): Ditto.
7107 (vect_analyze_group_access): Ditto.
7108 (vect_analyze_data_ref_access): Ditto.
7109 (vect_analyze_data_ref_accesses): Ditto.
7110 (vect_prune_runtime_alias_test_list): Ditto.
7111 (vect_analyze_data_refs): Ditto.
7112 (vect_create_addr_base_for_vector_ref): Ditto.
7113 (vect_create_data_ref_ptr): Ditto.
7114 (vect_grouped_store_supported): Ditto.
7115 (vect_grouped_load_supported): Ditto.
7116 * value-prof.c (check_counter): Ditto.
7117 (check_ic_target): Ditto.
7118 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7119 (vect_recog_widen_mult_pattern): Ditto.
7120 (vect_recog_widen_sum_pattern): Ditto.
7121 (vect_recog_over_widening_pattern): Ditto.
7122 (vect_recog_widen_shift_pattern): Ditto.
7123 (vect_recog_rotate_pattern): Ditto.
7124 (vect_recog_vector_vector_shift_pattern): Ditto.
7125 (vect_recog_divmod_pattern): Ditto.
7126 (vect_recog_mixed_size_cond_pattern): Ditto.
7127 (vect_recog_bool_pattern): Ditto.
7128 (vect_pattern_recog_1): Ditto.
7129 (vect_pattern_recog): Ditto.
7130 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7131 (vect_is_simple_iv_evolution): Ditto.
7132 (vect_analyze_scalar_cycles_1): Ditto.
7133 (vect_get_loop_niters): Ditto.
7134 (vect_analyze_loop_1): Ditto.
7135 (vect_analyze_loop_form): Ditto.
7136 (vect_analyze_loop_operations): Ditto.
7137 (vect_analyze_loop_2): Ditto.
7138 (vect_analyze_loop): Ditto.
7139 (report_vect_op): Ditto.
7140 (vect_is_slp_reduction): Ditto.
7141 (vect_is_simple_reduction_1): Ditto.
7142 (vect_get_known_peeling_cost): Ditto.
7143 (vect_estimate_min_profitable_iters): Ditto.
7144 (vect_model_reduction_cost): Ditto.
7145 (vect_model_induction_cost): Ditto.
7146 (get_initial_def_for_induction): Ditto.
7147 (vect_create_epilog_for_reduction): Ditto.
7148 (vectorizable_reduction): Ditto.
7149 (vectorizable_induction): Ditto.
7150 (vectorizable_live_operation): Ditto.
7151 (vect_loop_kill_debug_uses): Ditto.
7152 (vect_transform_loop): Ditto.
7153 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7154 (vect_stmt_relevant_p): Ditto.
7155 (process_use): Ditto.
7156 (vect_mark_stmts_to_be_vectorized): Ditto.
7157 (vect_model_simple_cost): Ditto.
7158 (vect_model_promotion_demotion_cost): Ditto.
7159 (vect_model_store_cost): Ditto.
7160 (vect_get_store_cost): Ditto.
7161 (vect_model_load_cost): Ditto.
7162 (vect_get_load_cost): Ditto.
7163 (vect_init_vector_1): Ditto.
7164 (vect_get_vec_def_for_operand): Ditto.
7165 (vect_finish_stmt_generation): Ditto.
7166 (vectorizable_call): Ditto.
7167 (vectorizable_conversion): Ditto.
7168 (vectorizable_assignment): Ditto.
7169 (vectorizable_shift): Ditto.
7170 (vectorizable_operation): Ditto.
7171 (vectorizable_store): Ditto.
7172 (vectorizable_load): Ditto.
7173 (vectorizable_condition): Ditto.
7174 (vect_analyze_stmt): Ditto.
7175 (vect_transform_stmt): Ditto.
7176 (vect_is_simple_use): Ditto.
7177 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7178 (vect_can_advance_ivs_p): Ditto.
7179 (vect_update_ivs_after_vectorizer): Ditto.
7180 (vect_do_peeling_for_loop_bound): Ditto.
7181 (vect_gen_niters_for_prolog_loop): Ditto.
7182 (vect_update_inits_of_drs): Ditto.
7183 (vect_create_cond_for_alias_checks): Ditto.
7184 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7185 (vect_build_slp_tree_1): Ditto.
7186 (vect_supported_load_permutation_p): Ditto.
7187 (vect_analyze_slp_instance): Ditto.
7188 (vect_analyze_slp): Ditto.
7189 (vect_make_slp_decision): Ditto.
7190 (vect_detect_hybrid_slp): Ditto.
7191 (vect_bb_vectorization_profitable_p): Ditto.
7192 (vect_slp_analyze_bb_1): Ditto.
7193 (vect_update_slp_costs_according_to_vf): Ditto.
7194 (vect_get_mask_element): Ditto.
7195 (vect_transform_slp_perm_load): Ditto.
7196 (vect_schedule_slp_instance): Ditto.
7197 (vect_schedule_slp): Ditto.
7198 (vect_slp_transform_bb): Ditto.
7199 * profile.c (read_profile_edge_counts): Ditto.
7200 (compute_branch_probabilities): Ditto.
7201 * coverage.c (get_coverage_counts): Ditto.
7203 2013-09-16 Diego Novillo <dnovillo@google.com>
7205 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7207 2013-09-16 Jan Hubicka <jh@suse.cz>
7209 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7210 abstract functions; for static functions check the presence of body.
7212 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7214 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7215 * config/aarch64/aarch64-simd.md
7216 (aarch64_mla_elt<mode>): New.
7217 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7218 (aarch64_mls_elt<mode>): Likewise.
7219 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7220 (aarch64_fma4_elt<mode>): Likewise.
7221 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7222 (aarch64_fma4_elt_to_128v2df): Likewise.
7223 (aarch64_fma4_elt_to_64df): Likewise.
7224 (fnma<mode>4): Likewise.
7225 (aarch64_fnma4_elt<mode>): Likewise.
7226 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7227 (aarch64_fnma4_elt_to_128v2df): Likewise.
7228 (aarch64_fnma4_elt_to_64df): Likewise.
7229 * config/aarch64/iterators.md (VDQSF): New.
7230 * config/aarch64/arm_neon.h
7231 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
7232 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
7234 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7236 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
7237 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
7238 (aarch64_mul3_elt_to_128df): Likewise.
7239 (aarch64_mul3_elt_to_64v2df): Likewise.
7240 * config/aarch64/iterators.md (VEL): Also handle DFmode.
7242 (VMUL_CHANGE_NLANES) Likewise.
7245 * config/aarch64/arm_neon.h
7246 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
7248 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7250 * config/aarch64/arm_neon.h
7251 (vcvtx_high_f32_f64): Fix parameters.
7253 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7254 Uros Bizjak <ubizjak@gmail.com>
7256 * config/alpha.c: Include tree-ssanames.h.
7258 2013-09-16 Richard Biener <rguenther@suse.de>
7260 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
7261 (dot_rdg_1): Handle control_dd.
7262 (create_edge_for_control_dependence): New function.
7263 (create_rdg_edges): Add control dependences if asked for.
7264 (build_rdg): Likewise.
7265 (generate_loops_for_partition): If there are not necessary
7266 control stmts remove all their dependencies.
7267 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
7268 (distribute_loop): Pass on control dependences.
7269 (tree_loop_distribution): Compute control dependences and remove
7270 restriction on number of loop nodes.
7272 2013-09-16 Jakub Jelinek <jakub@redhat.com>
7274 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
7277 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
7279 * cse.c (try_const_anchors): Punt on CC modes.
7281 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7283 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
7285 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
7288 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
7291 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7294 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
7296 (darwin_objc1_section): Likewise.
7297 (darwin_file_end): Emit Image Info section when required.
7299 2013-09-14 Jan Hubicka <jh@suse.cz>
7301 * tree-into-ssa.c (gate_into_ssa): New.
7302 (pass_data_build_ssa): Use it.
7303 * cgraph.h (expand_thunk): Update prototype.
7304 * cgraphunit.c (analyze_function): Expand thunks early.
7305 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
7306 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
7307 set lowered flag; do not add new function.
7308 (assemble_thunks_and_aliases): Update.
7309 * tree-ssa.c (gate_init_datastructures): New gate.
7310 (pass_data_init_datastructures): Use it.
7312 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7315 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
7316 xmm register set described in the psABI.
7318 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
7320 * dwarf2out.c (should_emit_struct_debug): Add check
7321 for type_decl variable is not NULL.
7323 2013-09-13 Jacek Caban <jacek@codeweavers.com>
7325 * config.gcc: Use new winnt-c.c target hooks
7326 * config/t-winnt: New file
7327 * config/winnt-c.c: New file
7328 * doc/tm.texi.in: Document new hook
7329 * doc/tm.texi: Regenerated
7331 2013-09-13 Jan Hubicka <jh@suse.cz>
7334 * ipa-inline.c (check_callers): New function.
7335 (check_caller_edge): Remove.
7336 (want_inline_function_to_all_callers_p): Also permit alises that are
7338 (inline_to_all_callers): Terminate the walk when devirtualization
7339 introduce new calls.
7341 2013-09-13 Jan Hubicka <jh@suse.cz>
7343 * ipa-inline-analysis.c (struct growth_data): Add node.
7344 (do_estimate_growth_1): Fix detection of recursion.
7346 2013-09-13 Jakub Jelinek <jakub@redhat.com>
7348 PR tree-optimization/58392
7349 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
7350 to avoid shadowing of outer loop variable. If
7351 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
7352 replace_by_duplicate_decl simduid of loops that have it set and
7353 set dest_cfun->has_simduid_loops and/or
7354 dest_cfun->has_force_vect_loops.
7355 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
7356 instead of maybe_lookup_decl.
7357 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
7358 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
7359 call. Copy over force_vect and copy and remap simduid. Set
7360 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
7361 (copy_cfg_body): Remove blocks_to_copy argument. Use
7362 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
7363 caller. Don't set cfun->has_simduid_loops and/or
7364 cfun->has_force_vect_loops here.
7365 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
7367 (expand_call_inline, tree_function_versioning): Adjust copy_body
7370 2013-09-13 Martin Jambor <mjambor@suse.cz>
7373 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
7374 the assert if the edge was a speculative one.
7376 2013-09-13 Richard Biener <rguenther@suse.de>
7378 * tree-data-ref.h (known_dependences_p): Move here ...
7379 * tree-loop-distribution.c (known_dependences_p): ... from here.
7380 (dump_rdg_component, debug_rdg_component): Remove.
7382 (generate_loops_for_partition): Use gimple_uid instead of
7383 relying on matching stmt visit order.
7384 (rdg_build_partitions): Take starting stmt vector.
7385 (ldist_gen): Merge into ...
7386 (distribute_loop): ... this function. Do not compute starting
7388 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
7390 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7392 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
7393 Set type attribute correctly. Set predicable_short_it attribute.
7394 (cmpsi_shiftsi): Remove %? from output template.
7396 2013-09-13 Richard Biener <rguenther@suse.de>
7398 * tree-loop-distribution.c (struct rdg_component,
7399 rdg_defs_used_in_other_loops_p, free_rdg_components,
7400 rdg_build_components): Remove.
7401 (stmts_from_loop): Do not record virtual PHIs.
7402 (generate_loops_for_partition): Skip virtual PHIs.
7403 (build_rdg_partition_for_component): Rename to ...
7404 (build_rdg_partition_for_vertex): ... this and adjust.
7405 (rdg_build_partitions): Take a vector of starting vertices
7406 instead of components. Remove unnecessary leftover handling.
7407 (ldist_gen): Do not build components or record other stores.
7408 (distribute_loop): Do not distribute loops containing stmts
7411 2013-09-13 Christian Bruel <christian.bruel@st.com>
7414 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
7416 2013-09-13 Kai Tietz <ktietz@redhat.com>
7418 * config.gcc: Separate cases for mingw and cygwin targets,
7419 and add 64-bit cygwin target case.
7421 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
7422 dll-export inline-functions.
7423 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
7426 2013-09-13 Jeff Law <law@redhat.com>
7430 2013-09-06 Jeff Law <law@redhat.com>
7432 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7433 edge implied equivalences into successor phis.
7435 2013-09-12 DJ Delorie <dj@redhat.com>
7437 * config/rl78/rl78-virt.md: Change from | to \; for asm line
7440 2013-09-12 Brooks Moses <bmoses@google.com>
7443 * Makefile.in: Do not install driver binaries in $(target)/bin.
7445 2013-09-12 DJ Delorie <dj@redhat.com>
7447 * config/rl78/rl78.opt (mrelax): New.
7448 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
7449 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
7451 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
7453 (rl78_expand_epilogue): Likewise.
7455 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
7458 * lra-eliminations.c (remove_reg_equal_offset_note): New.
7459 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
7460 pointer elimination with using remove_reg_equal_offset_note.
7462 2013-09-12 DJ Delorie <dj@redhat.com>
7464 * config/msp430/: New port.
7465 * config.gcc (msp430): Added.
7466 * doc/invoke.texi: Document MSP430 options.
7467 * doc/install.texi: Document msp430-elf
7468 * doc/md.texi: Document msp430-elf
7469 * doc/contrib.texi: Document msp430-elf
7471 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
7474 2013-09-12 Martin Jambor <mjambor@suse.cz>
7477 * ipa-prop.c (remove_described_reference): Give up if the edge in the
7478 reference descriptor is NULL.
7479 (ipa_edge_removal_hook): If owning a reference descriptor, set its
7482 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
7484 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
7485 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
7486 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
7487 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
7488 make_temp_ssa_name): move to tree-ssanames.h
7489 * tree-ssa-alias.h: Move prototype.
7490 * tree-ssa.h: Include tree-ssanames.h.
7491 * tree-ssanames.c (FREE_SSANAMES): Move to here.
7492 * tree-ssanames.h: New. Move items from tree-flow*.h
7493 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
7495 2013-09-12 Richard Biener <rguenther@suse.de>
7497 PR tree-optimization/58404
7498 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
7499 propagate non-invariant addresses into dereferences wrapped
7500 in component references.
7502 2013-09-12 Richard Biener <rguenther@suse.de>
7504 PR tree-optimization/58402
7505 * passes.def: Move pass_late_warn_uninitialized later.
7507 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
7509 * tree-ssa.h: New. Move content from tree-flow.h and
7511 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
7512 Move prototypes belonging to tree-ssa.c.
7513 * tree-flow-inline.h (redirect_edge_var_map_def,
7514 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
7516 * gimple.h: Adjust prototypes.
7517 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
7519 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7520 * tree.h: Move prototype to tree-ssa.h.
7521 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
7522 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
7523 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
7524 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
7525 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
7526 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
7527 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
7528 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
7529 graphite-blocking.c, graphite-clast-to-gimple.c,
7530 graphite-dependences.c, graphite-interchange.c,
7531 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
7532 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
7533 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
7534 ipa-reference.c, ipa-split.c, ipa-utils.c,
7535 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
7536 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
7537 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
7538 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
7539 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
7540 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
7541 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
7542 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
7543 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
7544 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
7545 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
7546 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
7547 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
7548 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
7549 value-prof.c, var-tracking.c,
7550 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
7552 2013-09-12 Richard Biener <rguenther@suse.de>
7554 PR tree-optimization/58396
7555 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
7556 (build_rdg): Take a loop-nest parameter, fix memleaks.
7557 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
7559 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
7561 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
7564 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
7566 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
7567 The FMA instruction names should have a 'v' prefix.
7569 2013-09-12 Richard Biener <rguenther@suse.de>
7571 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
7572 (dot_rdg): Use popen instead of system in optional code.
7573 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
7574 (already_processed_vertex_p): Adjust.
7575 (has_anti_or_output_dependence, predecessor_has_mem_write,
7576 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
7577 rdg_flag_uses): Remove.
7578 (rdg_flag_vertex): Simplify.
7579 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
7581 (build_rdg_partition_for_component): Process the first vertex
7582 of a component only.
7583 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
7585 2013-09-12 Alan Modra <amodra@gmail.com>
7587 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
7589 2013-09-11 DJ Delorie <dj@redhat.com>
7590 Nick Clifton <nickc@redhat.com>
7592 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
7593 (rl78_stack_based_mem): New.
7594 * config/rl78/constraints.md (Iv08): New.
7603 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
7605 (movsi): Change from expand to insn-and-split.
7606 (ashrsi3): Clobber AX.
7610 * config/rl78/rl78.md (CC_REG): Fix.
7611 (addsi3): Allow memory and immediate operands.
7612 (addsi3_internal): Split into...
7613 (addsi3_internal_virt): ...new, and ...
7614 (addsi3_internal_real): ...new.
7616 (subsi3_internal_virt): New.
7617 (subsi3_internal_real): New.
7618 (mulsi3): Add memory operand.
7619 (mulsi3_rl78): Likewise.
7620 (mulsi3_g13): Likewise.
7621 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
7622 (cbranchqi4_real): Add more constraint options.
7623 (cbranchhi4_real): Expand pattern.
7624 (cbranchhi4_real_signed): New.
7625 (cbranchhi4_real_inverted): New.
7626 (cbranchsi4_real_lt): New.
7627 (cbranchsi4_real_ge): New.
7628 (cbranchsi4_real_signed): New.
7629 (cbranchsi4_real): New.
7631 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
7633 (lshrsi3_virt): Likewise.
7634 (ashlsi3_virt): Likewise.
7635 (cbranchqi4_virt_signed): New.
7636 (cbranchhi4_virt_signed): New.
7637 (cbranchsi4_virt): New.
7638 * config/rl78/rl78.c: Whitespace fixes throughout.
7639 (move_elim_pass): New.
7640 (pass_data_rl78_move_elim): New.
7641 (pass_rl78_move_elim): New.
7642 (make_pass_rl78_move_elim): New.
7643 (rl78_devirt_info): Run devirt earlier.
7644 (rl78_move_elim_info): New.
7645 (rl78_asm_file_start): Register it.
7646 (rl78_split_movsi): New.
7647 (rl78_as_legitimate_address): Allow virtual base registers when
7649 (rl78_addr_space_convert): Remove spurious debug stuff.
7650 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
7651 (rl78_print_operand): More cases for not printing '#'.
7652 (rl78_expand_compare): Remove most of the logic.
7653 (content_memory): New.
7654 (clear_content_memory): New.
7655 (get_content_index): New.
7656 (get_content_name): New.
7657 (display_content_memory): New.
7658 (update_content): New.
7659 (record_content): New.
7660 (already_contains): New.
7661 (insn_ok_now): Re-recog insns with virtual registers.
7662 (add_postponed_content_update): New.
7663 (process_postponed_content_update): New.
7664 (gen_and_emit_move): New.
7665 (transcode_memory_rtx): Record new location content.
7666 Use gen_and_emit_move.
7667 (force_into_acc): New.
7668 (move_to_acc): Use gen_and_emit_move.
7669 (move_from_acc): Likewise.
7670 (move_acc_to_reg): Likewise.
7671 (move_to_x): Likewise.
7672 (move_to_hl): Likewise.
7673 (move_to_de): Likewise.
7674 (rl78_alloc_physical_registers_op1): Record location content.
7675 (has_constraint): New.
7676 (rl78_alloc_physical_registers_op2): Record location content.
7678 (rl78_alloc_physical_registers_ro1): Likewise.
7679 (rl78_alloc_physical_registers_cmp): Likewise.
7680 (rl78_alloc_physical_registers_umul): Likewise.
7681 (rl78_alloc_address_registers_macax): New.
7682 (rl78_alloc_physical_registers): Initialize and set location
7683 content memory as needed.
7684 (rl78_reorg): Make sure split2 is called.
7685 (rl78_rtx_costs): New.
7687 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
7689 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
7690 for (not (neg ...)) and (neg (not ...)) cases.
7692 2013-09-11 Richard Biener <rguenther@suse.de>
7695 * passes.def: Split critical edges before late uninit warning passes.
7696 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
7698 2013-09-11 Jakub Jelinek <jakub@redhat.com>
7700 PR tree-optimization/58385
7701 * fold-const.c (build_range_check): If both low and high are NULL,
7702 use omit_one_operand_loc to preserve exp side-effects.
7704 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7706 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
7708 2013-09-11 Richard Biener <rguenther@suse.de>
7710 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
7711 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
7712 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
7713 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
7714 stmts_from_loop, known_dependences_p, build_empty_rdg,
7715 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
7716 * tree-loop-distribution.c: ... here.
7717 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
7718 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
7719 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
7720 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
7721 * tree-loop-distribution.c: ... here.
7722 * tree-loop-distribution.c: Include gimple-pretty-print.h.
7723 (struct partition_s): Add loops member.
7724 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
7725 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
7726 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
7728 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7729 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7730 Sergey Lega <sergey.s.lega@intel.com>
7731 Anna Tikhonova <anna.tikhonova@intel.com>
7732 Ilya Tocar <ilya.tocar@intel.com>
7733 Andrey Turetskiy <andrey.turetskiy@intel.com>
7734 Ilya Verbin <ilya.verbin@intel.com>
7735 Kirill Yukhin <kirill.yukhin@intel.com>
7736 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7738 * config/i386/constraints.md (k): New.
7740 * config/i386/i386.c (const regclass_map): Add new mask registers.
7741 (dbx_register_map): Ditto.
7742 (dbx64_register_map): Ditto.
7743 (svr4_dbx_register_map): Ditto.
7744 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
7746 (ix86_preferred_reload_class): Disable constants for mask registers.
7747 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
7748 (ix86_hard_regno_mode_ok): Support new mask registers.
7749 (x86_order_regs_for_local_alloc): Ditto.
7750 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
7751 (FIXED_REGISTERS): Add new mask registers.
7752 (CALL_USED_REGISTERS): Ditto.
7753 (REG_ALLOC_ORDER): Ditto.
7754 (VALID_MASK_REG_MODE): New.
7755 (FIRST_MASK_REG): Ditto.
7756 (LAST_MASK_REG): Ditto.
7757 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
7758 (MAYBE_MASK_CLASS_P): New.
7759 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
7760 (REG_CLASS_CONTENTS): Ditto.
7761 (MASK_REGNO_P): New.
7762 (ANY_MASK_REG_P): Ditto.
7763 (HI_REGISTER_NAMES): Add new mask registers.
7764 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
7765 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
7767 (attribute "type"): Add mskmov, msklog.
7768 (attribute "length_immediate"): Support them.
7769 (attribute "memory"): Ditto.
7770 (attribute "prefix_0f"): Ditto.
7771 (*movhi_internal): Support new mask registers.
7772 (*movqi_internal): Ditto.
7773 (define_split): Split out clobber pattern is a logic
7774 insn on mask registers.
7775 (*k<logic><mode>): New.
7776 (*andhi_1): Extend to support mask regs.
7777 (*andqi_1): Extend to support mask regs.
7779 (define_split): Split and-not to and and not if operands
7781 (*<code><mode>_1): Separate HI mode to new pattern...
7782 (*<code>hi_1): This.
7783 (*<code>qi_1): Extend to support mask regs.
7785 (kortestzhi): Ditto.
7786 (kortestchi): Ditto.
7788 (*one_cmpl<mode>2_1): Remove HImode and handle it...
7789 (*one_cmplhi2_1): ...Here, now with mask registers support.
7790 (*one_cmplqi2_1): Support new mask registers.
7791 (HI/QImode arithmetics splitter): Don't split if mask registers
7793 (HI/QImode not splitter): Ditto.
7794 * config/i386/predicated.md (mask_reg_operand): New.
7795 (general_reg_operand): Ditto.
7797 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7799 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
7800 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
7802 2013-09-10 Jeff Law <law@redhat.com>
7804 PR tree-optimization/58380
7805 * tree-ssa-threadupdate.c (thread_block): Recognize another case
7806 of threading through a buried loop header.
7808 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
7809 return value for single successor case.
7811 2013-09-10 Jan Hubicka <jh@suse.cz>
7813 * ipa-devirt.c (ipa_devirt): Enable with LTO.
7815 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
7818 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
7819 support conditional execution.
7820 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
7822 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
7824 * lra.c (lra): Clear lra_optional_reload_pseudos before every
7826 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
7827 Check destination too to check move insn.
7828 (undo_optional_reloads): Add check that the original peudo did not
7829 changed its allocation and the optional reload was inherited on last
7830 inheritance pass. Break loop after deciding to keep optional reload.
7831 (lra_undo_inheritance): Add check that inherited pseudo still in
7834 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
7836 * config/aarch64/aarch64.md (generic_sched): New.
7837 * config/aarch64/aarch64-generic.md (load): Make conditional
7838 on generic_sched attribute.
7839 (nonload): Likewise.
7841 2013-09-10 Jan Hubicka <jh@suse.cz>
7843 * lto-cgraph.c: Include ipa-utils.h.
7844 (compute_ltrans_boundary): Also add possible targets into the boundary.
7846 2013-09-10 Jan Hubicka <jh@suse.cz>
7848 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
7849 VAR_DECL of vtable rather than full expression.
7851 2013-09-10 Jan Hubicka <jh@suse.cz>
7852 Paolo Carlini <paolo.carlini@oracle.com>
7854 * cgraphunit.c (analyze_functions): Save input_location, set it
7855 to UNKNOWN_LOCATION and restore it at the end.
7857 2013-09-10 Martin Jambor <mjambor@suse.cz>
7859 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
7860 represented by a thunk.
7862 2013-09-10 Jeff Law <law@redhat.com>
7864 PR tree-optimization/58343
7865 * tree-ssa-threadupdate.c (thread_block): Identify and disable
7866 jump threading requests through loop headers buried in the middle
7867 of a jump threading path.
7869 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
7870 in return value/type.
7872 2013-09-10 Jakub Jelinek <jakub@redhat.com>
7874 PR rtl-optimization/58365
7875 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
7876 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
7879 2013-09-10 Richard Biener <rguenther@suse.de>
7881 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
7882 * tree-data-ref.c (create_rdg_vertices): Collect all data
7883 references, signal failure to the caller, use data-ref API.
7884 (build_rdg): Compute data references only once. Maintain lifetime
7885 of data references and data dependences from within RDG.
7886 (free_rdg): Free dependence relations.
7887 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
7888 inventing extra dependences.
7889 (distribute_loop): Update for RDG API changes.
7891 2013-09-10 Kai Tietz <ktietz@redhat.com>
7893 * doc/invoke.texi (fms-extensions): Document changed
7894 behavior for ms-abi targets.
7895 * config/i386/i386.c (ix86_option_override_internal):
7896 Set default value of option -fms-extension for ms-abi targets.
7898 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7900 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
7902 2013-09-10 Alan Modra <amodra@gmail.com>
7905 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
7907 2013-09-10 Alan Modra <amodra@gmail.com>
7909 * config/rs6000/predicates.md (add_cint_operand): New.
7910 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
7911 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
7912 using add_cint_operand.
7913 (largetoc_high_plus_aix): Likewise.
7915 2013-09-09 Jakub Jelinek <jakub@redhat.com>
7917 PR tree-optimization/58364
7918 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
7919 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
7920 the current range can't be an unconditional true or false.
7922 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
7924 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
7926 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
7928 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
7930 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
7933 * doc/invoke.texi (-Wdelete-incomplete): Document it.
7935 2013-09-09 Ian Bolton <ian.bolton@arm.com>
7937 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
7938 NO_REGS for immediate that can't be moved directly into FP_REGS.
7940 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7942 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
7943 comparison with negated operand.
7944 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
7947 2013-09-09 Richard Biener <rguenther@suse.de>
7950 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
7951 of a subloop record all its block as affecting loop-closed SSA form.
7953 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7955 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
7956 of an rtx/bitpos pair.
7957 (store_fixed_bit_field): Update accordingly.
7959 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7961 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
7963 * builtins.c (expand_errno_check): Likewise.
7964 * dwarf2cfi.c (init_return_column_size): Likewise.
7965 * except.c (sjlj_mark_call_sites): Likewise.
7966 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
7967 * lra-constraints.c (emit_inc): Likewise.
7968 * ree.c (combine_set_extension): Likewise.
7969 * regmove.c (fixup_match_2): Likewise.
7970 * reload1.c (inc_for_reload): Likewise.
7972 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7974 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
7975 (force_to_mode, simplify_shift_const_1, simplify_comparison):
7976 Use gen_int_mode with the mode of the associated simplify_* call.
7977 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
7978 * expmed.c (expand_shift_1): Likewise.
7979 * function.c (instantiate_virtual_regs_in_insn): Likewise.
7980 * loop-iv.c (iv_number_of_iterations): Likewise.
7981 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
7982 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
7984 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7986 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
7987 of the associated expand_* call.
7988 (asan_emit_stack_protection): Likewise.
7989 * builtins.c (round_trampoline_addr): Likewise.
7990 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
7991 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
7992 (emit_store_flag): Likewise.
7993 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
7994 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
7996 * function.c (instantiate_virtual_regs_in_insn): Likewise.
7997 * ifcvt.c (noce_try_store_flag_constants): Likewise.
7998 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
7999 * modulo-sched.c (generate_prolog_epilog): Likewise.
8000 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8001 (expand_ctz, expand_ffs, expand_unop): Likewise.
8003 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8005 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8006 of the associated gen_rtx_* call.
8007 * caller-save.c (init_caller_save): Likewise.
8008 * combine.c (find_split_point, make_extraction): Likewise.
8009 (make_compound_operation): Likewise.
8010 * dwarf2out.c (mem_loc_descriptor): Likewise.
8011 * explow.c (plus_constant, probe_stack_range): Likewise.
8012 * expmed.c (expand_mult_const): Likewise.
8013 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8014 * reload1.c (init_reload): Likewise.
8015 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8016 * var-tracking.c (adjust_mems): Likewise.
8017 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8018 rather than gen_rtx_fmt_ee.
8020 2013-09-09 Jan Hubicka <jh@suse.cz>
8023 * value-prof.c (gimple_ic): Copy also abnormal edges.
8025 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8027 * asan.c (asan_shadow_cst): Use gen_int_mode.
8029 2013-09-08 Jan Hubicka <jh@suse.cz>
8031 * ipa-profile.c: Add toplevel comment.
8032 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8033 (contains_hot_call_p): New function.
8034 (ipa_propagate_frequency): Set frequencies based on counts when
8036 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8037 profile; do not tamper with profile after inlining if it is read.
8039 2013-09-08 Jan Hubicka <jh@suse.cz>
8041 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8044 2013-09-08 Jan Hubicka <jh@suse.cz>
8046 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8049 2013-09-08 Jeff Law <law@redhat.com>
8052 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8055 2013-09-08 Andi Kleen <ak@linux.intel.com>
8057 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8059 2013-09-08 Jan Hubicka <jh@suse.cz>
8061 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8063 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8065 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8066 for non-debug insns.
8067 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8068 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8069 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8072 2013-09-08 Jan Hubicka <jh@suse.cz>
8074 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8075 targets and devirtualize to BUILT_IN_UNREACHABLE.
8076 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8077 * ipa.c (walk_polymorphic_call_targets): New function.
8078 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8079 functions; use the new timevar.
8080 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8081 was removed from the program.
8082 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8083 not consider it in the walk when its vtable is dead.
8084 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8086 (devirt_variable_node_removal_hook): New function.
8087 (possible_polymorphic_call_targets): Also register
8088 devirt_variable_node_removal_hook.
8089 (ipa_devirt): Do not do non-speculative devirtualization.
8090 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8092 2013-09-08 Jan Hubicka <jh@suse.cz>
8094 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8095 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8096 varpool_remove_variable_insertion_hook): Declare.
8097 * varpool.c (varpool_node_hook_list): New structure.
8098 (first_varpool_node_removal_hook,
8099 first_varpool_variable_insertion_hook): New variables.
8100 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8101 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8102 varpool_remove_variable_insertion_hook,
8103 varpool_call_variable_insertion_hooks): New functions.
8104 (varpool_remove_node): Use it.
8106 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8109 * diagnostic.c (diagnostic_build_prefix): When s.file is
8110 "<built-in>" don't output line and column numbers.
8112 2013-09-06 Jan Hubicka <jh@suse.cz>
8114 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8115 * lto-streamer-out.c: Stream thunks, too.
8116 * lto-streamer-in.c (input_function): Pop cfun here
8117 (lto_read_body): Instead of here.
8119 2013-09-06 Caroline Tice <cmtice@google.com>
8121 * doc/install.texi: Add documentation for the --enable-vtable-verify
8122 and the --disable-libvtv configure options.
8124 2013-09-06 Jeff Law <law@redhat.com>
8126 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8127 edge implied equivalences into successor phis.
8129 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8131 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8133 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8135 * resource.c (mark_target_live_regs): Compute resources taking
8136 into account if a call is predicated or not.
8138 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8140 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8141 the compiler in the function names.
8143 2013-09-06 Jan Hubicka <jh@suse.cz>
8146 * ipa-inline.c (has_caller_p): New function.
8147 (want_inline_function_to_all_callers_p): Use it.
8148 (sum_callers, inline_to_all_callers): Break out from ...
8149 (ipa_inline): ... here.
8151 2013-09-06 Jan Hubicka <jh@suse.cz>
8153 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8154 only when AVX is enabled.
8156 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8158 * config/aarch64/aarch64.md
8159 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8160 is fpsimd_<load/store>2.
8161 (load_pair<mode>): Likewise.
8162 (store_pair<mode>): Likewise.
8164 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8166 * config/arm/types.md (type): Add "mrs" type.
8167 * config/aarch64/aarch64.md
8168 (aarch64_load_tp_hard): Make type "mrs".
8170 (load_tp_hard): Make type "mrs".
8171 * config/arm/cortex-a15.md: Update with new attributes.
8172 * config/arm/cortex-a5.md: Update with new attributes.
8173 * config/arm/cortex-a53.md: Update with new attributes.
8174 * config/arm/cortex-a7.md: Update with new attributes.
8175 * config/arm/cortex-a8.md: Update with new attributes.
8176 * config/arm/cortex-a9.md: Update with new attributes.
8177 * config/arm/cortex-m4.md: Update with new attributes.
8178 * config/arm/cortex-r4.md: Update with new attributes.
8179 * config/arm/fa526.md: Update with new attributes.
8180 * config/arm/fa606te.md: Update with new attributes.
8181 * config/arm/fa626te.md: Update with new attributes.
8182 * config/arm/fa726te.md: Update with new attributes.
8184 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8186 * config/aarch64/aarch64.md
8187 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8188 (*movtf_aarch64): Likewise.
8190 (thumb1_movdi_insn): Use "multiple" for type where more than one
8191 instruction is used for a move.
8192 (*arm32_movhf): Likewise.
8193 (*thumb_movdf_insn): Likewise.
8195 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8197 * config/arm/types.md (type): Rename fcpys to fmov.
8199 (*arm_movsi_vfp): Rename type fcpys as fmov.
8200 (*thumb2_movsi_vfp): Likewise
8201 (*movhf_vfp_neon): Likewise
8202 (*movhf_vfp): Likewise
8203 (*movsf_vfp): Likewise
8204 (*thumb2_movsf_vfp): Likewise
8205 (*movsfcc_vfp): Likewise
8206 (*thumb2_movsfcc_vfp): Likewise
8207 * config/aarch64/aarch64-simd.md
8208 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8209 * config/aarch64/aarch64.md
8210 (*movsi_aarch64): Replace type mov_reg with fmovs.
8211 (*movdi_aarch64): Likewise
8212 (*movsf_aarch64): Likewise
8213 (*movdf_aarch64): Likewise
8215 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
8216 * config/arm/iwmmxt.md
8217 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
8218 * config/arm/arm1020e.md: Update with new attributes.
8219 * config/arm/cortex-a15-neon.md: Update with new attributes.
8220 * config/arm/cortex-a5.md: Update with new attributes.
8221 * config/arm/cortex-a53.md: Update with new attributes.
8222 * config/arm/cortex-a7.md: Update with new attributes.
8223 * config/arm/cortex-a8-neon.md: Update with new attributes.
8224 * config/arm/cortex-a9.md: Update with new attributes.
8225 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8226 * config/arm/cortex-r4f.md: Update with new attributes.
8227 * config/arm/marvell-pj4.md: Update with new attributes.
8228 * config/arm/vfp11.md: Update with new attributes.
8230 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8232 * config/aarch64/aarch64.md
8233 (*madd<mode>): Fix type attribute.
8234 (*maddsi_uxtw): Likewise.
8235 (*msub<mode>): Likewise.
8236 (*msubsi_uxtw): Likewise.
8237 (<su_optab>maddsidi4): Likewise.
8238 (<su_optab>msubsidi4): Likewise.
8240 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8242 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
8243 * config/arm/arm.md (core_cycles): Remove fdiv.
8244 * config/arm/vfp.md:
8245 (*sqrtsf2_vfp): Update for attribute changes.
8246 (*sqrtdf2_vfp): Likewise.
8247 * config/aarch64/aarch64.md:
8248 (sqrt<mode>2): Update for attribute changes.
8249 * config/arm/arm1020e.md: Update with new attributes.
8250 * config/arm/cortex-a15-neon.md: Update with new attributes.
8251 * config/arm/cortex-a5.md: Update with new attributes.
8252 * config/arm/cortex-a53.md: Update with new attributes.
8253 * config/arm/cortex-a7.md: Update with new attributes.
8254 * config/arm/cortex-a8-neon.md: Update with new attributes.
8255 * config/arm/cortex-a9.md: Update with new attributes.
8256 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8257 * config/arm/cortex-r4f.md: Update with new attributes.
8258 * config/arm/marvell-pj4.md: Update with new attributes.
8259 * config/arm/vfp11.md: Update with new attributes.
8261 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8263 * config/arm/types.md
8264 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
8265 * config/aarch64/aarch64.md
8266 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
8268 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
8269 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
8270 (float<GPI:mode><GPF:mode>2): Likewise.
8272 (*truncsisf2_vfp): Update with new attributes.
8273 (*truncsidf2_vfp): Likewise.
8274 (fixuns_truncsfsi2): Likewise.
8275 (fixuns_truncdfsi2): Likewise.
8276 (*floatsisf2_vfp): Likewise.
8277 (*floatsidf2_vfp): Likewise.
8278 (floatunssisf2): Likewise.
8279 (floatunssidf2): Likewise.
8280 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
8281 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
8282 * config/arm/arm1020e.md: Update with new attributes.
8283 * config/arm/cortex-a15-neon.md: Update with new attributes.
8284 * config/arm/cortex-a5.md: Update with new attributes.
8285 * config/arm/cortex-a53.md: Update with new attributes.
8286 * config/arm/cortex-a7.md: Update with new attributes.
8287 * config/arm/cortex-a8-neon.md: Update with new attributes.
8288 * config/arm/cortex-a9.md: Update with new attributes.
8289 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8290 * config/arm/cortex-r4f.md: Update with new attributes.
8291 * config/arm/marvell-pj4.md: Update with new attributes.
8292 * config/arm/vfp11.md: Update with new attributes.
8294 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8296 * config/aarch64/arm_neon.h
8297 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
8298 (vqtbx<1,2,3,4><q>_s8): Likewise.
8300 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8302 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
8303 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
8304 (add<mode>3): Add type attribute.
8305 (add<mode>3): Likewise.
8306 (usadd<mode>3): Likewise.
8307 (ssadd<mode>3): Likewise.
8308 (sub<mode>3): Likewise.
8309 (sub<mode>3): Likewise.
8310 (ussub<mode>3): Likewise.
8311 (sssub<mode>3): Likewise.
8312 (ssmulsa3): Likewise.
8313 (usmulusa3): Likewise.
8314 (arm_usatsihi): Likewise.
8316 (*movdi_vfp): Add types for all instructions.
8317 (*movdi_vfp_cortexa8): Likewise.
8318 (*movhf_vfp_neon): Likewise.
8319 (*movhf_vfp): Likewise.
8320 (*movdf_vfp): Likewise.
8321 (*thumb2_movdf_vfp): Likewise.
8322 (*thumb2_movdfcc_vfp): Likewise.
8323 * config/arm/arm.md: Add type attribute to all insn patterns.
8324 (*thumb1_adddi3): Add type attribute.
8325 (*arm_adddi3): Likewise.
8326 (*adddi_sesidi_di): Likewise.
8327 (*adddi_zesidi_di): Likewise.
8328 (*thumb1_addsi3): Likewise.
8329 (addsi3_compare0): Likewise.
8330 (*addsi3_compare0_scratch): Likewise.
8331 (*compare_negsi_si): Likewise.
8332 (cmpsi2_addneg): Likewise.
8333 (*addsi3_carryin_<optab>): Likewise.
8334 (*addsi3_carryin_alt2_<optab>): Likewise.
8335 (*addsi3_carryin_clobercc_<optab>): Likewise.
8336 (*subsi3_carryin): Likewise.
8337 (*subsi3_carryin_const): Likewise.
8338 (*subsi3_carryin_compare): Likewise.
8339 (*subsi3_carryin_compare_const): Likewise.
8340 (*arm_subdi3): Likewise.
8341 (*thumb_subdi3): Likewise.
8342 (*subdi_di_zesidi): Likewise.
8343 (*subdi_di_sesidi): Likewise.
8344 (*subdi_zesidi_di): Likewise.
8345 (*subdi_sesidi_di): Likewise.
8346 (*subdi_zesidi_ze): Likewise.
8347 (thumb1_subsi3_insn): Likewise.
8348 (*arm_subsi3_insn): Likewise.
8349 (*anddi3_insn): Likewise.
8350 (*anddi_zesidi_di): Likewise.
8351 (*anddi_sesdi_di): Likewise.
8352 (*ne_zeroextracts): Likewise.
8353 (*ne_zeroextracts): Likewise.
8354 (*ite_ne_zeroextr): Likewise.
8355 (*ite_ne_zeroextr): Likewise.
8356 (*anddi_notdi_di): Likewise.
8357 (*anddi_notzesidi): Likewise.
8358 (*anddi_notsesidi): Likewise.
8359 (andsi_notsi_si): Likewise.
8360 (thumb1_bicsi3): Likewise.
8361 (*iordi3_insn): Likewise.
8362 (*iordi_zesidi_di): Likewise.
8363 (*iordi_sesidi_di): Likewise.
8364 (*thumb1_iorsi3_insn): Likewise.
8365 (*xordi3_insn): Likewise.
8366 (*xordi_zesidi_di): Likewise.
8367 (*xordi_sesidi_di): Likewise.
8368 (*arm_xorsi3): Likewise.
8369 (*andsi_iorsi3_no): Likewise.
8370 (*smax_0): Likewise.
8371 (*smax_m1): Likewise.
8372 (*arm_smax_insn): Likewise.
8373 (*smin_0): Likewise.
8374 (*arm_smin_insn): Likewise.
8375 (*arm_umaxsi3): Likewise.
8376 (*arm_uminsi3): Likewise.
8377 (*minmax_arithsi): Likewise.
8378 (*minmax_arithsi_): Likewise.
8379 (*satsi_<SAT:code>): Likewise.
8380 (arm_ashldi3_1bit): Likewise.
8381 (arm_ashrdi3_1bit): Likewise.
8382 (arm_lshrdi3_1bit): Likewise.
8383 (*arm_negdi2): Likewise.
8384 (*thumb1_negdi2): Likewise.
8385 (*arm_negsi2): Likewise.
8386 (*thumb1_negsi2): Likewise.
8387 (*negdi_extendsid): Likewise.
8388 (*negdi_zero_extend): Likewise.
8389 (*arm_abssi2): Likewise.
8390 (*thumb1_abssi2): Likewise.
8391 (*arm_neg_abssi2): Likewise.
8392 (*thumb1_neg_abss): Likewise.
8393 (one_cmpldi2): Likewise.
8394 (extend<mode>di2): Likewise.
8395 (*compareqi_eq0): Likewise.
8396 (*arm_extendhisi2addsi): Likewise.
8397 (*arm_movdi): Likewise.
8398 (*thumb1_movdi_insn): Likewise.
8399 (*arm_movt): Likewise.
8400 (*thumb1_movsi_insn): Likewise.
8401 (pic_add_dot_plus_four): Likewise.
8402 (pic_add_dot_plus_eight): Likewise.
8403 (tls_load_dot_plus_eight): Likewise.
8404 (*thumb1_movhi_insn): Likewise.
8405 (*thumb1_movsf_insn): Likewise.
8406 (*movdf_soft_insn): Likewise.
8407 (*thumb_movdf_insn): Likewise.
8408 (cbranchsi4_insn): Likewise.
8409 (cbranchsi4_scratch): Likewise.
8410 (*negated_cbranchsi4): Likewise.
8411 (*tbit_cbranch): Likewise.
8412 (*tlobits_cbranch): Likewise.
8413 (*tstsi3_cbranch): Likewise.
8414 (*cbranchne_decr1): Likewise.
8415 (*addsi3_cbranch): Likewise.
8416 (*addsi3_cbranch_scratch): Likewise.
8417 (*arm_cmpdi_insn): Likewise.
8418 (*arm_cmpdi_unsig): Likewise.
8419 (*arm_cmpdi_zero): Likewise.
8420 (*thumb_cmpdi_zero): Likewise.
8421 (*deleted_compare): Likewise.
8422 (*mov_scc): Likewise.
8423 (*mov_negscc): Likewise.
8424 (*mov_notscc): Likewise.
8425 (*cstoresi_eq0_thumb1_insn): Likewise.
8426 (cstoresi_nltu_thumb1): Likewise.
8427 (cstoresi_ltu_thu): Likewise.
8428 (thumb1_addsi3_addgeu): Likewise.
8429 (*arm_jump): Likewise.
8430 (*thumb_jump): Likewise.
8431 (*check_arch2): Likewise.
8432 (arm_casesi_internal): Likewise.
8433 (thumb1_casesi_dispatch): Likewise.
8434 (*arm_indirect_jump): Likewise.
8435 (*thumb1_indirect_jump): Likewise.
8437 (*and_scc): Likewise.
8438 (*ior_scc): Likewise.
8439 (*compare_scc): Likewise.
8440 (*cond_move): Likewise.
8441 (*cond_arith): Likewise.
8442 (*cond_sub): Likewise.
8443 (*cmp_ite0): Likewise.
8444 (*cmp_ite1): Likewise.
8445 (*cmp_and): Likewise.
8446 (*cmp_ior): Likewise.
8447 (*ior_scc_scc): Likewise.
8448 (*ior_scc_scc_cmp): Likewise.
8449 (*and_scc_scc): Likewise.
8450 (*and_scc_scc_cmp): Likewise.
8451 (*and_scc_scc_nod): Likewise.
8452 (*negscc): Likewise.
8453 (movcond_addsi): Likewise.
8454 (movcond): Likewise.
8455 (*ifcompare_plus_move): Likewise.
8456 (*if_plus_move): Likewise.
8457 (*ifcompare_move_plus): Likewise.
8458 (*if_move_plus): Likewise.
8459 (*ifcompare_arith_arith): Likewise.
8460 (*if_arith_arith): Likewise.
8461 (*ifcompare_arith_move): Likewise.
8462 (*if_arith_move): Likewise.
8463 (*ifcompare_move_arith): Likewise.
8464 (*if_move_arith): Likewise.
8465 (*ifcompare_move_not): Likewise.
8466 (*if_move_not): Likewise.
8467 (*ifcompare_not_move): Likewise.
8468 (*if_not_move): Likewise.
8469 (*ifcompare_shift_move): Likewise.
8470 (*if_shift_move): Likewise.
8471 (*ifcompare_move_shift): Likewise.
8472 (*if_move_shift): Likewise.
8473 (*ifcompare_shift_shift): Likewise.
8474 (*ifcompare_not_arith): Likewise.
8475 (*ifcompare_arith_not): Likewise.
8476 (*if_arith_not): Likewise.
8477 (*ifcompare_neg_move): Likewise.
8478 (*if_neg_move): Likewise.
8479 (*ifcompare_move_neg): Likewise.
8480 (*if_move_neg): Likewise.
8481 (prologue_thumb1_interwork): Likewise.
8482 (*cond_move_not): Likewise.
8483 (*sign_extract_onebit): Likewise.
8484 (*not_signextract_onebit): Likewise.
8485 (stack_tie): Likewise.
8486 (align_4): Likewise.
8487 (align_8): Likewise.
8488 (consttable_end): Likewise.
8489 (consttable_1): Likewise.
8490 (consttable_2): Likewise.
8491 (consttable_4): Likewise.
8492 (consttable_8): Likewise.
8493 (consttable_16): Likewise.
8494 (*thumb1_tablejump): Likewise.
8495 (prefetch): Likewise.
8496 (force_register_use): Likewise.
8497 (thumb_eh_return): Likewise.
8498 (load_tp_hard): Likewise.
8499 (load_tp_soft): Likewise.
8500 (tlscall): Likewise.
8501 (*arm_movtas_ze): Likewise.
8502 (*arm_rev): Likewise.
8503 (*arm_revsh): Likewise.
8504 (*arm_rev16): Likewise.
8505 * config/arm/thumb2.md
8506 (*thumb2_smaxsi3): Likewise.
8507 (*thumb2_sminsi3): Likewise.
8508 (*thumb32_umaxsi3): Likewise.
8509 (*thumb2_uminsi3): Likewise.
8510 (*thumb2_negdi2): Likewise.
8511 (*thumb2_abssi2): Likewise.
8512 (*thumb2_neg_abss): Likewise.
8513 (*thumb2_movsi_insn): Likewise.
8514 (tls_load_dot_plus_four): Likewise.
8515 (*thumb2_movhi_insn): Likewise.
8516 (*thumb2_mov_scc): Likewise.
8517 (*thumb2_mov_negs): Likewise.
8518 (*thumb2_mov_negs): Likewise.
8519 (*thumb2_mov_nots): Likewise.
8520 (*thumb2_mov_nots): Likewise.
8521 (*thumb2_movsicc_): Likewise.
8522 (*thumb2_movsfcc_soft_insn): Likewise.
8523 (*thumb2_indirect_jump): Likewise.
8524 (*thumb2_and_scc): Likewise.
8525 (*thumb2_ior_scc): Likewise.
8526 (*thumb2_ior_scc_strict_it): Likewise.
8527 (*thumb2_cond_move): Likewise.
8528 (*thumb2_cond_arith): Likewise.
8529 (*thumb2_cond_ari): Likewise.
8530 (*thumb2_cond_sub): Likewise.
8531 (*thumb2_negscc): Likewise.
8532 (*thumb2_movcond): Likewise.
8533 (thumb2_casesi_internal): Likewise.
8534 (thumb2_casesi_internal_pic): Likewise.
8535 (*thumb2_alusi3_short): Likewise.
8536 (*thumb2_mov<mode>_shortim): Likewise.
8537 (*thumb2_addsi_short): Likewise.
8538 (*thumb2_subsi_short): Likewise.
8539 (thumb2_addsi3_compare0): Likewise.
8540 (*thumb2_cbz): Likewise.
8541 (*thumb2_cbnz): Likewise.
8542 (*thumb2_one_cmplsi2_short): Likewise.
8543 (*thumb2_negsi2_short): Likewise.
8544 (*orsi_notsi_si): Likewise.
8545 * config/arm/arm1020e.md: Update with new attributes.
8546 * config/arm/arm1026ejs.md: Update with new attributes.
8547 * config/arm/arm1136jfs.md: Update with new attributes.
8548 * config/arm/arm926ejs.md: Update with new attributes.
8549 * config/arm/cortex-a15.md: Update with new attributes.
8550 * config/arm/cortex-a5.md: Update with new attributes.
8551 * config/arm/cortex-a53.md: Update with new attributes.
8552 * config/arm/cortex-a7.md: Update with new attributes.
8553 * config/arm/cortex-a8.md: Update with new attributes.
8554 * config/arm/cortex-a9.md: Update with new attributes.
8555 * config/arm/cortex-m4.md: Update with new attributes.
8556 * config/arm/cortex-r4.md: Update with new attributes.
8557 * config/arm/fa526.md: Update with new attributes.
8558 * config/arm/fa606te.md: Update with new attributes.
8559 * config/arm/fa626te.md: Update with new attributes.
8560 * config/arm/fa726te.md: Update with new attributes.
8562 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8564 * config/aarch64/aarch64-simd.md
8565 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
8566 <vwx> iterator to ensure correct register choice.
8567 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
8568 (aarch64_sqdmull_n<mode>): Likewise.
8569 (aarch64_sqdmull2_n<mode>_internal): Likewise.
8570 * config/aarch64/arm_neon.h
8571 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
8572 (vml<as><q>_n_<su>16): Likewise.
8573 (vml<as>l_high_lane<q>_<su>16): Likewise.
8574 (vml<as>l_high_n_<su>16): Likewise.
8575 (vml<as>l_lane<q>_<su>16): Likewise.
8576 (vml<as>l_n_<su>16): Likewise.
8577 (vmul<q>_lane<q>_<su>16): Likewise.
8578 (vmul<q>_n_<su>16): Likewise.
8579 (vmull_lane<q>_<su>16): Likewise.
8580 (vmull_n_<su>16): Likewise.
8581 (vmull_high_lane<q>_<su>16): Likewise.
8582 (vmull_high_n_<su>16): Likewise.
8583 (vqrdmulh<q>_n_s16): Likewise.
8585 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
8587 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
8588 have the correct lane parameter.
8590 2013-09-06 Richard Biener <rguenther@suse.de>
8592 * cfganal.c (control_dependences::~control_dependences):
8593 Properly free all of the vector.
8595 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
8598 * config/i386/i386.c (ix86_conditional_register_usage):
8599 Proper initialize extended SSE registers.
8601 2013-09-06 Jan Hubicka <jh@suse.cz>
8603 PR tree-optimization/58311
8604 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
8606 2013-09-06 Jan Hubicka <jh@suse.cz>
8608 * Makefile.in (tree-sra.o): Update dependencies.
8609 * tree-sra.c: Include ipa-utils.h
8610 (scan_function): Use recursive_call_p.
8611 (has_caller_p): New function.
8612 (cgraph_for_node_and_aliases): Count also callers of aliases.
8614 2013-09-06 Jan Hubicka <jh@suse.cz>
8617 * cgraph.h (symtab_semantically_equivalent_p): Declare.
8618 * tree-tailcall.c: Include ipa-utils.h.
8619 (find_tail_calls): Use it.
8620 * ipa-pure-const.c (check_call): Likewise.
8621 * ipa-utils.c (recursive_call_p): New function.
8622 * ipa-utils.h (recursive_call_p): Dclare.
8623 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
8624 (symtab_semantically_equivalent_p): New function.
8625 * Makefile.in (tree-tailcall.o): Update dependencies.
8627 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8629 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
8632 2013-09-06 Richard Biener <rguenther@suse.de>
8634 * lto-streamer.h (lto_global_var_decls): Remove.
8635 * Makefile.in (OBJS): Remove lto-symtab.o.
8636 (lto-symtab.o): Remove.
8637 (GTFILES): Remove lto-symtab.c
8638 * lto-symtab.c: Move to lto/
8640 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8642 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
8643 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
8644 (UNSPEC_FPINT_RINT): New constant definitions.
8645 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
8646 definition with 2 attributes.
8647 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
8648 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
8651 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8653 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
8654 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
8655 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
8656 attribute to "z196_alone".
8657 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
8660 2013-09-06 Richard Biener <rguenther@suse.de>
8662 * basic-block.h (class control_dependences): New.
8663 * tree-ssa-dce.c (control_dependence_map): Remove.
8665 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
8666 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
8667 find_pdom, find_control_dependence, find_all_control_dependences):
8669 (mark_control_dependent_edges_necessary,
8670 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
8671 tree_dce_done, perform_tree_ssa_dce): Adjust.
8672 * cfganal.c (set_control_dependence_map_bit,
8673 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
8674 find_all_control_dependences): Move from tree-ssa-dce.c and
8675 implement as methods of control_dependences class.
8676 (control_dependences::control_dependences): New.
8677 (control_dependences::~control_dependences): Likewise.
8678 (control_dependences::get_edges_dependent_on): Likewise.
8679 (control_dependences::get_edge): Likewise.
8681 2013-09-04 Jan Hubicka <jh@suse.cz>
8683 * tree.c (types_same_for_odr): Drop overactive check.
8684 * ipa-devirt.c (hash_type_name): Likewise.
8686 2013-09-04 Jan Hubicka <jh@suse.cz>
8688 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
8689 (analyze_functions): ... here.
8691 2013-09-04 Jan Hubicka <jh@suse.cz>
8694 * cgraphunit.c (analyze_functions): Clear AUX fields
8695 after processing; initialize assembler name has.
8697 2013-09-05 Jeff Law <law@redhat.com>
8699 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
8700 from thread_around_empty_block. Record threading path into PATH.
8701 Recurse if threading through the initial block is successful.
8702 (thread_across_edge): Corresponding changes to slightly simplify.
8704 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
8706 * config/aarch64/aarch64.md
8707 (type): Remove frecpe, frecps, frecpx.
8708 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
8709 fix to be a TARGET_SIMD instruction.
8710 (aarch64_frecps): Remove.
8711 * config/aarch64/aarch64-simd.md
8712 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
8713 (aarch64_frecps<mode>): Handle all float/vector of float modes.
8715 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
8716 Sofiane Naci <sofiane.naci@arm.com>
8718 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
8719 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
8720 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
8721 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
8722 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
8723 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
8724 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
8725 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
8726 into "clz, "rbit". Rename "shift" to "shift_imm".
8727 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
8728 changes. Update for attribute changes all occurrences of arlo_* and
8730 * config/arm/arm-fixed.md: Update for attribute changes
8731 all occurrences of arlo_* types.
8732 * config/arm/thumb2.md: Update for attribute changes all occurrences
8734 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
8735 (cortexa7_older_only): Likewise.
8736 (cortexa7_younger): Likewise.
8737 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8738 (1020alu_shift_op): Likewise.
8739 (1020alu_shift_reg_op): Likewise.
8740 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8741 (alu_shift_op): Likewise.
8742 (alu_shift_reg_op): Likewise.
8743 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8744 (11_alu_shift_op): Likewise.
8745 (11_alu_shift_reg_op): Likewise.
8746 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8747 (9_alu_shift_reg_op): Likewise.
8748 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
8750 (cortex_a15_alu_shift): Likewise.
8751 (cortex_a15_alu_shift_reg): Likewise.
8752 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
8754 (cortex_a5_alu_shift): Likewise.
8755 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
8757 (cortex_a53_alu_shift): Likewise.
8758 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
8760 (cortex_a7_alu_reg): Likewise.
8761 (cortex_a7_alu_shift): Likewise.
8762 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
8764 (cortex_a8_alu_shift): Likewise.
8765 (cortex_a8_alu_shift_reg): Likewise.
8766 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8767 (cortex_a9_dp_shift): Likewise.
8768 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
8770 * config/arm/cortex-r4.md
8771 (cortex_r4_alu): Update for attribute changes.
8772 (cortex_r4_mov): Likewise.
8773 (cortex_r4_alu_shift_reg): Likewise.
8774 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8775 (526_alu_shift_op): Likewise.
8776 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8777 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8778 (626te_alu_shift_op): Likewise.
8779 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
8780 (726te_alu_shift_op): Likewise.
8781 (726te_alu_shift_reg_op): Likewise.
8782 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8783 (mp626_alu_shift_op): Likewise.
8784 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
8785 (pj4_alu_conds): Likewise.
8786 (pj4_shift): Likewise.
8787 (pj4_shift_conds): Likewise.
8788 (pj4_alu_shift): Likewise.
8789 (pj4_alu_shift_conds): Likewise.
8790 * config/aarch64/aarch64.md: Update for attribute change
8791 all occurrences of arlo_* and shift* types.
8793 2013-09-05 Mike Stump <mikestump@comcast.net>
8795 * tree.h: Move documentation for tree_function_decl to tree-core.h
8796 with the declaration.
8798 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
8801 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
8802 looking for widest mode.
8804 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
8806 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
8808 2013-09-05 Richard Biener <rguenther@suse.de>
8810 PR tree-optimization/58137
8811 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
8812 Do not create vectors of pointers.
8813 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
8814 types for the components of the vector initializer.
8815 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
8816 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
8818 2013-09-05 Martin Jambor <mjambor@suse.cz>
8820 * ipa-prop.c (remove_described_reference): Accept missing references,
8821 return false if that hppens, otherwise return true.
8822 (cgraph_node_for_jfunc): New function.
8823 (try_decrement_rdesc_refcount): Likewise.
8824 (try_make_edge_direct_simple_call): Use them.
8825 (ipa_edge_removal_hook): Remove references from rdescs.
8826 (ipa_edge_duplication_hook): Clone rdescs and their references
8827 when the new edge has the same caller as the old one.
8828 * cgraph.c (cgraph_resolve_speculation): Remove speculative
8829 reference before removing any edges.
8831 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
8833 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
8835 * thumb2.md (thumb2_storewb_parisi): New pattern.
8837 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
8839 * config/aarch64/aarch64-option-extensions.def: Add
8840 AARCH64_OPT_EXTENSION of 'crc'.
8841 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
8842 (AARCH64_ISA_CRC): Ditto.
8843 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
8844 description of the CRC extension.
8846 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
8848 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
8849 * config/rs6000/linux.h: Ditto.
8850 * alpha/linux.h: Ditto.
8851 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
8852 no_c99_libc_has_function.
8853 * config/c6x/uclinux-elf.h: Ditto.
8854 * config/lm32/uclinux-elf.h: Ditto.
8855 * config/m68k/uclinux.h: Ditto.
8856 * config/moxie/uclinux.h: Ditto.
8857 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
8858 (crisv32-*-linux*, cris-*-linux*): Ditto.
8859 * config/bfin/bfin.c: Include "tm_p.h".
8861 2013-09-05 Richard Biener <rguenther@suse.de>
8863 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
8864 check for a definition without a basic-block.
8866 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
8867 Sofiane Naci <sofiane.naci@arm.com>
8869 * config/aarch64/aarch64.md
8870 (*movti_aarch64): Rename r_2_f and f_2_r.
8871 (*movsf_aarch64): Likewise.
8872 (*movdf_aarch64): Likewise.
8873 (*movtf_aarch64): Likewise.
8874 (aarch64_movdi_<mode>low): Likewise.
8875 (aarch64_movdi_<mode>high): Likewise.
8876 (aarch64_mov<mode>high_di): Likewise.
8877 (aarch64_mov<mode>low_di): Likewise.
8878 (aarch64_movtilow_tilow): Likewise.
8879 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
8880 values to config/arm/types.md
8881 (attribute "conds"): Update for attribute change.
8882 (anddi3_insn): Likewise.
8883 (iordi3_insn): Likewise.
8884 (xordi3_insn): Likewise.
8885 (one_cmpldi2): Likewise.
8886 * config/arm/types.md (type): Add Neon types.
8887 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
8888 use "type" attribute.
8889 (movmisalign<mode>_neon_store): Likewise.
8890 (movmisalign<mode>_neon_load): Likewise.
8891 (vec_set<mode>_internal): Likewise.
8892 (vec_setv2di_internal): Likewise.
8893 (vec_extract<mode>): Likewise.
8894 (vec_extractv2di): Likewise.
8895 (add<mode>3_neon): Likewise.
8896 (adddi3_neon): Likewise.
8897 (sub<mode>3_neon): Likewise.
8898 (subdi3_neon): Likewise.
8899 (mul<mode>3_neon): Likewise.
8900 (mul<mode>3add<mode>_neon): Likewise.
8901 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
8902 (fma<VCVTF:mode>4)): Likewise.
8903 (fma<VCVTF:mode>4_intrinsic): Likewise.
8904 (fmsub<VCVTF:mode>4)): Likewise.
8905 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
8906 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
8907 (ior<mode>3): Likewise.
8908 (and<mode>3): Likewise.
8909 (anddi3_neon): Likewise.
8910 (orn<mode>3_neon): Likewise.
8911 (orndi3_neon): Likewise.
8912 (bic<mode>3_neon): Likewise.
8913 (bicdi3_neon): Likewise.
8914 (xor<mode>3): Likewise.
8915 (one_cmpl<mode>2): Likewise.
8916 (abs<mode>2): Likewise.
8917 (neg<mode>2): Likewise.
8918 (umin<mode>3_neon): Likewise.
8919 (umax<mode>3_neon): Likewise.
8920 (smin<mode>3_neon): Likewise.
8921 (smax<mode>3_neon): Likewise.
8922 (vashl<mode>3): Likewise.
8923 (vashr<mode>3_imm): Likewise.
8924 (vlshr<mode>3_imm): Likewise.
8925 (ashl<mode>3_signed): Likewise.
8926 (ashl<mode>3_unsigned): Likewise.
8927 (neon_load_count): Likewise.
8928 (ashldi3_neon_noclobber): Likewise.
8929 (signed_shift_di3_neon): Likewise.
8930 (unsigned_shift_di3_neon): Likewise.
8931 (ashrdi3_neon_imm_noclobber): Likewise.
8932 (lshrdi3_neon_imm_noclobber): Likewise.
8933 (widen_ssum<mode>3): Likewise.
8934 (widen_usum<mode>3): Likewise.
8935 (quad_halves_<code>v4si): Likewise.
8936 (quad_halves_<code>v4sf): Likewise.
8937 (quad_halves_<code>v8hi): Likewise.
8938 (quad_halves_<code>v16qi): Likewise.
8939 (reduc_splus_v2di): Likewise.
8940 (neon_vpadd_internal<mode>): Likewise.
8941 (neon_vpsmin<mode>): Likewise.
8942 (neon_vpsmax<mode>): Likewise.
8943 (neon_vpumin<mode>): Likewise.
8944 (neon_vpumax<mode>): Likewise.
8945 (ss_add<mode>_neon): Likewise.
8946 (us_add<mode>_neon): Likewise.
8947 (ss_sub<mode>_neon): Likewise.
8948 (us_sub<mode>_neon): Likewise.
8949 (neon_vadd<mode>_unspec): Likewise.
8950 (neon_vaddl<mode>): Likewise.
8951 (neon_vaddw<mode>): Likewise.
8952 (neon_vhadd<mode>): Likewise.
8953 (neon_vqadd<mode>): Likewise.
8954 (neon_vaddhn<mode>): Likewise.
8955 (neon_vmul<mode>): Likewise.
8956 (neon_vmla<mode>): Likewise.
8957 (neon_vmlal<mode>): Likewise.
8958 (neon_vmls<mode>): Likewise.
8959 (neon_vmlsl<mode>): Likewise.
8960 (neon_vqdmulh<mode>): Likewise.
8961 (neon_vqdmlal<mode>): Likewise.
8962 (neon_vqdmlsl<mode>): Likewise.
8963 (neon_vmull<mode>): Likewise.
8964 (neon_vqdmull<mode>): Likewise.
8965 (neon_vsub<mode>_unspec): Likewise.
8966 (neon_vsubl<mode>): Likewise.
8967 (neon_vsubw<mode>): Likewise.
8968 (neon_vqsub<mode>): Likewise.
8969 (neon_vhsub<mode>): Likewise.
8970 (neon_vsubhn<mode>): Likewise.
8971 (neon_vceq<mode>): Likewise.
8972 (neon_vcge<mode>): Likewise.
8973 (neon_vcgeu<mode>): Likewise.
8974 (neon_vcgt<mode>): Likewise.
8975 (neon_vcgtu<mode>): Likewise.
8976 (neon_vcle<mode>): Likewise.
8977 (neon_vclt<mode>): Likewise.
8978 (neon_vcage<mode>): Likewise.
8979 (neon_vcagt<mode>): Likewise.
8980 (neon_vtst<mode>): Likewise.
8981 (neon_vabd<mode>): Likewise.
8982 (neon_vabdl<mode>): Likewise.
8983 (neon_vaba<mode>): Likewise.
8984 (neon_vabal<mode>): Likewise.
8985 (neon_vmax<mode>): Likewise.
8986 (neon_vmin<mode>): Likewise.
8987 (neon_vpaddl<mode>): Likewise.
8988 (neon_vpadal<mode>): Likewise.
8989 (neon_vpmax<mode>): Likewise.
8990 (neon_vpmin<mode>): Likewise.
8991 (neon_vrecps<mode>): Likewise.
8992 (neon_vrsqrts<mode>): Likewise.
8993 (neon_vqabs<mode>): Likewise.
8994 (neon_vqneg<mode>): Likewise.
8995 (neon_vcls<mode>): Likewise.
8996 (clz<mode>2): Likewise.
8997 (popcount<mode>2): Likewise.
8998 (neon_vrecpe): Likewise.
8999 (neon_vrsqrte): Likewise.
9000 (neon_vget_lane<mode>_sext_internal): Likewise.
9001 (neon_vget_lane<mode>_zext_internal): Likewise.
9002 (neon_vdup_n<mode>): Likewise.
9003 (neon_vdup_nv2di): Likewise.
9004 (neon_vdpu_lane<mode>_internal): Likewise.
9005 (neon_vswp<mode>): Likewise.
9006 (float<mode><V_cvtto>2): Likewise.
9007 (floatuns<mode><V_cvtto>2): Likewise.
9008 (fix_trunc<mode><V_cvtto>)2): Likewise
9009 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9010 (neon_vcvt<mode>): Likewise.
9011 (neon_vcvtv4sfv4hf): Likewise.
9012 (neon_vcvtv4hfv4sf): Likewise.
9013 (neon_vcvt_n<mode>): Likewise.
9014 (neon_vmovn<mode>): Likewise.
9015 (neon_vqmovn<mode>): Likewise.
9016 (neon_vqmovun<mode>): Likewise.
9017 (neon_vmovl<mode>): Likewise.
9018 (neon_vmul_lane<mode>): Likewise.
9019 (neon_vmull_lane<mode>): Likewise.
9020 (neon_vqdmull_lane<mode>): Likewise.
9021 (neon_vqdmulh_lane<mode>): Likewise.
9022 (neon_vmla_lane<mode>): Likewise.
9023 (neon_vmlal_lane<mode>): Likewise.
9024 (neon_vqdmlal_lane<mode>): Likewise.
9025 (neon_vmls_lane<mode>): Likewise.
9026 (neon_vmlsl_lane<mode>): Likewise.
9027 (neon_vqdmlsl_lane<mode>): Likewise.
9028 (neon_vext<mode>): Likewise.
9029 (neon_vrev64<mode>): Likewise.
9030 (neon_vrev32<mode>): Likewise.
9031 (neon_vrev16<mode>): Likewise.
9032 (neon_vbsl<mode>_internal): Likewise.
9033 (neon_vshl<mode>): Likewise.
9034 (neon_vqshl<mode>): Likewise.
9035 (neon_vshr_n<mode>): Likewise.
9036 (neon_vshrn_n<mode>): Likewise.
9037 (neon_vqshrn_n<mode>): Likewise.
9038 (neon_vqshrun_n<mode>): Likewise.
9039 (neon_vshl_n<mode>): Likewise.
9040 (neon_vqshl_n<mode>): Likewise.
9041 (neon_vqshlu_n<mode>): Likewise.
9042 (neon_vshll_n<mode>): Likewise.
9043 (neon_vsra_n<mode>): Likewise.
9044 (neon_vsri_n<mode>): Likewise.
9045 (neon_vsli_n<mode>): Likewise.
9046 (neon_vtbl1v8qi): Likewise.
9047 (neon_vtbl2v8qi): Likewise.
9048 (neon_vtbl3v8qi): Likewise.
9049 (neon_vtbl4v8qi): Likewise.
9050 (neon_vtbx1v8qi): Likewise.
9051 (neon_vtbx2v8qi): Likewise.
9052 (neon_vtbx3v8qi): Likewise.
9053 (neon_vtbx4v8qi): Likewise.
9054 (neon_vtrn<mode>_internal): Likewise.
9055 (neon_vzip<mode>_internal): Likewise.
9056 (neon_vuzp<mode>_internal): Likewise.
9057 (neon_vld1<mode>): Likewise.
9058 (neon_vld1_lane<mode>): Likewise.
9059 (neon_vld1_dup<mode>): Likewise.
9060 (neon_vld1_dupv2di): Likewise.
9061 (neon_vst1<mode>): Likewise.
9062 (neon_vst1_lane<mode>): Likewise.
9063 (neon_vld2<mode>): Likewise.
9064 (neon_vld2_lane<mode>): Likewise.
9065 (neon_vld2_dup<mode>): Likewise.
9066 (neon_vst2<mode>): Likewise.
9067 (neon_vst2_lane<mode>): Likewise.
9068 (neon_vld3<mode>): Likewise.
9069 (neon_vld3qa<mode>): Likewise.
9070 (neon_vld3qb<mode>): Likewise.
9071 (neon_vld3_lane<mode>): Likewise.
9072 (neon_vld3_dup<mode>): Likewise.
9073 (neon_vst3<mode>): Likewise.
9074 (neon_vst3qa<mode>): Likewise.
9075 (neon_vst3qb<mode>): Likewise.
9076 (neon_vst3_lane<mode>): Likewise.
9077 (neon_vld4<mode>): Likewise.
9078 (neon_vld4qa<mode>): Likewise.
9079 (neon_vld4qb<mode>): Likewise.
9080 (neon_vld4_lane<mode>): Likewise.
9081 (neon_vld4_dup<mode>): Likewise.
9082 (neon_vst4<mode>): Likewise.
9083 (neon_vst4qa<mode>): Likewise.
9084 (neon_vst4qb<mode>): Likewise.
9085 (neon_vst4_lane<mode>): Likewise.
9086 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9087 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9088 (neon_vec_<US>mult_lo_<mode>): Likewise.
9089 (neon_vec_<US>mult_hi_<mode>): Likewise.
9090 (neon_vec_<US>shiftl_<mode>): Likewise.
9091 (neon_unpack<US>_<mode>): Likewise.
9092 (neon_vec_<US>mult_<mode>): Likewise.
9093 (vec_pack_trunc_<mode>): Likewise.
9094 (neon_vec_pack_trunk_<mode>): Likewise.
9095 (neon_vabd<mode>_2): Likewise.
9096 (neon_vabd<mode>_3): Likewise.
9097 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9098 (thumb2_movsi_vfp): Likewise.
9099 (movdi_vfp): Likewise.
9100 (movdi_vfp_cortexa8): Likewise.
9101 (movhf_vfp_neon): Likewise.
9102 (movhf_vfp): Likewiwse.
9103 (movsf_vfp): Likewiwse.
9104 (thumb2_movsf_vfp): Likewiwse.
9105 (movdf_vfp): Likewise.
9106 (thumb2_movdf_vfp): Likewise.
9107 (movsfcc_vfp): Likewise.
9108 (thumb2_movsfcc_vfp): Likewise.
9109 (movdfcc_vfp): Likewise.
9110 (thumb2_movdfcc_vfp): Likewise.
9111 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9112 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9113 (v10_v2c): Likewise.
9114 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9116 (cortex_a15_neon_int_2): Likewise.
9117 (cortex_a15_neon_int_3): Likewise.
9118 (cortex_a15_neon_int_4): Likewise.
9119 (cortex_a15_neon_int_5): Likewise.
9120 (cortex_a15_neon_vqneg_vqabs): Likewise.
9121 (cortex_a15_neon_vmov): Likewise.
9122 (cortex_a15_neon_vaba): Likewise.
9123 (cortex_a15_neon_vaba_qqq): Likewise.
9124 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9125 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9126 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9127 scalar_64_32_long_scalar): Likewise.
9128 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9129 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9130 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9131 lotype_qdd_64_32_long): Likewise.
9132 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9133 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9134 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9135 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9136 (cortex_a15_neon_shift_1): Likewise.
9137 (cortex_a15_neon_shift_2): Likewise.
9138 (cortex_a15_neon_shift_3): Likewise.
9139 (cortex_a15_neon_vshl_ddd): Likewise.
9140 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9141 (cortex_a15_neon_vsra_vrsra): Likewise.
9142 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9143 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9144 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9145 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9146 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9147 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9148 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9149 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9150 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9151 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9152 (cortex_a15_neon_bp_simple): Likewise.
9153 (cortex_a15_neon_bp_2cycle): Likewise.
9154 (cortex_a15_neon_bp_3cycle): Likewise.
9155 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9156 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9157 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9158 (cortex_a15_neon_vld2_4_regs): Likewise.
9159 (cortex_a15_neon_vld3_vld4): Likewise.
9160 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9161 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9162 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9163 (cortex_a15_neon_vst3_vst4): Likewise.
9164 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9165 (cortex_a15_neon_vld3_vld4_lane" 10
9166 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9167 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9168 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9169 (cortex_a15_neon_ldm_2): Likewise.0
9170 (cortex_a15_neon_stm_2): Likewise.
9171 (cortex_a15_neon_mcr): Likewise.
9172 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9173 (cortex_a15_neon_mrc): Likewise.
9174 (cortex_a15_neon_mrrc): Likewise.
9175 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9177 (cortex_a15_alu_shift): Likewise.
9178 (cortex_a15_alu_shift_reg): Likewise.
9179 (cortex_a15_mult32): Likewise.
9180 (cortex_a15_mult64): Likewise.
9181 (cortex_a15_block): Likewise.
9182 (cortex_a15_branch): Likewise.
9183 (cortex_a15_load1): Likewise.
9184 (cortex_a15_load3): Likewise.
9185 (cortex_a15_store1): Likewise.
9186 (cortex_a15_store3): Likewise.
9187 (cortex_a15_call): Likewise.
9188 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9189 (cortex_a5_f2r): Likewise.
9190 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9192 (cortex_a53_f2r): Likewise.
9193 * config/arm/cortex-a7.md
9194 (cortex_a7_branch): Update for attribute change.
9195 (cortex_a7_call): Likewise.
9196 (cortex_a7_alu_imm): Likewise.
9197 (cortex_a7_alu_reg): Likewise.
9198 (cortex_a7_alu_shift): Likewise.
9199 (cortex_a7_mul): Likewise.
9200 (cortex_a7_load1): Likewise.
9201 (cortex_a7_store1): Likewise.
9202 (cortex_a7_load2): Likewise.
9203 (cortex_a7_store2): Likewise.
9204 (cortex_a7_load3): Likewise.
9205 (cortex_a7_store3): Likewise.
9206 (cortex_a7_load4): Likewise.
9207 (cortex_a7_store4): Likewise.
9208 (cortex_a7_fpalu): Likewise.
9209 (cortex_a7_fconst): Likewise.
9210 (cortex_a7_fpmuls): Likewise.
9211 (cortex_a7_neon_mul): Likewise.
9212 (cortex_a7_fpmacs): Likewise.
9213 (cortex_a7_neon_mla: Likewise.
9214 (cortex_a7_fpmuld: Likewise.
9215 (cortex_a7_fpmacd: Likewise.
9216 (cortex_a7_fpfmad: Likewise.
9217 (cortex_a7_fdivs: Likewise.
9218 (cortex_a7_fdivd: Likewise.
9219 (cortex_a7_r2f: Likewise.
9220 (cortex_a7_f2r: Likewise.
9221 (cortex_a7_f_flags: Likewise.
9222 (cortex_a7_f_loads: Likewise.
9223 (cortex_a7_f_loadd: Likewise.
9224 (cortex_a7_f_stores: Likewise.
9225 (cortex_a7_f_stored: Likewise.
9226 (cortex_a7_neon): Likewise.
9227 * config/arm/cortex-a8-neon.md
9228 (cortex_a8_neon_mrc): Update for attribute change.
9229 (cortex_a8_neon_mrrc): Likewise.
9230 (cortex_a8_neon_int_1): Likewise.
9231 (cortex_a8_neon_int_2): Likewise.
9232 (cortex_a8_neon_int_3): Likewise.
9233 (cortex_a8_neon_int_4): Likewise.
9234 (cortex_a8_neon_int_5): Likewise.
9235 (cortex_a8_neon_vqneg_vqabs): Likewise.
9236 (cortex_a8_neon_vmov): Likewise.
9237 (cortex_a8_neon_vaba): Likewise.
9238 (cortex_a8_neon_vaba_qqq): Likewise.
9239 (cortex_a8_neon_vsma): Likewise.
9240 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9241 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9242 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9243 long_scalar): Likewise.
9244 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9245 (cortex_a8_neon_mla_qqq_8_16): Likewise.
9246 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9247 long_scalar_qdd_64_32_long): Likewise.
9248 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9249 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9250 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
9251 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9252 (cortex_a8_neon_shift_1): Likewise.
9253 (cortex_a8_neon_shift_2): Likewise.
9254 (cortex_a8_neon_shift_3): Likewise.
9255 (cortex_a8_neon_vshl_ddd): Likewise.
9256 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9257 (cortex_a8_neon_vsra_vrsra): Likewise.
9258 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
9259 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
9260 (cortex_a8_neon_fp_vsum): Likewise.
9261 (cortex_a8_neon_fp_vmul_ddd): Likewise.
9262 (cortex_a8_neon_fp_vmul_qqd): Likewise.
9263 (cortex_a8_neon_fp_vmla_ddd): Likewise.
9264 (cortex_a8_neon_fp_vmla_qqq): Likewise.
9265 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
9266 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
9267 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9268 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9269 (cortex_a8_neon_bp_simple): Likewise.
9270 (cortex_a8_neon_bp_2cycle): Likewise.
9271 (cortex_a8_neon_bp_3cycle): Likewise.
9272 (cortex_a8_neon_ldr): Likewise.
9273 (cortex_a8_neon_str): Likewise.
9274 (cortex_a8_neon_vld1_1_2_regs): Likewise.
9275 (cortex_a8_neon_vld1_3_4_regs): Likewise.
9276 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9277 (cortex_a8_neon_vld2_4_regs): Likewise.
9278 (cortex_a8_neon_vld3_vld4): Likewise.
9279 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9280 (cortex_a8_neon_vst1_3_4_regs): Likewise.
9281 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
9282 (cortex_a8_neon_vst3_vst4): Likewise.
9283 (cortex_a8_neon_vld1_vld2_lane): Likewise.
9284 (cortex_a8_neon_vld3_vld4_lane): Likewise.
9285 (cortex_a8_neon_vst1_vst2_lane): Likewise.
9286 (cortex_a8_neon_vst3_vst4_lane): Likewise.
9287 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
9288 (cortex_a8_neon_mcr): Likewise.
9289 (cortex_a8_neon_mcr_2_mcrr): Likewise.
9290 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9291 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
9293 (ca9_neon_mrrc): Likewise.
9294 (cortex_a9_neon_int_1): Likewise.
9295 (cortex_a9_neon_int_2): Likewise.
9296 (cortex_a9_neon_int_3): Likewise.
9297 (cortex_a9_neon_int_4): Likewise.
9298 (cortex_a9_neon_int_5): Likewise.
9299 (cortex_a9_neon_vqneg_vqabs): Likewise.
9300 (cortex_a9_neon_vmov): Likewise.
9301 (cortex_a9_neon_vaba): Likewise.
9302 (cortex_a9_neon_vaba_qqq): Likewise.
9303 (cortex_a9_neon_vsma): Likewise.
9304 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9305 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9306 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9307 long_scalar): Likewise.
9308 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9309 (cortex_a9_neon_mla_qqq_8_16): Likewise.
9310 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9311 long_scalar_qdd_64_32_long): Likewise.
9312 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9313 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9314 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
9315 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9316 (cortex_a9_neon_shift_1): Likewise.
9317 (cortex_a9_neon_shift_2): Likewise.
9318 (cortex_a9_neon_shift_3): Likewise.
9319 (cortex_a9_neon_vshl_ddd): Likewise.
9320 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9321 (cortex_a9_neon_vsra_vrsra): Likewise.
9322 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
9323 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
9324 (cortex_a9_neon_fp_vsum): Likewise.
9325 (cortex_a9_neon_fp_vmul_ddd): Likewise.
9326 (cortex_a9_neon_fp_vmul_qqd): Likewise.
9327 (cortex_a9_neon_fp_vmla_ddd): Likewise.
9328 (cortex_a9_neon_fp_vmla_qqq): Likewise.
9329 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
9330 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
9331 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9332 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9333 (cortex_a9_neon_bp_simple): Likewise.
9334 (cortex_a9_neon_bp_2cycle): Likewise.
9335 (cortex_a9_neon_bp_3cycle): Likewise.
9336 (cortex_a9_neon_ldr): Likewise.
9337 (cortex_a9_neon_str): Likewise.
9338 (cortex_a9_neon_vld1_1_2_regs): Likewise.
9339 (cortex_a9_neon_vld1_3_4_regs): Likewise.
9340 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9341 (cortex_a9_neon_vld2_4_regs): Likewise.
9342 (cortex_a9_neon_vld3_vld4): Likewise.
9343 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9344 (cortex_a9_neon_vst1_3_4_regs): Likewise.
9345 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
9346 (cortex_a9_neon_vst3_vst4): Likewise.
9347 (cortex_a9_neon_vld1_vld2_lane): Likewise.
9348 (cortex_a9_neon_vld3_vld4_lane): Likewise.
9349 (cortex_a9_neon_vst1_vst2_lane): Likewise.
9350 (cortex_a9_neon_vst3_vst4_lane): Likewise.
9351 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
9352 (cortex_a9_neon_mcr): Likewise.
9353 (cortex_a9_neon_mcr_2_mcrr): Likewise.
9354 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9355 (cortex_a9_fps): Likewise.
9356 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
9358 (cortex_m4_fmuls): Likewise.
9359 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
9361 (cortex_r4_mrc): Likewise.
9362 * config/arm/iterators.md: Update comment referring to neon_type.
9363 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
9364 (iwmmxt_movsi_insn): Likewise.
9365 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
9367 (pj4_core_to_vfp): Likewise.
9368 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
9370 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
9371 (vfp_fstore): Likewise.
9372 * doc/md.texi: Change references to neon_type to refer to type.
9374 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9376 * tree.h (DECL_BUILT_IN): Fix typo in comment.
9378 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
9380 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
9383 2013-09-04 Easwaran Raman <eraman@google.com>
9386 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
9387 (build_and_add_sum): Use it.
9388 (appears_later_in_bb): Simplify code.
9390 2013-09-04 Teresa Johnson <tejohnson@google.com>
9392 * dumpfile.c (dump_finish): Don't close stderr/stdout.
9394 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
9396 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
9398 2013-09-04 Jan Hubicka <jh@suse.cz>
9400 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
9401 * ipa-devirt.c: Include diganostic.h
9402 (odr_type_d): Add types and types_set.
9403 (hash_type_name): Work for types with vtables during LTO.
9404 (odr_hasher::remove): Fix comment; destroy types_set.
9405 (add_type_duplicate): New function,
9406 (get_odr_type): Use it.
9407 (dump_type_inheritance_graph): Dump type duplicates.
9408 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
9410 * tree.c (types_same_for_odr): Give exact answers on types with
9413 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9415 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
9416 explaining their differences.
9418 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
9420 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
9422 2013-09-03 Jeff Law <law@redhat.com>
9424 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
9425 when not threading through a joiner block. Pass joiner/no joiner
9426 state to register_jump_thread.
9427 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
9428 state from argument rather than implying on path length.
9429 Dump the entire jump thread path into debugging dump.
9430 * tree-flow.h (register_jump_thread): Update prototype.
9432 2013-08-29 Xinliang David Li <davidxl@google.com>
9434 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
9435 Remove a trivial gcc_assert.
9437 2013-08-29 Xinliang David Li <davidxl@google.com>
9439 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
9440 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
9441 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
9442 Delay base decl alignment adjustment.
9443 * tree-vectorizer.c (vect_destroy_datarefs): New function.
9444 * tree-vectorizer.h: New data structure.
9445 (set_dr_misalignment): New function.
9446 (dr_misalignment): Ditto.
9447 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
9448 (vectorizable_load): Ditto.
9449 (ensure_base_align): New function.
9450 (vectorize_loops): Add dbg_cnt support.
9451 (execute_vect_slp): Ditto.
9452 * dbgcnt.def: New debug counter.
9453 * Makefile: New dependency.
9455 2013-09-03 Meador Inge <meadori@codesourcery.com>
9459 2013-08-30 Meador Inge <meadori@codesourcery.com>
9461 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
9463 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
9465 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
9466 function descriptor.
9468 2013-09-03 Richard Biener <rguenther@suse.de>
9470 * tree-affine.c (add_elt_to_tree): Fix association issue,
9471 avoid useless converts and make sure to always return a
9472 properly typed result.
9474 2013-09-03 Richard Biener <rguenther@suse.de>
9477 * fold-const.c (negate_expr_p): Fix division case.
9478 (negate_expr): Likewise.
9480 2013-09-03 Richard Biener <rguenther@suse.de>
9483 * tree-streamer-out.c: Include tm.h.
9484 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
9486 2013-09-03 Jan Hubicka <jh@suse.cz>
9488 * tree-profile.c (tree_profiling): Cleanup CFG when done.
9490 2013-09-03 Alan Modra <amodra@gmail.com>
9492 * config.gcc (powerpc*-*-linux*): Add support for little-endian
9493 multilibs to big-endian target and vice versa.
9494 * config/rs6000/t-linux64: Use := assignment on all vars.
9495 (MULTILIB_EXTRA_OPTS): Remove fPIC.
9496 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
9497 * config/rs6000/t-linux64le: New file.
9498 * config/rs6000/t-linux64bele: New file.
9499 * config/rs6000/t-linux64lebe: New file.
9501 2013-09-02 Jan Hubicka <jh@suse.cz>
9503 * ipa-inline-transform.c (inline_transform): Do not
9504 optimize_inline_calls when not optimizing.
9506 2013-09-02 Jan Hubicka <jh@suse.cz>
9508 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
9509 duplicated nodes for assembler names.
9510 * symtab.c (symtab_unregister_node): Do not attempt to unlink
9511 hard registers from assembler name hash.
9513 2013-09-02 Jan Hubicka <jh@suse.cz>
9515 * ipa-split.c (execute_split_functions): Split externally visible
9516 functions called once.
9518 2013-09-02 Martin Jambor <mjambor@suse.cz>
9521 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
9522 linked list. When finding the correct duplicate, also consider also
9523 the caller in additon to its inlined_to node.
9525 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
9527 * config/aarch64/aarch64-simd-builtins.def
9528 (dup_lane_scalar): Remove.
9529 * config/aarch64/aarch64-simd.md
9530 (aarch64_simd_dup): Add 'w->w' alternative.
9531 (aarch64_dup_lane<mode>): Allow for VALL.
9532 (aarch64_dup_lane_scalar<mode>): Remove.
9533 (aarch64_dup_lane_<vswap_width_name><mode>): New.
9534 (aarch64_get_lane_signed<mode>): Add w->w altenative.
9535 (aarch64_get_lane_unsigned<mode>): Likewise.
9536 (aarch64_get_lane<mode>): Likewise.
9537 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
9538 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
9539 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
9540 (VCON): Change container of V2SF.
9541 (vswap_width_name): Likewise.
9542 * config/aarch64/arm_neon.h
9543 (__aarch64_vdup_lane_any): New.
9544 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
9545 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
9546 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
9548 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
9551 * expr.c (emit_move_complex): Do not move complex FP values as parts if
9552 the source or the destination is a single hard register.
9554 2013-09-02 Richard Biener <rguenther@suse.de>
9557 * tree-scalar-evolution.c (instantiate_scev_name): Allow
9560 2013-09-02 Richard Biener <rguenther@suse.de>
9562 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
9563 arithmetic to sizetype.
9565 2013-09-02 Bin Cheng <bin.cheng@arm.com>
9567 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
9568 Find auto-increment use both before and after candidate.
9570 2013-09-02 Marek Polacek <polacek@redhat.com>
9572 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
9574 2013-09-01 Jan Hubicka <jh@suse.cz>
9576 * Makefile.in: Add ipa-profile.o
9577 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
9578 * cgraph.c (struct cgraph_propagate_frequency_data,
9579 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
9580 ipa-profile.c; replace cgraph_ by ipa_ prefix.
9581 * cgraph.h (cgraph_propagate_frequency): Remove.
9582 * ipa-inline-analysis.c: Include ipa-utils.h;
9583 drop duplicated cfgloop.h.
9584 (inline_update_callee_summaries): Update.
9585 * ipa-profile.c: New file.
9586 * ipa-utils.h (ipa_propagate_frequency): Declare.
9587 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
9588 data-streamer.h, value-prof.h.
9589 (symtab_remove_unreachable_nodes): Update profile.
9590 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
9591 account_time_size, cmp_counts, dump_histogram,
9592 ipa_profile_generate_summary, ipa_profile_write_summary,
9593 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
9594 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
9595 Move to ipa-profile.c.
9597 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
9599 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
9601 2013-09-01 Jan Hubicka <jh@suse.cz>
9603 * common.opt (fdevirtualize-speculatively): New function.
9604 * invoke.texi (fdevirtualize-speculatively): Document.
9605 * ipa-devirt.c: Include ipa-inline.h
9606 (likely_target_p): New function.
9607 (ipa_devirt): New function.
9608 (gate_ipa_devirt): New function.
9609 (pass_data_ipa_devirt): New static var.
9610 (pass_ipa_devirt): Likewise.
9611 (make_pass_ipa_devirt): New function.
9612 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
9613 (common_handle_option): Disable devirtualization when
9614 value range profiling is available.
9615 * passes.def (pass_ipa_devirt): Add.
9616 * timever.def (TV_IPA_DEVIRT): New timevar.
9617 * tree-pass.h (make_pass_ipa_devirt):
9619 2013-09-01 Iain Sandoe <iain@codesourcery.com>
9621 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
9622 include sanitize(undefined).
9624 2013-08-31 Diego Novillo <dnovillo@google.com>
9626 * Makefile.in (TREE_CORE_H): Define.
9628 (GTFILES): Add tree-core.h.
9629 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
9631 * tree-core.h: New file.
9632 Move all data structures, enum, typedefs, global
9633 declarations and constants from ...
9636 2013-08-31 Jan Hubicka <jh@suse.cz>
9638 * bulitins.c (expand_builtin): Do not early exit for gcov
9639 instrumented functions.
9641 2013-08-31 Marek Polacek <polacek@redhat.com>
9643 * ubsan.c: Include tm_p.h.
9645 2013-08-31 Jan Hubicka <jh@suse.cz>
9647 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
9650 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
9651 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
9652 * tree-cfg.c (verify_gimple_label): ... here.
9653 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
9654 (ipa_merge_profiles): New function.
9655 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
9656 (lto_input_function_body): Likewise.
9657 * ipa-utils.h (ipa_merge_profiles): Declare.
9658 * lto-streamer.h (lto_input_function_body): Update prototype.
9659 (emit_label_in_global_context_p): Remove.
9660 * lto-symtab.c: Include ipa-utils.h
9661 (lto_cgraph_replace_node): Use ipa_merge_profiles.
9663 2013-08-31 Jan Hubicka <jh@suse.cz>
9665 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
9667 2013-08-31 Jan Hubicka <jh@suse.cz>
9669 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
9671 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
9673 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
9674 "cmp" RTX before signed_comparison_operator check to account
9677 2013-08-30 Jan Hubicka <jh@suse.cz>
9679 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
9680 (detect_type_change_1): Rename to ...
9681 (detect_type_change): ... this one; early return on non-polymorphic
9683 (detect_type_change_ssa): Add comp_type parameter; update
9684 use of detect_type_change.
9685 (compute_complex_assign_jump_func): Add param_type parameter;
9686 update use of detect_type_change_ssa.
9687 (compute_complex_ancestor_jump_func): Likewise.
9688 (ipa_get_callee_param_type): New function.
9689 (ipa_compute_jump_functions_for_edge): Compute parameter type;
9690 update calls to the jump function computation functions.
9692 2013-08-30 Teresa Johnson <tejohnson@google.com>
9693 Steven Bosscher <steven@gcc.gnu.org>
9695 * cfgrtl.c (fixup_new_cold_bb): New routine.
9696 (commit_edge_insertions): Invoke fixup_partitions.
9697 (find_partition_fixes): New routine.
9698 (fixup_partitions): Ditto.
9699 (verify_hot_cold_block_grouping): Update comments.
9700 (rtl_verify_edges): Invoke find_partition_fixes.
9701 (rtl_verify_bb_pointers): Update comments.
9702 (rtl_verify_bb_layout): Ditto.
9703 * basic-block.h (probably_never_executed_edge_p): Declare.
9704 (fixup_partitions): Ditto.
9705 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
9706 * bb-reorder.c (sanitize_hot_paths): New function.
9707 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
9709 * predict.c (probably_never_executed_edge_p): New routine.
9710 * cfg.c (check_bb_profile): Add partition insanity warnings.
9712 2013-08-30 Meador Inge <meadori@codesourcery.com>
9714 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
9716 2013-08-30 Marek Polacek <polacek@redhat.com>
9718 * Makefile.in (ubsan.o): Add.
9719 (c-family/c-ubsan.o): Add.
9720 (builtins.o): Add ubsan.h dependency.
9721 * ubsan.h: New file.
9722 * ubsan.c: New file.
9723 * common.opt: Add -fsanitize=undefined option.
9724 (flag_sanitize): Add variable.
9725 (fsanitize=): Add option. Add Driver.
9726 (fsanitize=thread): Remove option.
9727 (fsanitize=address): Likewise.
9728 (static-libubsan): New option.
9729 * doc/invoke.texi: Document the new flag and -static-libubsan.
9730 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
9731 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
9732 * builtin-attrs.def (ATTR_COLD): Define.
9733 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
9734 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
9735 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
9736 * flag-types.h (sanitize_code): New enum.
9737 * opts.c (common_handle_option): Parse command line arguments
9738 of -fsanitize=. Add -fsanitize=unreachable option.
9739 * varasm.c (get_variable_section): Adjust.
9740 (assemble_noswitch_variable): Likewise.
9741 (assemble_variable): Likewise.
9742 (output_constant_def_contents): Likewise.
9743 (categorize_decl_for_section): Likewise.
9744 (place_block_symbol): Likewise.
9745 (output_object_block): Likewise.
9746 * builtins.def: Likewise.
9747 * toplev.c (compile_file): Likewise.
9748 (process_options): Likewise.
9749 * cppbuiltin.c: Likewise.
9750 * tsan.c (tsan_pass): Likewise.
9751 (tsan_gate): Likewise.
9752 (tsan_gate_O0): Likewise.
9753 * cfgexpand.c (partition_stack_vars): Likewise.
9754 (expand_stack_vars): Likewise.
9755 (defer_stack_allocation): Likewise.
9756 (expand_used_vars): Likewise.
9757 * cfgcleanup.c (old_insns_match_p): Likewise.
9758 * asan.c (asan_finish_file): Likewise.
9759 (asan_instrument): Likewise.
9760 (gate_asan): Likewise.
9761 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
9762 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
9763 (asan_global_struct): Use pointer_sized_int_node instead
9764 calling build_nonstandard_integer_type.
9765 (initialize_sanitizer_builtins): Likewise.
9766 (asan_finish_file): Likewise.
9767 * gcc.c: Document %{%:function(args):X}.
9768 (static_spec_functions): Add sanitize.
9769 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
9770 store funcval != NULL there.
9771 (do_spec_1): Adjust handle_spec_function caller.
9772 (handle_braces): Allow %:function(args) as condition.
9773 (sanitize_spec_function): New function.
9774 (ADD_STATIC_LIBUBSAN_LIBS): Define.
9775 (LIBUBSAN_SPEC): Likewise.
9776 (LIBUBSAN_EARLY_SPEC): Likewise.
9777 (SANITIZER_SPEC): Handle libubsan.
9778 (SANITIZER_EARLY_SPEC): Likewise.
9779 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
9780 instead of fsanitize=address.
9781 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
9782 instead of fsanitize=address*.
9783 * builtins.c: Include ubsan.h.
9784 (fold_builtin_0): Instrument __builtin_unreachable.
9785 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
9786 instead of flag_asan.
9787 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
9788 (pointer_sized_int_node): Define.
9789 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
9791 2013-08-30 Mike Stump <mikestump@comcast.net>
9793 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
9796 2013-08-29 Jan Hubicka <jh@suse.cz>
9798 * cgraph.c (cgraph_function_body_availability): Handle weakref
9800 * passes.def: Remove pass_fixup_cfg.
9801 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
9802 track when we need to remove functions.
9803 (gate_ipa_inline): Execute inlining always; add comment why.
9804 (pass_data_ipa_inline): Remove TODO_remove_functions.
9805 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
9806 do not produce summaries.
9807 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
9808 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
9809 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
9810 constant pool and vtable.
9812 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
9814 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
9815 New arm_neon.h's internal macros to specify 64-bit constants.
9816 Avoid using stdint.h's macros.
9818 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
9820 * recog.c (verify_changes): Verify that changes[i].old is non-zero
9821 before applying REG_P.
9823 2013-08-30 Jakub Jelinek <jakub@redhat.com>
9825 PR tree-optimization/58277
9826 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
9827 after seeing too many stmts with vdef in between dombb and current
9828 bb, invalidate everything.
9830 2013-08-30 Richard Biener <rguenther@suse.de>
9832 * fold-const.c (fold_single_bit_test): Fix overflow test.
9834 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
9836 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
9837 and which can live in a register, always retrieve the value on entry.
9838 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
9839 passed by invisible reference specially.
9840 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
9841 (vt_add_function_parameter): Correctly deal with a parameter passed by
9842 invisible reference.
9844 2013-08-30 Jan Hubicka <jh@suse.cz>
9846 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
9848 2013-08-30 Richard Biener <rguenther@suse.de>
9850 PR tree-optimization/58228
9851 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
9852 allow invariant loads in nested loop vectorization.
9854 2013-08-30 Richard Biener <rguenther@suse.de>
9856 PR tree-optimization/58223
9857 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
9858 (has_anti_or_output_dependence): ... this and adjust to also
9859 look for output dependences.
9860 (mark_nodes_having_upstream_mem_writes): Adjust.
9861 (rdg_flag_uses): Likewise.
9863 2013-08-30 Richard Biener <rguenther@suse.de>
9865 PR tree-optimization/58010
9866 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
9867 assert that we have a loop-closed PHI.
9869 2013-08-29 Jan Hubicka <jh@suse.cz>
9871 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
9872 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
9873 * lto-section-in.c (lto_free_function_in_decl_state): New function.
9874 (lto_free_function_in_decl_state_for_node): New function.
9876 2013-08-29 Xinliang David Li <davidxl@google.com>
9878 * loop-unroll.c (report_unroll_peel): Minor message change.
9879 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
9880 Emit alignment peeling message with default -fopt-info.
9881 (vect_loop_versioning): Emit loop version info message.
9882 * tree-vectorizer.c (vectorize_loops): Minor message change.
9883 (execute_vect_slp): Ditto.
9885 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
9887 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
9888 of the clone from the DECL_NAME of the original function.
9890 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
9892 * passes.c (register_pass): Add overload.
9893 * tree-pass.h (register_pass): Forward declare it. Add comment.
9895 2013-08-29 Jan Hubicka <jh@suse.cz>
9897 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
9898 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
9899 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
9900 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
9901 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
9902 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
9903 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
9904 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
9906 2013-08-29 Teresa Johnson <tejohnson@google.com>
9908 * dumpfile.c (dump_loc): Output column number.
9909 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
9910 * doc/invoke.texi: Document optall -fopt-info flag.
9911 * profile.c (read_profile_edge_counts): Use new dump framework.
9912 (compute_branch_probabilities): Ditto.
9913 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
9914 when pass not in any opt group.
9915 * pass_manager.h (pass_manager::get_pass_profile): New method.
9916 * value-prof.c (check_counter): Use new dump framework.
9917 (check_ic_target): Ditto.
9918 * coverage.c (get_coverage_counts): Ditto.
9919 (coverage_init): Setup new dump framework.
9921 2013-08-29 Richard Biener <rguenther@suse.de>
9923 PR tree-optimization/58246
9924 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
9925 handle the dominance check inside a basic-block.
9927 2013-08-29 Richard Biener <rguenther@suse.de>
9930 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
9931 of default defs that appear in abnormal PHI nodes.
9933 2013-08-29 Richard Biener <rguenther@suse.de>
9935 PR tree-optimization/57685
9936 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
9937 single-use operands to avoid exponential complexity.
9939 2013-08-28 Dehao Chen <dehao@google.com>
9941 * ipa-inline.c (edge_badness): Fix integer underflow.
9943 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
9945 * gtm-builtins.def (_ITM_free): Declare leaf.
9947 2013-08-28 Jakub Jelinek <jakub@redhat.com>
9950 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
9951 (*tls_local_dynamic_base_64_largepic): Likewise.
9952 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
9953 Remove predicate from call operand.
9954 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
9955 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
9957 2013-08-28 Jeff Law <law@redhat.com>
9959 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
9960 checks for the number of predecessors and successors allowed.
9961 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
9962 which require copying a joiner block if there is a request which
9963 is a subpath that requires no joiner block copying.
9965 2013-08-28 Jan Hubicka <jh@suse.cz>
9967 * lto-streamer-out.c (DFS_write_tree_body): Drop
9968 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
9969 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
9970 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
9971 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
9972 Do not read DECL_ERROR_ISSUED.
9973 (unpack_ts_decl_with_vis_value_fields): Do not read
9975 (lto_input_ts_binfo_tree_pointers): Do not read
9976 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
9977 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
9978 write DECL_ERROR_ISSUED..
9979 (pack_ts_decl_with_vis_value_fields): Do not write
9981 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
9982 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
9983 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
9984 * tree.h (tree_decl_common): Update comment.
9985 (DECL_ERROR_ISSUED): Remove.
9987 2013-08-28 Jakub Jelinek <jakub@redhat.com>
9990 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
9992 2013-08-28 Jan Hubicka <jh@suse.cz>
9994 * builtins.def (free): Declare leaf.
9996 2013-08-27 David Malcolm <dmalcolm@redhat.com>
9999 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10000 * configure: Regenerate.
10002 2013-08-27 Martin Jambor <mjambor@suse.cz>
10004 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10005 (ipa_ancestor_jf_data): Likewise.
10006 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10007 (ipa_get_jf_pass_through_type_preserved): New function.
10008 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10009 (ipa_get_jf_ancestor_type_preserved): New function.
10010 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10011 (ipa_get_jf_ancestor_result): Likewise.
10012 (propagate_vals_accross_pass_through): Use
10013 ipa_get_jf_pass_through_result to do all the value mappings.
10014 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10015 type_preserved flag.
10016 (ipa_set_jf_cst_copy): New function.
10017 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10018 (ipa_set_jf_arith_pass_through): Likewise.
10019 (ipa_set_ancestor_jf): Likewise.
10020 (compute_complex_assign_jump_func): Set type_preserved instead of
10022 (ipa_compute_jump_functions_for_edge): Likewise.
10023 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10024 (update_jump_functions_after_inlining): Update type_preserved.
10025 Explicitely create jump functions when combining one with pass_through.
10026 (ipa_write_jump_function): Stream the type_preserved flags.
10027 (ipa_read_jump_function): Likewise.
10029 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10030 Aldy Hernandez <aldyh@redhat.com>
10032 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10033 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10034 * function.h (struct function): Add has_force_vect_loops and
10036 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10037 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10039 * gimple.def: Update CLAUSES comments.
10040 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10041 (gimple_build_omp_for): Add argument to prototype.
10042 (gimple_omp_for_kind): New.
10043 (gimple_omp_for_set_kind): New.
10044 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10045 GOVD_DATA_SHARE_CLASS.
10046 (enum omp_region_type): Add ORT_SIMD.
10047 (gimple_add_tmp_var): Handle ORT_SIMD.
10048 (gimplify_var_or_parm_decl): Same.
10049 (is_gimple_stmt): Same.
10050 (omp_firstprivatize_variable): Same.
10051 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10052 (omp_notice_variable): Handle ORT_SIMD.
10053 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10054 (omp_check_private): Handle ORT_SIMD.
10055 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10056 OMP_CLAUSE_SAFELEN.
10057 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10058 Handle OMP_CLAUSE_LASTPRIVATE.
10059 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10060 OMP_CLAUSE_SAFELEN.
10061 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10062 (gimplify_expr): Handle OMP_SIMD.
10063 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10064 (expand_GOMP_SIMD_VF): New.
10065 (expand_GOMP_SIMD_LAST_LANE): New.
10066 * internal-fn.def (GOMP_SIMD_LANE): New.
10067 (GOMP_SIMD_VF): New.
10068 (GOMP_SIMD_LAST_LANE): New.
10069 * omp-low.c: Include target.h.
10070 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10071 OMP_CLAUSE_SAFELEN.
10072 (check_omp_nesting_restrictions): Same.
10074 (lower_rec_simd_input_clauses): New.
10075 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10077 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10078 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10079 (expand_omp_build_assign): New.
10080 (expand_omp_for_init_counts): New.
10081 (expand_omp_for_init_vars): New.
10082 (extract_omp_for_update_vars): New.
10083 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10084 and rewrite accordingly.
10085 (expand_omp_simd): New.
10086 (expand_omp_for): Use expand_omp_simd.
10087 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10088 (lower_omp_for): Do not lower the body.
10089 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10090 in their own loops.
10091 * tree-flow.h (find_omp_clause): Remove prototype.
10092 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10093 forcing vectorization of the loop, or if flag_tree_vectorize.
10094 (gate_tree_if_conversion): Similarly.
10095 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10096 gimple_build_omp_for.
10097 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10098 * tree-parloops (create_parallel_loop): Pass kind argument to
10099 gimple_build_omp_for.
10100 * tree-pretty-print.c (dump_omp_clause): Add cases for
10101 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10102 OMP_CLAUSE__SIMDUID_.
10103 (dump_generic_node): Handle OMP_SIMD.
10104 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10105 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10106 unroll OMP_SIMD loops here.
10107 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10108 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10110 (vect_analyze_data_refs): Handle simd loops.
10111 * tree-vect-loop.c (vectorizable_live_operation): Handle
10113 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10114 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10115 (vectorizable_load): Same.
10116 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10117 (struct simduid_to_vf): New.
10118 (simduid_to_vf::hash): New.
10119 (simduid_to-vf::equal): New.
10120 (struct simd_array_to_simduid): New.
10121 (simd_array_to_simduid::hash): New.
10122 (simd_array_to_simduid::equal): New.
10123 (adjust_simduid_builtins): New.
10124 (struct note_simd_array_uses_struct): New.
10125 (note_simd_array_uses_cb): New.
10126 (note_simd_array_uses): New.
10127 (vectorize_loops): Handle simd hints and adjust simd builtins
10129 * tree-vectorizer.h (struct _stmt_vec_info): Add
10130 simd_lane_access_p field.
10131 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10132 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10133 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10134 (omp_clause_code_name): Same.
10135 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10136 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10137 * tree.def (OMP_SIMD): New entry.
10138 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10139 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10140 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10141 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10142 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10143 (OMP_CLAUSE_LINEAR_STEP): New.
10144 (OMP_CLAUSE_SAFELEN_EXPR): New.
10145 (OMP_CLAUSE__SIMDUID__DECL): New.
10146 (find_omp_clause): New prototype.
10148 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10150 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10151 Haswell processor detection.
10153 2013-08-27 Christian Widmer <shadow@umbrox.de>
10156 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10157 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10158 AVX2 capable processors.
10160 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10162 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10163 of vget_low_* with implementations in terms of other intrinsics.
10165 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10167 PR middle-end/57219
10168 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10169 values to -1, 0 and 1.
10171 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10173 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10174 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10175 (<optab><mode>3_insn): Remove.
10176 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10177 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10178 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10179 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10180 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10181 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10182 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10183 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10184 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10185 (ror<mode>3_insn): Likewise.
10186 * config/aarch64/predicates.md (aarch64_simd_register): New.
10188 2013-08-27 Richard Biener <rguenther@suse.de>
10190 PR tree-optimization/57521
10191 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10192 one edge is non-critical.
10193 (find_phi_replacement_condition): Make sure to use a non-critical
10194 edge. Cleanup and remove old bug workarounds.
10195 (bb_postdominates_preds): Remove.
10196 (if_convertible_loop_p_1): Do not compute post-dominators.
10197 (combine_blocks): Do not free post-dominators.
10198 (main_tree_if_conversion): Likewise.
10199 (pass_data_if_conversion): Add TODO_verify_ssa.
10201 2013-08-27 DJ Delorie <dj@redhat.com>
10203 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10205 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10207 * function.c (assign_parm_find_data_types): Set passed_mode and
10208 nominal_mode to the TYPE_MODE of nominal_type for the built
10209 pointer type in case of the struct-pass-by-reference.
10211 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10213 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10214 (UINT16_TYPE): Change default to "unsigned int".
10216 * config/avr/avr.opt (mfract-convert-truncate): New option.
10217 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
10218 is set, round negative fractional integers according to n1169
10219 when converting to integer types.
10221 2013-08-26 Jan Hubicka <jh@suse.cz>
10223 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
10224 methods can not be called indirectly when their address is not taken.
10226 2013-08-26 Jan Hubicka <jh@suse.cz>
10228 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
10231 2013-08-26 Jan Hubicka <jh@suse.cz>
10233 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
10236 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10238 * reload.c (find_valid_class): Allow classes that do not include
10239 FIRST_PSEUDO_REGISTER - 1.
10241 2013-08-26 Jan Hubicka <jh@suse.cz>
10243 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
10244 fix edge count/frequency when speculation failed; fix type check
10245 for the direct call.
10247 2013-08-26 Jan Hubicka <jh@suse.cz>
10249 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
10251 2013-08-26 Jan Hubicka <jh@suse.cz>
10253 * ipa-inline-transform.c (inline_transform): Be ready for basic block
10254 to be changed by edge redirection.
10256 2013-08-26 Jan Hubicka <jh@suse.cz>
10258 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
10259 formating; add sanity check.
10260 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
10261 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
10262 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
10263 (resolve_noninline_speculation): Update callee keys, too.
10265 2013-08-26 Jan Hubicka <jh@suse.cz>
10267 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
10268 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
10270 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10272 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
10275 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
10277 * config/i386/i386.c (ix86_debug_options): Remove prototype.
10278 (x86_64_elf_select_section): Ditto.
10279 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
10281 (ix86_pass_by_reference): Ditto.
10282 (output_set_got): Ditto.
10283 (ix86_unary_operator_ok): Ditto.
10284 (ix86_expand_builtin): Ditto.
10286 2013-08-23 Jan Hubicka <jh@suse.cz>
10288 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
10290 2013-08-23 Jan Hubicka <jh@suse.cz>
10292 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
10293 (tree_decl_with_vis): Add FINAL field.
10295 2013-08-23 Jeff Law <law@redhat.com>
10297 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
10298 the debugging dump when the expression is fully redundant.
10300 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10302 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
10303 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
10304 * pretty-print.c (pp_formatted_text_data): Likewise.
10305 (pp_write_text_to_stream): Likewise.
10306 (pp_write_text_as_dot_label_to_stream): Likewise.
10307 (pp_append_r): Likewise.
10308 (pp_format): Likewise.
10309 (pp_flush): Likewise.
10310 (pp_clear_output_area): Likewise.
10311 (pp_append_text): Likewise.
10312 (pp_formatted_text): Likewise.
10313 (pp_remaining_character_count_for_line): Likewise.
10314 (pp_newline): Likewise.
10315 (pp_character): Likewise.
10316 (output_buffer::~output_buffer): Define.
10317 (pretty_printer::~pretty_printer): Destruct output buffer.
10318 * pretty-print.h (output_buffer::~output_buffer): Declare.
10319 (pretty_printer::~pretty_printer): Declare virtual.
10321 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
10324 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
10325 HOST_WIDE_INT_M1U): New macros.
10326 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
10327 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
10328 unsigned -1 for lshift.
10329 * cse.c (cse_insn): Likewise.
10330 * double-int.c (rshift_double, lshift_double): Likewise.
10331 * builtins.c (fold_builtin_bitop): Likewise.
10332 * combine.c (force_to_mode): Likewise.
10333 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
10334 * simplify-rtx.c (simplify_const_unary_operation,
10335 simplify_const_binary_operation): Likewise.
10336 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
10337 check_va_list_escapes): Likewise.
10338 * rtlanal.c (nonzero_bits1): Likewise.
10339 * expmed.c (expand_smod_pow2): Likewise.
10340 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
10342 2013-08-23 Jan Hubicka <jh@suse.cz>
10344 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
10345 as having address taken.
10347 2013-08-23 Jan Hubicka <jh@suse.cz>
10349 * ipa-utils.h (method_class_type): Declare.
10350 * ipa-devirt.c (method_class_type): Export.
10352 * cgraphunit.c (analyze_functions): Do basic devirtualization;
10353 do not walk base classes of anonymous types.
10355 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
10357 PR rtl-optimization/58220
10358 PR regression/58221
10359 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
10360 handle SEQUENCE insns properly.
10362 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10364 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
10366 (pp_newline_and_indent): Likewise.
10367 (pp_separate_with): Likewise.
10368 * pretty-print.c (pp_newline_and_flush): Define.
10369 (pp_newline_and_indent): Likewise.
10370 (pp_separate_with): Likewise.
10372 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10375 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
10376 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
10378 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
10380 * config/i386/predicates.md (ext_sse_reg_operand): New.
10381 * config/i386/i386.md (*movti_internal): Use
10382 predicate to determine if EVEX is needed.
10383 (*movsi_internal): Ditto.
10384 (*movdf_internal): Ditto.
10385 (*movsf_internal): Ditto.
10386 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
10388 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10390 PR tree-optimization/58209
10391 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
10392 (find_tail_calls): Give up for pointer result types if m is non-NULL.
10393 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
10394 emit POINTER_PLUS_EXPR.
10395 (create_tailcall_accumulator): For pointer result type accumulate in
10398 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
10400 * configure.ac: Add backslashes missing from the last change.
10401 * configure: Regenerate.
10403 2013-08-22 Jan Hubicka <jh@susue.cz>
10405 * ipa.c (function_and_variable_visibility): First remember function
10406 was global and then make it local.
10408 2013-08-22 Julian Brown <julian@codesourcery.com>
10410 * configure.ac: Add aarch64 to list of arches which use "nop" in
10412 * configure: Regenerate.
10414 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10416 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
10417 gnu_libc_has_function.
10418 * config/s390/tpf.h: Likewise.
10420 2013-08-22 Jan Hubicka <jh@susue.cz>
10422 * timevar.c (validate_phases): Add cast.
10424 2013-08-22 Jan Hubicka <jh@susue.cz>
10426 * timevar.c (validate_phases): Use size_t for memory.
10427 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
10429 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
10431 * pretty-print.h (output_buffer::output_buffer): Declare.
10432 (pretty_printer::pretty_printer): Likewise.
10433 (pp_construct): Remove.
10434 * pretty-print.c (output_buffer::output_buffer): Define.
10435 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
10436 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
10437 (print_gimple_expr): Likewise.
10438 (print_gimple_seq): Likewise.
10439 (gimple_dump_bb): Likewise.
10440 * sched-vis.c (dump_value_slim): Likewise.
10441 (dump_insn_slim): Likewise.
10442 (dump_rtl_slim): Likewise.
10443 (str_pattern_slim): Likewise.
10444 * tree-mudflap.c (mf_varname_tree): Likewise.
10445 * graph.c (print_graph_cfg): Likewise.
10446 (start_graph_dump): Likewise.
10447 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
10449 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
10450 pretty printer initialization.
10451 * coretypes.h (diagnostic_context): Remove superflous type alias
10453 (pretty_printer): Likewise. Declare directly as a class.
10454 (pretty_print_info): Remove declaration as class.
10455 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
10456 and pp_clear_output_area.
10457 (asan_add_global): Likewise.
10459 2013-08-22 Jan Hubicka <jh@suse.cz>
10461 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
10462 * ipa-utils.h (update_type_inheritance_graph): Declare.
10463 (possible_polymorphic_call_target_p): Declare.
10464 (possible_polymorphic_call_target_p): New.
10465 * ipa-devirt.c: Update toplevel comments.
10466 (cached_polymorphic_call_targets): Move up.
10467 (odr_type_d): Move ID down.
10468 (polymorphic_type_binfo_p): Update comment.
10469 (odr_hasher::remove): Likewise;
10470 (get_odr_type): Set anonymous_namespace.
10471 (dump_odr_type): Dump it.
10472 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
10473 (maybe_record_node): Record node in cached_polymorphic_call_targets.
10474 (record_binfo): Add comment.
10475 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
10477 (devirt_node_removal_hook): Do not iCE when cache is freed.
10478 (possible_polymorphic_call_target_p): New predicate.
10479 (update_type_inheritance_graph): New function.
10481 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
10482 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
10483 Sergey Lega <sergey.s.lega@intel.com>
10484 Anna Tikhonova <anna.tikhonova@intel.com>
10485 Ilya Tocar <ilya.tocar@intel.com>
10486 Andrey Turetskiy <andrey.turetskiy@intel.com>
10487 Ilya Verbin <ilya.verbin@intel.com>
10488 Kirill Yukhin <kirill.yukhin@intel.com>
10489 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10491 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
10492 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
10493 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
10494 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
10495 (OPTION_MASK_ISA_AVX2_UNSET): Update.
10496 (OPTION_MASK_ISA_AVX512F_UNSET): New.
10497 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
10498 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
10499 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
10500 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
10501 OPT_mavx512pf, OPT_mavx512er cases.
10502 * config/i386/constraints.md (v): New constraint.
10503 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
10504 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
10505 (bit_AVX512CD): New.
10506 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
10507 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
10508 * config/i386/i386-c.c (ix86_target_macros_internal):
10509 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
10511 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
10512 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
10513 * config/i386/i386.c (regclass_map, dbx_register_map)
10514 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
10515 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
10516 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
10517 -mavx512pf options.
10518 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
10519 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
10520 -mavx512pf options. Fix formatting.
10521 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
10522 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
10523 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
10524 -mavx512cd, -mavx512pf options.
10525 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
10526 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
10527 (ix86_preferred_output_reload_class): Replace SSE_REGS with
10529 (ix86_hard_regno_mode_ok): Support 512-bit registers.
10530 (ix86_set_reg_reg_cost): Ditto.
10531 (x86_order_regs_for_local_alloc): Ditto.
10532 (MAX_VECT_LEN): Extend to 64-byte.
10533 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
10534 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
10535 (TARGET_AVX512ER, TARGET_AVX512CD): New.
10536 (BIGGEST_ALIGNMENT): Extend to 512-bits.
10537 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
10538 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
10539 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
10540 (SSE_REG_MODE_P): Support new modes.
10541 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
10542 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
10543 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
10544 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
10545 (REG_CLASS_CONTENTS): Add new registers.
10546 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
10547 (EXT_REX_SSE_REGNO_P): New.
10548 (HI_REGISTER_NAMES): Add new registers.
10549 * config/i386/i386.md: Define constants for new registers.
10550 (mode): Add new 512-bit modes.
10551 (prefix): Support evex prefix.
10552 (isa): Support avx512f, noavx512f, fma_avx512f.
10553 (ssemodesuffix): Add new 512-bit modes.
10555 (*movxi_internal_avx512f): Ditto.
10556 (*movdi_internal): Replace constraint "x" with the new constraint "v".
10558 (*movsi_internal): Likewise.
10559 (*movdf_internal): Likewise.
10560 (*movsf_internal): Likewise.
10561 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
10562 (<code><mode>3): Likewise.
10563 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
10565 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
10566 with the new constraint "v".
10567 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
10569 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
10570 with the new constraint "v".
10571 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
10572 (<sse2>_storedqu<avxsizesuffix>): Likewise.
10573 (*<plusminus_insn><mode>3): Likewise.
10574 (<sse>_vm<plusminus_insn><mode>3): Likewise.
10575 (*mul<mode>3): Likewise.
10576 (<sse>_vmmul<mode>3): Likewise.
10577 (<sse>_div<mode>3): Likewise.
10578 (<sse>_vmdiv<mode>3): Likewise.
10579 (<sse>_sqrt<mode>2): Likewise.
10580 (<sse>_vmsqrt<mode>2): Likewise.
10581 (*<code><mode>3_finite): Likewise.
10582 (*<code><mode>3) <smaxmin>: Likewise.
10583 (<sse>_vm<code><mode>3): Likewise.
10584 (*<code><mode>3) <any_logic>: Likewise.
10585 (*fma_fmadd_<mode>): Likewise.
10586 (*fma_fmsub_<mode>): Likewise.
10587 (*fma_fnmadd_<mode>): Likewise.
10588 (*fma_fnmsub_<mode>): Likewise.
10589 (*fma_fmaddsub_<mode>): Likewise.
10590 (*fma_fmsubadd_<mode>): Likewise.
10591 (*fmai_fmadd_<mode>): Likewise.
10592 (*fmai_fmsub_<mode>): Likewise.
10593 (*fmai_fnmadd_<mode>): Likewise.
10594 (*fmai_fnmsub_<mode>): Likewise.
10595 (sse_cvtsi2ss): Likewise.
10596 (sse_cvtsi2ssq): Likewise.
10597 (sse_cvtss2si): Likewise.
10598 (sse_cvtss2si_2): Likewise.
10599 (sse_cvtss2siq): Likewise.
10600 (sse_cvtss2siq_2): Likewise.
10601 (sse_cvttss2si): Likewise.
10602 (sse_cvtss2siq_2): Likewise.
10603 (float<sseintvecmodelower><mode>2): Likewise.
10604 (sse2_cvtsd2si_2): Likewise.
10605 (sse2_cvtsd2siq_2): Likewise.
10606 (*<plusminus_insn><mode>3): Likewise.
10607 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
10608 (*<sse4_1_avx2>_mul<mode>3): Likewise.
10609 (ashr<mode>3): Likewise.
10610 (<shift_insn><mode>3): Likewise.
10611 (avx2_<code><mode>3): Likewise.
10612 (*avx2_<code><mode>3): Likewise.
10613 (*andnot<mode>3): Likewise.
10614 (*<code><mode>3) <any_logic>: Likewise.
10615 (abs<mode>2): Likewise.
10616 (avx2_permvar<mode>): Likewise.
10617 (avx2_perm<mode>_1): Likewise.
10618 (*avx_vpermilp<mode>): Likewise.
10619 (avx_vpermilvar<mode>3): Likewise.
10620 (avx2_ashrv<mode>): Likewise.
10621 (avx2_<shift_insn>v<mode>): Likewise.
10622 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
10624 * doc/rtl.texi: Document XImode.
10626 2013-08-21 Jeff Law <law@redhat.com>
10628 * tree-flow.h (register_jump_thread): Pass vector of edges
10629 instead of each important edge.
10630 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
10631 thread path into a vector and pass that to register_jump_thread.
10632 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
10633 passed in edge vector to the current 3-edge form.
10636 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
10638 * dce.c (fini_dce): Call df_analyze again just in case
10639 delete_unmarked_insns removed anything.
10641 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
10643 * reload.h (struct reg_equivs): Rename to ..
10644 (struct reg_equivs_s): .. this.
10646 2013-08-20 Martin Liska <marxin.liska@gmail.com>
10648 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
10650 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10652 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
10654 2013-08-21 Jeff Law <law@redhat.com>
10656 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
10657 simplify assignments too. If the RHS collapses to a singleton
10658 range, then return the value for the range.
10660 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
10662 * config/i386/sse.md (V16): Rename to...
10664 (mov<mode>): Update iterator name.
10665 (*mov<mode>_internal): Ditto.
10666 (push<mode>1): Ditto.
10667 (movmisalign<mode>): Ditto.
10669 2013-08-20 Jan Hubicka <jh@suse.cz>
10672 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
10673 entry for direct edges.
10674 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
10676 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10678 Revert my last two changes, r201865 and r201864:
10681 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10683 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
10684 instances can own GC refs.
10686 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
10687 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
10688 (gcc::context::gt_pch_nx): Likewise.
10689 (gcc::context::gt_pch_nx): Likewise.
10690 * ggc.h (gt_ggc_mx <T>): New.
10691 (gt_pch_nx_with_op <T>): New.
10692 (gt_pch_nx <T>): New.
10693 * passes.c (opt_pass::gt_ggc_mx): New.
10694 (opt_pass::gt_pch_nx): New.
10695 (opt_pass::gt_pch_nx_with_op): New.
10696 (pass_manager::gt_ggc_mx): New.
10697 (pass_manager::gt_pch_nx): New.
10698 (pass_manager::gt_pch_nx_with_op): New.
10699 (pass_manager::operator new): Use
10700 ggc_internal_cleared_alloc_stat rather than xcalloc.
10701 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
10702 (pass_manager::gt_ggc_mx): New.
10703 (pass_manager::gt_pch_nx): New.
10704 (pass_manager::gt_pch_nx_with_op): New.
10705 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
10706 (opt_pass::operator new): New.
10707 (opt_pass::gt_ggc_mx): New.
10708 (opt_pass::gt_pch_nx): New.
10709 (opt_pass::gt_pch_nx_with_op): New.
10712 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10714 * Makefile.in (GTFILES): Add context.h.
10715 * context.c (gcc::context::operator new): New.
10716 (gcc::context::gt_ggc_mx): New.
10717 (gcc::context::gt_pch_nx): New.
10718 (gcc::context::gt_pch_nx): New.
10719 * context.h (gcc::context): Add GTY((user)) marking.
10720 (gcc::context::operator new): New.
10721 (gcc::context::gt_ggc_mx): New.
10722 (gcc::context::gt_pch_nx): New.
10723 (gcc::context::gt_pch_nx): New.
10724 (g): Add GTY marking.
10725 (gt_ggc_mx (gcc::context *)): New.
10726 (gt_pch_nx (gcc::context *)): New.
10727 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
10728 void *cookie)): New.
10729 * gengtype.c (open_base_files) <ifiles>: Add context.h.
10731 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
10733 * dce.c (fini_dce): Call df_analyze again just in case
10734 delete_unmarked_insns removed anything.
10736 2013-08-20 Teresa Johnson <tejohnson@google.com>
10738 PR rtl-optimizations/57451
10739 * final.c (reemit_insn_block_notes): Prevent lexical blocks
10740 from crossing split section boundaries.
10742 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10744 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
10745 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
10746 with MULTLIB_DEFAULTS.
10748 2013-08-20 Nick Clifton <nickc@redhat.com>
10750 * target.def (narrow_volatile_bitfield): Note that the default
10751 value is false, not !TARGET_STRICT_ALIGN.
10752 * doc/tm.texi: Regenerate.
10754 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
10756 Fix LIB_SPEC for systems without libpthread.
10758 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
10759 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
10761 * config/i386/linux-common.h: Likewise.
10762 * config/mips/linux-common.h: Likewise.
10764 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
10766 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
10769 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10771 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
10772 instances can own GC refs.
10774 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
10775 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
10776 (gcc::context::gt_pch_nx): Likewise.
10777 (gcc::context::gt_pch_nx): Likewise.
10778 * ggc.h (gt_ggc_mx <T>): New.
10779 (gt_pch_nx_with_op <T>): New.
10780 (gt_pch_nx <T>): New.
10781 * passes.c (opt_pass::gt_ggc_mx): New.
10782 (opt_pass::gt_pch_nx): New.
10783 (opt_pass::gt_pch_nx_with_op): New.
10784 (pass_manager::gt_ggc_mx): New.
10785 (pass_manager::gt_pch_nx): New.
10786 (pass_manager::gt_pch_nx_with_op): New.
10787 (pass_manager::operator new): Use
10788 ggc_internal_cleared_alloc_stat rather than xcalloc.
10789 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
10790 (pass_manager::gt_ggc_mx): New.
10791 (pass_manager::gt_pch_nx): New.
10792 (pass_manager::gt_pch_nx_with_op): New.
10793 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
10794 (opt_pass::operator new): New.
10795 (opt_pass::gt_ggc_mx): New.
10796 (opt_pass::gt_pch_nx): New.
10797 (opt_pass::gt_pch_nx_with_op): New.
10799 2013-08-20 David Malcolm <dmalcolm@redhat.com>
10801 * Makefile.in (GTFILES): Add context.h.
10802 * context.c (gcc::context::operator new): New.
10803 (gcc::context::gt_ggc_mx): New.
10804 (gcc::context::gt_pch_nx): New.
10805 (gcc::context::gt_pch_nx): New.
10806 * context.h (gcc::context): Add GTY((user)) marking.
10807 (gcc::context::operator new): New.
10808 (gcc::context::gt_ggc_mx): New.
10809 (gcc::context::gt_pch_nx): New.
10810 (gcc::context::gt_pch_nx): New.
10811 (g): Add GTY marking.
10812 (gt_ggc_mx (gcc::context *)): New.
10813 (gt_pch_nx (gcc::context *)): New.
10814 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
10815 void *cookie)): New.
10816 * gengtype.c (open_base_files) <ifiles>: Add context.h.
10818 2013-08-20 Alan Modra <amodra@gmail.com>
10821 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
10822 (rs6000_emit_epilogue): Likewise.
10824 2013-08-19 Dehao Chen <dehao@google.com>
10826 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
10828 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
10829 Jakub Jelinek <jakub@redhat.com>
10831 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
10832 (BUILT_IN_FABSD64): Likewise.
10833 (BUILT_IN_FABSD128): Likewise.
10834 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
10835 (fold_builtin_1): Likewise.
10836 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
10837 destination and source operands.
10838 (*abstd2_fpr): Likewise.
10839 (*nabstd2_fpr): Likewise.
10841 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
10843 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
10846 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
10848 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
10851 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
10853 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
10854 * builtins.c (default_libc_has_function): New.
10855 (gnu_libc_has_function): Ditto.
10856 (no_c99_libc_has_function): Ditto.
10857 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
10858 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
10859 (fold_builtin_sincos): Likewise.
10860 (fold_builtin_cexp): Likewise.
10861 * builtins.def (DEF_C94_BUILTIN): Likewise.
10862 (DEF_C99_BUILTIN): Likewise.
10863 (DEF_C99_C90RES_BUILTIN): Likewise.
10864 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
10865 definitions to using this define.
10866 * config/darwin-protos.h (darwin_libc_has_function): New.
10867 * config/darwin.c (darwin_libc_has_function): Ditto.
10868 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
10869 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
10870 * config/darwin.h: Ditto.
10871 * config/elfos.h: Ditto.
10872 * config/freebsd.h: Ditto.
10873 * config/i386/cygming.h: Ditto.
10874 * config/i386/djgpp.h: Ditto.
10875 * config/i386/i386-interix.h: Ditto.
10876 * config/microblaze/microblaze.h: Ditto.
10877 * config/mmix/mmix.h: Ditto.
10878 * config/gnu-user.h: Ditto.
10879 * config/ia64/hpux.h: Ditto.
10880 * config/pa/pa-hpux.h: Ditto.
10881 * config/pdp11/pdp11.h: Ditto.
10882 * config/picochip/picochip.h: Ditto.
10883 * config/linux.h: Ditto.
10884 * config/netbsd.h: Ditto.
10885 * config/openbsd.h: Ditto.
10886 * config/rs6000/aix43.h: Ditto.
10887 * config/rs6000/aix51.h: Ditto.
10888 * config/rs6000/aix52.h: Ditto.
10889 * config/rs6000/aix53.h: Ditto.
10890 * config/rs6000/aix61.h: Ditto.
10891 * config/rs6000/darwin.h: Ditto.
10892 * config/rs6000/linux.h: Ditto.
10893 * config/rs6000/linux64.h: Ditto.
10894 * config/s390/tpf.h: Ditto.
10895 * config/sol2-10.h: Ditto.
10896 * config/sol2.h: Ditto.
10897 * config/vms/vms.h: Ditto.
10898 * config/vxworks.h: Ditto.
10899 * config/linux-android.c (linux_android_libc_has_function):
10900 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
10901 * config/linux-protos.h (linux_android_libc_has_function):
10903 * config/i386/i386.c (ix86_libc_has_function): New.
10904 * config/i386/i386-protos.h
10905 (ix86_libc_has_function): New declaration.
10906 * config/i386/i386.md
10907 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
10908 ("isinf<mode>2): Likewise.
10909 * convert.c (convert_to_integer): Using new target hook
10910 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
10911 TARGET_C99_FUNCTIONS.
10912 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
10913 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
10914 * coretypes.h (function_class): New enum for different
10915 classes of functions.
10916 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
10917 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
10918 (TARGET_HAS_SINCOS): Likewise.
10919 (TARGET_LIBC_HAS_FUNCTION): New.
10920 * doc/tm.texi: Regenerated.
10921 * targhooks.h (default_libc_has_function): New declaration.
10922 (no_c99_libc_has_function): Ditto.
10923 (gnu_libc_has_function): Ditto.
10924 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
10925 and TARGET_HAS_SINCOS.
10927 2013-08-18 Jan Hubicka <jh@suse.cz>
10929 * Makeifle-in (ipa-devirt.o): New.
10930 (GTFILES): Add ipa-utils.h and ipa-devirt.c
10931 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
10932 (analyze_functions): Look into possible targets of polymorphic call.
10933 * dumpfile.c (dump_files): Add type-inheritance dump.
10934 * dumpfile.h (TDI_inheritance): New.
10935 * ipa-devirt.c: New file.
10936 * ipa-utils.h (odr_type_d): Forward declare.
10937 (odr_type): New type.
10938 (build_type_inheritance_graph): Declare.
10939 (possible_polymorphic_call_targets): Declare and introduce inline
10940 variant when only edge is pased.
10941 (dump_possible_polymorphic_call_targets): Likewise.
10942 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
10943 * tree.c (type_in_anonymous_namespace_p): Break out from ...
10944 (types_same_for_odr): ... here.
10945 * tree.h (type_in_anonymous_namespace_p): Declare.
10947 2013-08-18 Jakub Jelinek <jakub@redhat.com>
10949 PR tree-optimization/58006
10950 * tree-parloops.c (take_address_of): Don't ICE if get_name
10952 (eliminate_local_variables_stmt): Remove clobber stmts.
10954 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
10956 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
10957 error message is issued for an alias to undefined symbol.
10959 2013-08-18 Jan Hubicka <jh@suse.cz>
10961 * cgraph.c (cgraph_create_indirect_edge): Discover
10962 polymorphic calls and record basic info into indirect_info.
10963 * gimple-fold.c (gimple_fold_call): When doing BINFO based
10964 devirtualization, ignore objc function calls.
10965 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
10966 call with no parm index info.
10967 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
10968 * tree.c (virtual_method_call_p): New function.
10969 * tree.h (virtual_method_call_p): Declare.
10971 2013-08-16 Jan Hubicka <jh@suse.cz>
10973 PR middle-end/58179
10974 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
10976 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
10978 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
10981 2013-08-16 David Malcolm <dmalcolm@redhat.com>
10983 * gengtype.c (type_for_name): Add special-case support for
10984 locating types within the "gcc::" namespace.
10985 (open_base_files): Emit a "using namespace gcc" directive.
10987 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10990 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
10991 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
10993 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
10994 array instead of each individual operand as a separate argument.
10995 (emit_fusion_gpr_load): Likewise.
10996 (expand_fusion_gpr_load): Add new function declaration.
10998 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
10999 signature to have the operands passed as an array, instead of as
11000 separate arguments. Allow ZERO_EXTEND to be in the memory
11001 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11002 depend on the register live/dead flags when peepholes are run.
11003 (expand_fusion_gpr_load): New function to be called from the
11004 peephole2 pass, to change the register that addis sets to be the
11006 (emit_fusion_gpr_load): Change the calling signature to have the
11007 operands passed as an array, instead of as separate arguments.
11008 Allow ZERO_EXTEND to be in the memory address, and also
11009 SIGN_EXTEND if -mpower8-fusion-sign.
11011 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11012 unspec enumeration.
11013 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11014 adjust the register addis loads up in the peephole2 pass. Do not
11015 depend on the register live/dead state when the peephole pass is done.
11017 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11019 * gengtype.c (create_user_defined_type): Ensure that the kind
11020 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11021 declaration is seen before the GTY((user)) marking.
11023 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11026 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11028 2013-08-16 Jan Hubicka <jh@suse.cz>
11030 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11031 arugment expected_type.
11032 (gimple_fold_call): Use it.
11033 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11034 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11035 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11036 (try_make_edge_direct_virtual_call): Likewise.
11037 * tree.c (obj_type_ref_class): New.
11038 * tree.h (obj_type_ref_class): Use it.
11040 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11042 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11043 (rtl_slim_pp): Likewise.
11044 (init_rtl_slim_pretty_print): Likewise.
11045 (dump_value_slim): Don't call it. Use local pretty printer.
11046 (dump_insn_slim): Likewise.
11047 (dump_rtl_slim): Likewise.
11048 (str_pattern_slim): Likewise.
11049 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11052 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11054 PR tree-optimization/58164
11055 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11056 walk gimple_goto_dest of GIMPLE_GOTO.
11058 PR tree-optimization/58165
11059 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11060 bi_call must be the last stmt in a bb, don't split_block, instead
11061 use fallthru edge from it and give up if there is none.
11062 Release conds vector when returning early.
11064 2013-08-14 Xinliang David Li <davidxl@google.com>
11066 * config/i386/i386.c (ix86_option_override_internal):
11067 Remove unused variable and field.
11069 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11072 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11073 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11074 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11075 and Linux, correct BLKmode alignment when 128-bit alignment is
11076 required and compatibility flag is not set.
11077 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11078 for zero-size arguments when compatibility flag is not set.
11080 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11082 PR tree-optimization/58145
11083 * tree-sra.c (build_ref_for_offset): If prev_base has
11084 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11086 2013-08-14 Xinliang David Li <davidxl@google.com>
11088 * config/i386/i386.c (ix86_option_override_internal):
11089 Fix uninitialized variable error.
11091 2013-08-14 Xinliang David Li <davidxl@google.com>
11093 * config/i386/i386.opt: Define two new options.
11094 * config/i386/x86-tune.def: Add arch selector field in macros.
11095 * config/i386/i386.h: Adjust macro definition.
11096 * config/i386/i386.c (ix86_option_override_internal):
11098 (parse_mtune_ctrl_str): New function.
11099 (set_ix86_tune_features): New function.
11100 (ix86_function_specific_restore): Call the new helper function.
11102 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11104 PR rtl-optimization/57662
11105 * sel-sched.c (code_motion_process_successors): When the current insn
11106 is removed after the recursive traversal, break from the loop.
11107 Add comments and debug printouts.
11109 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11110 Alexandre Oliva <aoliva@redhat.com>
11113 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11114 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11115 there also UNSPEC_PLTOFF.
11117 2013-08-14 Marek Polacek <polacek@redhat.com>
11119 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11120 NUM_CONDITIONS bit positions.
11122 2013-08-13 Cary Coutant <ccoutant@google.com>
11124 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11125 (attr_checksum): Hash vector contents instead of pointer.
11126 (attr_checksum_ordered): Likewise.
11128 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11130 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11131 when Pmode != word_mode. Add length_address attribute.
11132 (sse3_monitor_<mode>): Merge from sse3_monitor and
11133 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11134 Pmode != word_mode. Update insn length attribute.
11135 * config/i386/i386.c (ix86_option_override_internal): Update
11136 ix86_gen_monitor selection for merged sse3_monitor insn.
11138 2013-08-13 Julian Brown <julian@codesourcery.com>
11140 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11141 perform invalid legitimization on greater-than-word-size modes for
11142 TARGET_E500_DOUBLE.
11144 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11146 * ira.c (setup_class_translate_array): Use aclass instead of cl
11147 for classes not fully covered by allocno classes.
11149 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11151 PR tree-optimization/57661
11152 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11153 * tree-inline.c (tree_function_versioning): Initialize it.
11154 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11155 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11156 that is not being copied.
11159 * asan.c (instrument_strlen_call): Fix typo in comment.
11160 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11162 2013-08-13 Steve Ellcey <sellcey@mips.com>
11164 * config/mips/mips.md (prefetch): Use lw instead of ld on
11165 loongson in 32bit mode.
11167 2013-08-13 Nick Clifton <nickc@redhat.com>
11169 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11171 2013-08-13 Jan Hubicka <jh@suse.cz>
11173 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11174 introduced edge; fix typo in sanity check.
11175 (cgraph_resolve_speculation): Export; improve diagnostic.
11176 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11177 speculation at type mismatch.
11178 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11179 (cgraph_resolve_speculation): Declare.
11180 (symtab_can_be_discarded): New function.
11181 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11182 * ipa-inline-transform.c (speculation_removed): New global var.
11183 (clone_inlined_nodes): See if speculation can be removed.
11184 (inline_call): If speculations was removed, we growths may not match.
11185 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11186 (speculation_useful_p): New function.
11187 (resolve_noninline_speculation): New function.
11188 (inline_small_functions): Resolve useless speculations.
11189 * ipa-inline.h (speculation_useful_p): Declare
11190 * ipa.c (can_replace_by_local_alias): Simplify.
11191 (ipa_profile): Produce speculative calls in non-lto, too;
11192 add simple cost model; produce local aliases.
11194 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11196 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11199 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11201 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11203 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11205 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11206 iterator instead of X87MODEF.
11208 2013-08-12 Perez Read <netfirewall@gmail.com>
11211 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11212 operand 0 for intel asm alternative.
11213 (*movabs<mode>_2): Ditto for operand 1.
11215 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
11217 * config/aarch64/arm_none.h
11218 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
11220 2013-08-12 Nick Clifton <nickc@redhat.com>
11222 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
11224 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
11226 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
11227 expand for QI/HImode operand to produce more effictive code for
11228 unsigned char(short) --> float(double) conversion.
11230 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
11232 * doc/invoke.texi: Mention that -ftls-model does not force the final
11235 2013-08-12 Marek Polacek <polacek@redhat.com>
11236 Marc Glisse <marc.glisse@inria.fr>
11238 PR tree-optimization/57980
11239 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
11240 when creating -1 constant.
11242 2013-08-10 Jan Hubicka <jh@suse.cz>
11244 Workaround binutils PR14342.
11245 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
11246 (gimple_init_edge_profiler): Likewise.
11247 (gimple_gen_ic_func_profiler): Likewise.
11249 2013-08-09 Jan Hubicka <jh@suse.cz>
11251 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
11253 2013-08-09 Xinliang David Li <davidxl@google.com>
11255 * config/i386/stringop.def: New file.
11256 * config/i386/stringop.opt: New file.
11257 * config/i386/i386-opts.h: Include stringopt.def.
11258 * config/i386/i386.opt: Include stringopt.opt.
11259 * config/i386/i386.c (ix86_option_override_internal):
11260 Override default size based stringop inline strategies with options.
11261 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
11264 2013-08-09 Jan Hubicka <jh@suse.cz>
11266 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
11268 2013-08-09 Jan Hubicka <jh@suse.cz>
11270 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
11272 (dump_cgraph_node): Dump profile-id.
11273 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
11274 and common_target_probability.
11275 * lto-cgraph.c (lto_output_edge): Stream common targets.
11276 (lto_output_node): Stream profile ids.
11277 (input_node): Stream profile ids.
11278 (input_edge): Stream common targets.
11279 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
11280 * ipa.c: Include value-prof.h
11281 (ipa_profile_generate_summary): Turn indirect call statement histograms
11282 into common targets.
11283 (ipa_profile): Turn common targets into speculative edges.
11285 2013-08-09 Jan Hubicka <jh@suse.cz>
11287 * cgraph.h (cgraph_node): Add profile_id.
11288 * value-prof.c (cgraph_node_map): Turn into pointer_map.
11289 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
11290 (del_node_map): Update.
11291 (find_func_by_funcdef_no): Replace by ...
11292 (find_func_by_profile_id): ... this one.
11293 (gimple_ic_transform): Do not remove useful histograms when
11294 speculation is not done; dump info when indirect call removal
11296 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
11297 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
11298 (__gcov_indirect_call_profiler_v2): .. this one.
11299 * profile.h (init_node_map): Update.
11300 * coverage.c (coverage_compute_profile_id): New function.
11301 * coverage.h (coverage_compute_profile_id): Declare.
11302 * tree-profile.c (init_ic_make_global_vars): Make
11303 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
11304 (gimple_init_edge_profiler): Update prototype of
11305 __gcov_indirect_call_profiler.
11306 (gimple_gen_ic_func_profiler): Simplify.
11307 (tree_profiling): Use init_node_map
11309 2013-08-09 Jan Hubicka <jh@suse.cz>
11311 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
11312 non-speculative refs.
11313 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
11314 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
11315 (cgraph_set_call_stmt): Likewise.
11316 (cgraph_create_edge_1): Fix release checking compilatoin;
11317 clear lto_stmt_uid.
11318 (cgraph_free_edge): Free indirect info.
11319 (cgraph_turn_edge_to_speculative): New function.
11320 (cgraph_speculative_call_info): New function.
11321 (cgraph_make_edge_direct): Return direct edge; handle speculation.
11322 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
11323 (dump_cgraph_node): Dump speculation.
11324 (verify_edge_count_and_frequency): Accept speculative edges.
11325 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
11326 (verify_cgraph_node): Handle speculation.
11327 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
11328 (cgraph_set_call_stmt): Update prototype.
11329 (cgraph_make_edge_direct): Update prototype.
11330 (cgraph_speculative_call_info): Declare.
11331 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
11332 to change; update call of ipa_find_references.
11333 * ipa-ref.c (ipa_record_reference): Fix return value; clear
11334 lto_stmt_uid and speculative flags.
11335 (ipa_dump_references): Dump speculation.
11336 (ipa_clone_references): Clone speculative flag.
11337 (ipa_clone_referring): Likewise.
11338 (ipa_clone_ref): New function.
11339 (ipa_find_reference): Look into lto_stmt_uids
11340 (ipa_clear_stmts_in_references): Do not clear speculative calls.
11341 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
11342 (ipa_find_reference): Update declaration.
11343 (ipa_clone_ref): Declare.
11344 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
11345 stream speculative flag.
11346 (lto_output_ref): Stream statements uids and speculation.
11347 (input_ref): Likewise.
11348 (input_edge): Stream speuclation.
11349 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
11350 (cgraph_set_call_stmt_including_clones): Handle speculation.
11351 * ipa-inline.c (heap_edge_removal_hook): New function.
11352 (inline_small_functions): Register it.
11353 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
11354 also initialize refs.
11355 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
11357 (try_make_edge_direct_simple_call): Likewise.
11358 (try_make_edge_direct_simple_call): Likewise.
11359 (update_indirect_edges_after_inlining): Likewise.
11360 (remove_described_reference): Look proper lto_stmt_uid.
11361 (propagate_controlled_uses): Likewise.
11362 (propagate_controlled_uses): Liekwise.
11363 * tree-inline.c (copy_bb): Copy speculative edges.
11364 (redirect_all_calls): New function.
11365 (copy_cfg_body): Do redirection after loop info is updated.
11366 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
11368 2013-08-09 Jan Hubicka <jh@suse.cz>
11370 * lto-streamer-out.c (output_function): Renumber PHIs.
11371 * lto-streamer-in.c (input_function): Likewise.
11373 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
11375 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
11376 (get_lane_unsigned): Likewise.
11377 (dup_lane_scalar): Likewise.
11378 (get_lane): enable for VALL.
11379 * config/aarch64/aarch64-simd.md
11380 (aarch64_dup_lane_scalar<mode>): Remove.
11381 (aarch64_get_lane_signed<mode>): Likewise.
11382 (aarch64_get_lane_unsigned<mode>): Likewise.
11383 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
11384 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
11385 (aarch64_get_lane<mode>): Enable for all vector modes.
11386 (aarch64_get_lanedi): Remove misleading constraints.
11387 * config/aarch64/arm_neon.h
11388 (__aarch64_vget_lane_any): Define.
11389 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
11390 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
11391 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
11392 * config/aarch64/iterators.md (VDQQH): New.
11393 (VDQQHS): Likewise.
11394 (vwcore): Likewise.
11396 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
11398 * configure.ac: Add GAS check for LEON instructions on SPARC.
11399 * configure: Regenerate.
11400 * config.in: Likewise.
11401 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
11403 * config/sparc/sparc.opt (LEON, LEON3): New masks.
11404 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
11406 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
11407 (AS_LEON_FLAG): New macro.
11408 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
11409 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
11410 Deal with LEON and LEON3 for the memory model.
11411 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
11412 (atomic_compare_and_swap<mode>_1): Likewise.
11413 (*atomic_compare_and_swap<mode>_1): Likewise.
11415 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11417 * config/arm/neon.md (vcond): Fix floating-point vector
11418 comparisons against 0.
11420 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
11422 * lra-constraints.c (emit_spill_move): Remove assert.
11423 (process_alt_operands): Add more debugging
11424 output. Increase reject for spilling into memory. Decrease
11425 reject for reloading scratch.
11426 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
11428 2013-08-08 Steve Ellcey <sellcey@mips.com>
11430 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
11431 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
11432 micromips incompatible. Add nan2008.
11433 (MULTILIB_DIRNAMES): Add nan2008.
11434 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
11435 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
11436 and micromips incompatible. Add nan2008.
11437 (MULTILIB_DIRNAMES): Add nan2008.
11438 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
11440 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
11442 PR rtl-optimization/58079
11443 * combine.c (combine_simplify_rtx): Avoid using SUBST if
11444 simplify_comparison has widened a comparison with an integer.
11446 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11448 * config/arm/neon.md (movmisalign<mode>): Disable when we
11449 don't allow unaligned accesses.
11450 (*movmisalign<mode>_neon_store): Likewise.
11451 (*movmisalign<mode>_neon_load): Likewise.
11452 (*movmisalign<mode>_neon_store): Likewise.
11453 (*movmisalign<mode>_neon_load): Likewise.
11455 2013-08-08 Jan Hubicka <jh@suse.cz>
11457 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
11458 (make_pass_rebuild_cgraph_edges): Also clear references.
11459 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
11460 * ipa-inline-transform.c (inline_transform): Remove all references
11462 * cgraphunit.c (expand_function): Remove all references after
11464 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
11465 (ipa_find_reference): Rewrite to iterator.
11466 (remove_stmt_references): Likewise.
11467 (ipa_clear_stmts_in_references): New function.
11468 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
11469 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
11471 * ipa-split.c (split_function): Remove references in split function.
11473 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
11476 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
11477 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
11479 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
11482 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
11483 suggested mode for the assignment isn't compatible with the
11484 registers required.
11486 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11489 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
11491 2013-08-07 Xinliang David Li <davidxl@google.com>
11493 * config/i386/i386.opt: New option -mtune-ctrl=.
11494 * config/i386/x86-tune.def: New file.
11495 * config/i386/i386.h: include x86-tune.def.
11496 * config/i386/i386.c (ix86_option_override_internal):
11497 Parsing -mtune-ctrl= option and set tune features.
11499 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
11502 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
11503 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
11506 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
11508 * rtl.h (update_alignments): Declare.
11509 * final.c (grow_label_align): New function extracted from...
11510 (shorten_branches): ...here. Call it.
11511 (update_alignments): New function.
11512 * reorg.c (sibling_labels): New variable.
11513 (get_label_before): Add SIBLING parameter. If it is non-zero, push
11514 the new label along with it onto the sibling_labels vector.
11515 (fill_simple_delay_slots): Adjust call to get_label_before.
11516 (fill_slots_from_thread): Likewise.
11517 (relax_delay_slots): Likewise.
11518 (make_return_insns): Likewise.
11519 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
11521 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
11523 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
11524 document its semantics.
11525 (diagnostic_report_diagnostic): Adjust accordingly.
11527 2013-08-07 David Malcolm <dmalcolm@redhat.com>
11529 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
11530 (sparc_option_override): ...and port to new C++ pass API.
11531 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
11533 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
11535 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
11537 2013-08-06 Caroline Tice <cmtice@google.com>
11539 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
11540 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
11541 * tree-pass.h: Add pass_vtable_verify.
11542 * varasm.c (assemble_variable): Add code to properly set the comdat
11543 section and name for the .vtable_map_vars section.
11544 (assemble_vtyv_preinit_initializer): New function.
11545 (default_sectin_type_flags): Make sure .vtable_map_vars section has
11547 * output.h: Add function decl for assemble_vtv_preinit_initializer.
11548 * vtable-verify.c: New file.
11549 * vtable-verify.h: New file.
11550 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
11551 initialiation levels.
11552 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
11553 * passes.def: Insert pass_vtable_verify.
11554 * aclocal.m4: Reorder includes.
11555 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
11556 -fvtv-counts options.
11557 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
11558 as appropriate, if -fvtable-verify=... is used.
11559 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
11560 -fvtable-verify=... is used.
11561 * Makefile.in (OBJS): Add vtable-verify.o to list.
11562 (vtable-verify.o): Add new build rule.
11563 (GTFILES): Add vtable-verify.c to list.
11564 * common.opt (fvtable-verify=): New flag.
11565 (vtv_priority): Values for fvtable-verify= flag.
11566 (fvtv-counts): New flag.
11567 (fvtv-debug): New flag.
11568 * tree.h (save_vtable_map_decl): New extern function decl.
11570 2013-08-07 David Malcolm <dmalcolm@redhat.com>
11572 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
11573 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
11574 (pass_data_rl78_devirt): ...new pass_data instance and...
11575 (make_pass_rl78_devirt): ...new function.
11576 (rl78_asm_file_start): Port pass registration to new C++ API.
11578 2013-08-07 David Malcolm <dmalcolm@redhat.com>
11580 * coretypes.h (rtl_opt_pass): Add.
11581 (gcc::context): Add.
11582 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
11583 (epiphany_init): Port to new C++ pass API.
11584 (epiphany_optimize_mode_switching): Likewise.
11585 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
11586 (pass_manager::get_pass_mode_switching): New.
11587 (pass_manager::get_pass_peephole2): New.
11588 * mode-switching.c (pass_mode_switching): Add clone method.
11589 * recog.c (pass_peephole2): Add clone method.
11590 (pass_split_all_insns): Add clone method.
11592 2013-08-06 David Malcolm <dmalcolm@redhat.com>
11594 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
11595 (mips_option_override): ...here, porting to new C++ API for passes.
11597 2013-08-06 Jan Hubicka <jh@suse.cz>
11599 * cgraph.c (cgraph_get_body): New function based on lto.c
11601 * cgraph.h (cgraph_get_body): Declare.
11602 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
11604 * cgraphunit.c (expand_function): Get body prior expanding.
11605 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
11606 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
11608 * passes.c (do_per_function_toporder): Get body.
11609 * tree-inline.c (expand_call_inline): Get body prior inlining it.
11610 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
11612 2013-08-06 Martin Jambor <mjambor@suse.cz>
11615 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
11616 is not re-finalized. Rename second parameter to no_collect.
11618 2013-08-06 Martin Jambor <mjambor@suse.cz>
11620 PR middle-end/58041
11621 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
11622 MEM_REF has proper alignment information.
11624 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
11627 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
11629 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
11630 rtx (*) (rtx, ...) with insn_gen_fn.
11631 * genoutput.c (output_insn_data): Cast gen_? function pointers to
11632 insn_gen_fn::stored_funcptr. Add initializer braces.
11634 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11636 Rewrite how instances of passes are cloned to remove assumptions
11637 about their sizes (thus allowing pass subclasses to have
11638 additional data fields, albeit non-GC-managed ones at this point).
11640 * passes.c (make_pass_instance): Now that passes have clone
11641 methods, rewrite this function to eliminate XNEW and memcpy
11642 calls that used hardcoded sizes. Since this function no longer
11643 creates pass instances, rename it to...
11644 (add_pass_instance): ...this. Document the old way that passes were
11645 numbered and flagged, and rework this function to continue using it.
11646 (next_pass_1): Add an initial_pass argument for use by
11648 (position_pass): When adding multiple instances of a pass, use
11649 the pass's clone method, rather than relying on the XNEW/memcpy
11650 within the former make_pass_instance (now add_pass_instance).
11651 (pass_manager::pass_manager): When invoking next_pass_1, also supply
11652 the initial instance of the current pass within the pass manager.
11654 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11656 This is the automated part of the conversion of passes from C
11657 structs to C++ classes.
11659 Patch autogenerated by refactor_passes.py from
11660 https://github.com/davidmalcolm/gcc-refactoring-scripts
11661 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
11663 * asan.c (pass_asan): Convert from a global struct to a subclass of
11664 gimple_opt_pass along with...
11665 (pass_data_asan): ...new pass_data instance and...
11666 (make_pass_asan): ...new function.
11667 (pass_asan_O0): Convert from a global struct to a subclass of
11668 gimple_opt_pass along with...
11669 (pass_data_asan_O0): ...new pass_data instance and...
11670 (make_pass_asan_O0): ...new function.
11671 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
11672 subclass of rtl_opt_pass along with...
11673 (pass_data_inc_dec): ...new pass_data instance and...
11674 (make_pass_inc_dec): ...new function.
11675 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
11676 a subclass of rtl_opt_pass along with...
11677 (pass_data_reorder_blocks): ...new pass_data instance and...
11678 (make_pass_reorder_blocks): ...new function.
11679 (pass_duplicate_computed_gotos): Convert from a global struct to a
11680 subclass of rtl_opt_pass along with...
11681 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
11682 (make_pass_duplicate_computed_gotos): ...new function.
11683 (pass_partition_blocks): Convert from a global struct to a subclass of
11684 rtl_opt_pass along with...
11685 (pass_data_partition_blocks): ...new pass_data instance and...
11686 (make_pass_partition_blocks): ...new function.
11687 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
11688 struct to a subclass of rtl_opt_pass along with...
11689 (pass_data_branch_target_load_optimize1): ...new pass_data instance
11691 (make_pass_branch_target_load_optimize1): ...new function.
11692 (pass_branch_target_load_optimize2): Convert from a global struct to a
11693 subclass of rtl_opt_pass along with...
11694 (pass_data_branch_target_load_optimize2): ...new pass_data instance
11696 (make_pass_branch_target_load_optimize2): ...new function.
11697 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
11698 of rtl_opt_pass along with...
11699 (pass_data_jump): ...new pass_data instance and...
11700 (make_pass_jump): ...new function.
11701 (pass_jump2): Convert from a global struct to a subclass of
11702 rtl_opt_pass along with...
11703 (pass_data_jump2): ...new pass_data instance and...
11704 (make_pass_jump2): ...new function.
11705 * cfgexpand.c (pass_expand): Convert from a global struct to a
11706 subclass of rtl_opt_pass along with...
11707 (pass_data_expand): ...new pass_data instance and...
11708 (make_pass_expand): ...new function.
11709 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
11710 of rtl_opt_pass along with...
11711 (pass_data_free_cfg): ...new pass_data instance and...
11712 (make_pass_free_cfg): ...new function.
11713 (pass_into_cfg_layout_mode): Convert from a global struct to a
11714 subclass of rtl_opt_pass along with...
11715 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
11716 (make_pass_into_cfg_layout_mode): ...new function.
11717 (pass_outof_cfg_layout_mode): Convert from a global struct to a
11718 subclass of rtl_opt_pass along with...
11719 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
11720 (make_pass_outof_cfg_layout_mode): ...new function.
11721 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
11722 struct to a subclass of gimple_opt_pass along with...
11723 (pass_data_build_cgraph_edges): ...new pass_data instance and...
11724 (make_pass_build_cgraph_edges): ...new function.
11725 (pass_rebuild_cgraph_edges): Convert from a global struct to a
11726 subclass of gimple_opt_pass along with...
11727 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
11728 (make_pass_rebuild_cgraph_edges): ...new function.
11729 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
11730 subclass of gimple_opt_pass along with...
11731 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
11733 (make_pass_remove_cgraph_callee_edges): ...new function.
11734 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
11735 struct to a subclass of rtl_opt_pass along with...
11736 (pass_data_stack_adjustments): ...new pass_data instance and...
11737 (make_pass_stack_adjustments): ...new function.
11738 * combine.c (pass_combine): Convert from a global struct to a subclass
11739 of rtl_opt_pass along with...
11740 (pass_data_combine): ...new pass_data instance and...
11741 (make_pass_combine): ...new function.
11742 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
11743 global struct to a subclass of rtl_opt_pass along with...
11744 (pass_data_compare_elim_after_reload): ...new pass_data instance
11746 (make_pass_compare_elim_after_reload): ...new function.
11747 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
11748 of rtl_opt_pass along with...
11749 (pass_data_rtl_cprop): ...new pass_data instance and...
11750 (make_pass_rtl_cprop): ...new function.
11751 * cse.c (pass_cse): Convert from a global struct to a subclass of
11752 rtl_opt_pass along with...
11753 (pass_data_cse): ...new pass_data instance and...
11754 (make_pass_cse): ...new function.
11755 (pass_cse2): Convert from a global struct to a subclass of
11756 rtl_opt_pass along with...
11757 (pass_data_cse2): ...new pass_data instance and...
11758 (make_pass_cse2): ...new function.
11759 (pass_cse_after_global_opts): Convert from a global struct to a
11760 subclass of rtl_opt_pass along with...
11761 (pass_data_cse_after_global_opts): ...new pass_data instance and...
11762 (make_pass_cse_after_global_opts): ...new function.
11763 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
11764 of rtl_opt_pass along with...
11765 (pass_data_ud_rtl_dce): ...new pass_data instance and...
11766 (make_pass_ud_rtl_dce): ...new function.
11767 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
11768 rtl_opt_pass along with...
11769 (pass_data_fast_rtl_dce): ...new pass_data instance and...
11770 (make_pass_fast_rtl_dce): ...new function.
11771 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
11772 a subclass of rtl_opt_pass along with...
11773 (pass_data_df_initialize_opt): ...new pass_data instance and...
11774 (make_pass_df_initialize_opt): ...new function.
11775 (pass_df_initialize_no_opt): Convert from a global struct to a
11776 subclass of rtl_opt_pass along with...
11777 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
11778 (make_pass_df_initialize_no_opt): ...new function.
11779 (pass_df_finish): Convert from a global struct to a subclass of
11780 rtl_opt_pass along with...
11781 (pass_data_df_finish): ...new pass_data instance and...
11782 (make_pass_df_finish): ...new function.
11783 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
11784 rtl_opt_pass along with...
11785 (pass_data_rtl_dse1): ...new pass_data instance and...
11786 (make_pass_rtl_dse1): ...new function.
11787 (pass_rtl_dse2): Convert from a global struct to a subclass of
11788 rtl_opt_pass along with...
11789 (pass_data_rtl_dse2): ...new pass_data instance and...
11790 (make_pass_rtl_dse2): ...new function.
11791 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
11792 subclass of rtl_opt_pass along with...
11793 (pass_data_dwarf2_frame): ...new pass_data instance and...
11794 (make_pass_dwarf2_frame): ...new function.
11795 * except.c (pass_set_nothrow_function_flags): Convert from a global
11796 struct to a subclass of rtl_opt_pass along with...
11797 (pass_data_set_nothrow_function_flags): ...new pass_data instance
11799 (make_pass_set_nothrow_function_flags): ...new function.
11800 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
11801 subclass of rtl_opt_pass along with...
11802 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
11804 (make_pass_convert_to_eh_region_ranges): ...new function.
11805 * final.c (pass_compute_alignments): Convert from a global struct to a
11806 subclass of rtl_opt_pass along with...
11807 (pass_data_compute_alignments): ...new pass_data instance and...
11808 (make_pass_compute_alignments): ...new function.
11809 (pass_final): Convert from a global struct to a subclass of
11810 rtl_opt_pass along with...
11811 (pass_data_final): ...new pass_data instance and...
11812 (make_pass_final): ...new function.
11813 (pass_shorten_branches): Convert from a global struct to a subclass of
11814 rtl_opt_pass along with...
11815 (pass_data_shorten_branches): ...new pass_data instance and...
11816 (make_pass_shorten_branches): ...new function.
11817 (pass_clean_state): Convert from a global struct to a subclass of
11818 rtl_opt_pass along with...
11819 (pass_data_clean_state): ...new pass_data instance and...
11820 (make_pass_clean_state): ...new function.
11821 * function.c (pass_instantiate_virtual_regs): Convert from a global
11822 struct to a subclass of rtl_opt_pass along with...
11823 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
11824 (make_pass_instantiate_virtual_regs): ...new function.
11825 (pass_leaf_regs): Convert from a global struct to a subclass of
11826 rtl_opt_pass along with...
11827 (pass_data_leaf_regs): ...new pass_data instance and...
11828 (make_pass_leaf_regs): ...new function.
11829 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
11830 subclass of rtl_opt_pass along with...
11831 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
11833 (make_pass_thread_prologue_and_epilogue): ...new function.
11834 (pass_match_asm_constraints): Convert from a global struct to a
11835 subclass of rtl_opt_pass along with...
11836 (pass_data_match_asm_constraints): ...new pass_data instance and...
11837 (make_pass_match_asm_constraints): ...new function.
11838 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
11839 subclass of rtl_opt_pass along with...
11840 (pass_data_rtl_fwprop): ...new pass_data instance and...
11841 (make_pass_rtl_fwprop): ...new function.
11842 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
11843 rtl_opt_pass along with...
11844 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
11845 (make_pass_rtl_fwprop_addr): ...new function.
11846 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
11847 rtl_opt_pass along with...
11848 (pass_data_rtl_pre): ...new pass_data instance and...
11849 (make_pass_rtl_pre): ...new function.
11850 (pass_rtl_hoist): Convert from a global struct to a subclass of
11851 rtl_opt_pass along with...
11852 (pass_data_rtl_hoist): ...new pass_data instance and...
11853 (make_pass_rtl_hoist): ...new function.
11854 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
11855 subclass of gimple_opt_pass along with...
11856 (pass_data_lower_cf): ...new pass_data instance and...
11857 (make_pass_lower_cf): ...new function.
11858 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
11859 from a global struct to a subclass of gimple_opt_pass along with...
11860 (pass_data_strength_reduction): ...new pass_data instance and...
11861 (make_pass_strength_reduction): ...new function.
11862 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
11863 of rtl_opt_pass along with...
11864 (pass_data_rtl_ifcvt): ...new pass_data instance and...
11865 (make_pass_rtl_ifcvt): ...new function.
11866 (pass_if_after_combine): Convert from a global struct to a subclass of
11867 rtl_opt_pass along with...
11868 (pass_data_if_after_combine): ...new pass_data instance and...
11869 (make_pass_if_after_combine): ...new function.
11870 (pass_if_after_reload): Convert from a global struct to a subclass of
11871 rtl_opt_pass along with...
11872 (pass_data_if_after_reload): ...new pass_data instance and...
11873 (make_pass_if_after_reload): ...new function.
11874 * init-regs.c (pass_initialize_regs): Convert from a global struct to
11875 a subclass of rtl_opt_pass along with...
11876 (pass_data_initialize_regs): ...new pass_data instance and...
11877 (make_pass_initialize_regs): ...new function.
11878 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
11879 of ipa_opt_pass_d along with...
11880 (pass_data_ipa_cp): ...new pass_data instance and...
11881 (make_pass_ipa_cp): ...new function.
11882 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
11883 global struct to a subclass of gimple_opt_pass along with...
11884 (pass_data_inline_parameters): ...new pass_data instance and...
11885 (make_pass_inline_parameters): ...new function.
11886 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
11887 subclass of gimple_opt_pass along with...
11888 (pass_data_early_inline): ...new pass_data instance and...
11889 (make_pass_early_inline): ...new function.
11890 (pass_ipa_inline): Convert from a global struct to a subclass of
11891 ipa_opt_pass_d along with...
11892 (pass_data_ipa_inline): ...new pass_data instance and...
11893 (make_pass_ipa_inline): ...new function.
11894 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
11895 struct to a subclass of gimple_opt_pass along with...
11896 (pass_data_local_pure_const): ...new pass_data instance and...
11897 (make_pass_local_pure_const): ...new function.
11898 (pass_ipa_pure_const): Convert from a global struct to a subclass of
11899 ipa_opt_pass_d along with...
11900 (pass_data_ipa_pure_const): ...new pass_data instance and...
11901 (make_pass_ipa_pure_const): ...new function.
11902 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
11903 to a subclass of ipa_opt_pass_d along with...
11904 (pass_data_ipa_reference): ...new pass_data instance and...
11905 (make_pass_ipa_reference): ...new function.
11906 * ipa-split.c (pass_split_functions): Convert from a global struct to
11907 a subclass of gimple_opt_pass along with...
11908 (pass_data_split_functions): ...new pass_data instance and...
11909 (make_pass_split_functions): ...new function.
11910 (pass_feedback_split_functions): Convert from a global struct to a
11911 subclass of gimple_opt_pass along with...
11912 (pass_data_feedback_split_functions): ...new pass_data instance and...
11913 (make_pass_feedback_split_functions): ...new function.
11914 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
11915 global struct to a subclass of simple_ipa_opt_pass along with...
11916 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
11918 (make_pass_ipa_function_and_variable_visibility): ...new function.
11919 (pass_ipa_free_inline_summary): Convert from a global struct to a
11920 subclass of simple_ipa_opt_pass along with...
11921 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
11922 (make_pass_ipa_free_inline_summary): ...new function.
11923 (pass_ipa_whole_program_visibility): Convert from a global struct to a
11924 subclass of ipa_opt_pass_d along with...
11925 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
11927 (make_pass_ipa_whole_program_visibility): ...new function.
11928 (pass_ipa_profile): Convert from a global struct to a subclass of
11929 ipa_opt_pass_d along with...
11930 (pass_data_ipa_profile): ...new pass_data instance and...
11931 (make_pass_ipa_profile): ...new function.
11932 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
11933 ipa_opt_pass_d along with...
11934 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
11935 (make_pass_ipa_cdtor_merge): ...new function.
11936 * ira.c (pass_ira): Convert from a global struct to a subclass of
11937 rtl_opt_pass along with...
11938 (pass_data_ira): ...new pass_data instance and...
11939 (make_pass_ira): ...new function.
11940 (pass_reload): Convert from a global struct to a subclass of
11941 rtl_opt_pass along with...
11942 (pass_data_reload): ...new pass_data instance and...
11943 (make_pass_reload): ...new function.
11944 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
11945 subclass of rtl_opt_pass along with...
11946 (pass_data_cleanup_barriers): ...new pass_data instance and...
11947 (make_pass_cleanup_barriers): ...new function.
11948 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
11949 of rtl_opt_pass along with...
11950 (pass_data_loop2): ...new pass_data instance and...
11951 (make_pass_loop2): ...new function.
11952 (pass_rtl_loop_init): Convert from a global struct to a subclass of
11953 rtl_opt_pass along with...
11954 (pass_data_rtl_loop_init): ...new pass_data instance and...
11955 (make_pass_rtl_loop_init): ...new function.
11956 (pass_rtl_loop_done): Convert from a global struct to a subclass of
11957 rtl_opt_pass along with...
11958 (pass_data_rtl_loop_done): ...new pass_data instance and...
11959 (make_pass_rtl_loop_done): ...new function.
11960 (pass_rtl_move_loop_invariants): Convert from a global struct to a
11961 subclass of rtl_opt_pass along with...
11962 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
11963 (make_pass_rtl_move_loop_invariants): ...new function.
11964 (pass_rtl_unswitch): Convert from a global struct to a subclass of
11965 rtl_opt_pass along with...
11966 (pass_data_rtl_unswitch): ...new pass_data instance and...
11967 (make_pass_rtl_unswitch): ...new function.
11968 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
11969 subclass of rtl_opt_pass along with...
11970 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
11972 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11973 (pass_rtl_doloop): Convert from a global struct to a subclass of
11974 rtl_opt_pass along with...
11975 (pass_data_rtl_doloop): ...new pass_data instance and...
11976 (make_pass_rtl_doloop): ...new function.
11977 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
11978 a subclass of rtl_opt_pass along with...
11979 (pass_data_lower_subreg): ...new pass_data instance and...
11980 (make_pass_lower_subreg): ...new function.
11981 (pass_lower_subreg2): Convert from a global struct to a subclass of
11982 rtl_opt_pass along with...
11983 (pass_data_lower_subreg2): ...new pass_data instance and...
11984 (make_pass_lower_subreg2): ...new function.
11985 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
11986 struct to a subclass of ipa_opt_pass_d along with...
11987 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
11988 (make_pass_ipa_lto_gimple_out): ...new function.
11989 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
11990 of ipa_opt_pass_d along with...
11991 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
11992 (make_pass_ipa_lto_finish_out): ...new function.
11993 * mode-switching.c (pass_mode_switching): Convert from a global struct
11994 to a subclass of rtl_opt_pass along with...
11995 (pass_data_mode_switching): ...new pass_data instance and...
11996 (make_pass_mode_switching): ...new function.
11997 * modulo-sched.c (pass_sms): Convert from a global struct to a
11998 subclass of rtl_opt_pass along with...
11999 (pass_data_sms): ...new pass_data instance and...
12000 (make_pass_sms): ...new function.
12001 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12002 subclass of gimple_opt_pass along with...
12003 (pass_data_expand_omp): ...new pass_data instance and...
12004 (make_pass_expand_omp): ...new function.
12005 (pass_lower_omp): Convert from a global struct to a subclass of
12006 gimple_opt_pass along with...
12007 (pass_data_lower_omp): ...new pass_data instance and...
12008 (make_pass_lower_omp): ...new function.
12009 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12010 of gimple_opt_pass along with...
12011 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12012 (make_pass_diagnose_omp_blocks): ...new function.
12013 * passes.c (pass_early_local_passes): Convert from a global struct to
12014 a subclass of simple_ipa_opt_pass along with...
12015 (pass_data_early_local_passes): ...new pass_data instance and...
12016 (make_pass_early_local_passes): ...new function.
12017 (pass_all_early_optimizations): Convert from a global struct to a
12018 subclass of gimple_opt_pass along with...
12019 (pass_data_all_early_optimizations): ...new pass_data instance and...
12020 (make_pass_all_early_optimizations): ...new function.
12021 (pass_all_optimizations): Convert from a global struct to a subclass
12022 of gimple_opt_pass along with...
12023 (pass_data_all_optimizations): ...new pass_data instance and...
12024 (make_pass_all_optimizations): ...new function.
12025 (pass_all_optimizations_g): Convert from a global struct to a subclass
12026 of gimple_opt_pass along with...
12027 (pass_data_all_optimizations_g): ...new pass_data instance and...
12028 (make_pass_all_optimizations_g): ...new function.
12029 (pass_rest_of_compilation): Convert from a global struct to a subclass
12030 of rtl_opt_pass along with...
12031 (pass_data_rest_of_compilation): ...new pass_data instance and...
12032 (make_pass_rest_of_compilation): ...new function.
12033 (pass_postreload): Convert from a global struct to a subclass of
12034 rtl_opt_pass along with...
12035 (pass_data_postreload): ...new pass_data instance and...
12036 (make_pass_postreload): ...new function.
12037 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12038 subclass of rtl_opt_pass along with...
12039 (pass_data_gcse2): ...new pass_data instance and...
12040 (make_pass_gcse2): ...new function.
12041 * postreload.c (pass_postreload_cse): Convert from a global struct to
12042 a subclass of rtl_opt_pass along with...
12043 (pass_data_postreload_cse): ...new pass_data instance and...
12044 (make_pass_postreload_cse): ...new function.
12045 * predict.c (pass_profile): Convert from a global struct to a subclass
12046 of gimple_opt_pass along with...
12047 (pass_data_profile): ...new pass_data instance and...
12048 (make_pass_profile): ...new function.
12049 (pass_strip_predict_hints): Convert from a global struct to a subclass
12050 of gimple_opt_pass along with...
12051 (pass_data_strip_predict_hints): ...new pass_data instance and...
12052 (make_pass_strip_predict_hints): ...new function.
12053 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12054 of rtl_opt_pass along with...
12055 (pass_data_peephole2): ...new pass_data instance and...
12056 (make_pass_peephole2): ...new function.
12057 (pass_split_all_insns): Convert from a global struct to a subclass of
12058 rtl_opt_pass along with...
12059 (pass_data_split_all_insns): ...new pass_data instance and...
12060 (make_pass_split_all_insns): ...new function.
12061 (pass_split_after_reload): Convert from a global struct to a subclass
12062 of rtl_opt_pass along with...
12063 (pass_data_split_after_reload): ...new pass_data instance and...
12064 (make_pass_split_after_reload): ...new function.
12065 (pass_split_before_regstack): Convert from a global struct to a
12066 subclass of rtl_opt_pass along with...
12067 (pass_data_split_before_regstack): ...new pass_data instance and...
12068 (make_pass_split_before_regstack): ...new function.
12069 (pass_split_before_sched2): Convert from a global struct to a subclass
12070 of rtl_opt_pass along with...
12071 (pass_data_split_before_sched2): ...new pass_data instance and...
12072 (make_pass_split_before_sched2): ...new function.
12073 (pass_split_for_shorten_branches): Convert from a global struct to a
12074 subclass of rtl_opt_pass along with...
12075 (pass_data_split_for_shorten_branches): ...new pass_data instance
12077 (make_pass_split_for_shorten_branches): ...new function.
12078 * ree.c (pass_ree): Convert from a global struct to a subclass of
12079 rtl_opt_pass along with...
12080 (pass_data_ree): ...new pass_data instance and...
12081 (make_pass_ree): ...new function.
12082 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12083 subclass of rtl_opt_pass along with...
12084 (pass_data_stack_regs): ...new pass_data instance and...
12085 (make_pass_stack_regs): ...new function.
12086 (pass_stack_regs_run): Convert from a global struct to a subclass of
12087 rtl_opt_pass along with...
12088 (pass_data_stack_regs_run): ...new pass_data instance and...
12089 (make_pass_stack_regs_run): ...new function.
12090 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12091 subclass of rtl_opt_pass along with...
12092 (pass_data_cprop_hardreg): ...new pass_data instance and...
12093 (make_pass_cprop_hardreg): ...new function.
12094 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12095 subclass of rtl_opt_pass along with...
12096 (pass_data_reginfo_init): ...new pass_data instance and...
12097 (make_pass_reginfo_init): ...new function.
12098 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12099 of rtl_opt_pass along with...
12100 (pass_data_regmove): ...new pass_data instance and...
12101 (make_pass_regmove): ...new function.
12102 * regrename.c (pass_regrename): Convert from a global struct to a
12103 subclass of rtl_opt_pass along with...
12104 (pass_data_regrename): ...new pass_data instance and...
12105 (make_pass_regrename): ...new function.
12106 * reorg.c (pass_delay_slots): Convert from a global struct to a
12107 subclass of rtl_opt_pass along with...
12108 (pass_data_delay_slots): ...new pass_data instance and...
12109 (make_pass_delay_slots): ...new function.
12110 (pass_machine_reorg): Convert from a global struct to a subclass of
12111 rtl_opt_pass along with...
12112 (pass_data_machine_reorg): ...new pass_data instance and...
12113 (make_pass_machine_reorg): ...new function.
12114 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12115 of rtl_opt_pass along with...
12116 (pass_data_sched): ...new pass_data instance and...
12117 (make_pass_sched): ...new function.
12118 (pass_sched2): Convert from a global struct to a subclass of
12119 rtl_opt_pass along with...
12120 (pass_data_sched2): ...new pass_data instance and...
12121 (make_pass_sched2): ...new function.
12122 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12123 to a subclass of rtl_opt_pass along with...
12124 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12125 (make_pass_stack_ptr_mod): ...new function.
12126 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12127 to a subclass of rtl_opt_pass along with...
12128 (pass_data_rtl_store_motion): ...new pass_data instance and...
12129 (make_pass_rtl_store_motion): ...new function.
12130 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12131 of gimple_opt_pass along with...
12132 (pass_data_tracer): ...new pass_data instance and...
12133 (make_pass_tracer): ...new function.
12134 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12135 to a subclass of gimple_opt_pass along with...
12136 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12137 (make_pass_diagnose_tm_blocks): ...new function.
12138 (pass_lower_tm): Convert from a global struct to a subclass of
12139 gimple_opt_pass along with...
12140 (pass_data_lower_tm): ...new pass_data instance and...
12141 (make_pass_lower_tm): ...new function.
12142 (pass_tm_init): Convert from a global struct to a subclass of
12143 gimple_opt_pass along with...
12144 (pass_data_tm_init): ...new pass_data instance and...
12145 (make_pass_tm_init): ...new function.
12146 (pass_tm_mark): Convert from a global struct to a subclass of
12147 gimple_opt_pass along with...
12148 (pass_data_tm_mark): ...new pass_data instance and...
12149 (make_pass_tm_mark): ...new function.
12150 (pass_tm_edges): Convert from a global struct to a subclass of
12151 gimple_opt_pass along with...
12152 (pass_data_tm_edges): ...new pass_data instance and...
12153 (make_pass_tm_edges): ...new function.
12154 (pass_tm_memopt): Convert from a global struct to a subclass of
12155 gimple_opt_pass along with...
12156 (pass_data_tm_memopt): ...new pass_data instance and...
12157 (make_pass_tm_memopt): ...new function.
12158 (pass_ipa_tm): Convert from a global struct to a subclass of
12159 simple_ipa_opt_pass along with...
12160 (pass_data_ipa_tm): ...new pass_data instance and...
12161 (make_pass_ipa_tm): ...new function.
12162 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12163 subclass of gimple_opt_pass along with...
12164 (pass_data_call_cdce): ...new pass_data instance and...
12165 (make_pass_call_cdce): ...new function.
12166 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12167 subclass of gimple_opt_pass along with...
12168 (pass_data_build_cfg): ...new pass_data instance and...
12169 (make_pass_build_cfg): ...new function.
12170 (pass_split_crit_edges): Convert from a global struct to a subclass of
12171 gimple_opt_pass along with...
12172 (pass_data_split_crit_edges): ...new pass_data instance and...
12173 (make_pass_split_crit_edges): ...new function.
12174 (pass_warn_function_return): Convert from a global struct to a
12175 subclass of gimple_opt_pass along with...
12176 (pass_data_warn_function_return): ...new pass_data instance and...
12177 (make_pass_warn_function_return): ...new function.
12178 (pass_warn_function_noreturn): Convert from a global struct to a
12179 subclass of gimple_opt_pass along with...
12180 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12181 (make_pass_warn_function_noreturn): ...new function.
12182 (pass_warn_unused_result): Convert from a global struct to a subclass
12183 of gimple_opt_pass along with...
12184 (pass_data_warn_unused_result): ...new pass_data instance and...
12185 (make_pass_warn_unused_result): ...new function.
12186 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12187 a subclass of gimple_opt_pass along with...
12188 (pass_data_merge_phi): ...new pass_data instance and...
12189 (make_pass_merge_phi): ...new function.
12190 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12191 a subclass of gimple_opt_pass along with...
12192 (pass_data_lower_complex): ...new pass_data instance and...
12193 (make_pass_lower_complex): ...new function.
12194 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12195 gimple_opt_pass along with...
12196 (pass_data_lower_complex_O0): ...new pass_data instance and...
12197 (make_pass_lower_complex_O0): ...new function.
12198 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12199 subclass of gimple_opt_pass along with...
12200 (pass_data_lower_eh): ...new pass_data instance and...
12201 (make_pass_lower_eh): ...new function.
12202 (pass_refactor_eh): Convert from a global struct to a subclass of
12203 gimple_opt_pass along with...
12204 (pass_data_refactor_eh): ...new pass_data instance and...
12205 (make_pass_refactor_eh): ...new function.
12206 (pass_lower_resx): Convert from a global struct to a subclass of
12207 gimple_opt_pass along with...
12208 (pass_data_lower_resx): ...new pass_data instance and...
12209 (make_pass_lower_resx): ...new function.
12210 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12211 of gimple_opt_pass along with...
12212 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12213 (make_pass_lower_eh_dispatch): ...new function.
12214 (pass_cleanup_eh): Convert from a global struct to a subclass of
12215 gimple_opt_pass along with...
12216 (pass_data_cleanup_eh): ...new pass_data instance and...
12217 (make_pass_cleanup_eh): ...new function.
12218 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
12219 to a subclass of simple_ipa_opt_pass along with...
12220 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
12221 (make_pass_ipa_lower_emutls): ...new function.
12222 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
12223 a subclass of gimple_opt_pass along with...
12224 (pass_data_if_conversion): ...new pass_data instance and...
12225 (make_pass_if_conversion): ...new function.
12226 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
12227 subclass of gimple_opt_pass along with...
12228 (pass_data_build_ssa): ...new pass_data instance and...
12229 (make_pass_build_ssa): ...new function.
12230 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
12231 global struct to a subclass of gimple_opt_pass along with...
12232 (pass_data_loop_distribution): ...new pass_data instance and...
12233 (make_pass_loop_distribution): ...new function.
12234 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
12235 subclass of gimple_opt_pass along with...
12236 (pass_data_mudflap_1): ...new pass_data instance and...
12237 (make_pass_mudflap_1): ...new function.
12238 (pass_mudflap_2): Convert from a global struct to a subclass of
12239 gimple_opt_pass along with...
12240 (pass_data_mudflap_2): ...new pass_data instance and...
12241 (make_pass_mudflap_2): ...new function.
12242 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
12243 subclass of gimple_opt_pass along with...
12244 (pass_data_mudflap_1): ...new pass_data instance and...
12245 (make_pass_mudflap_1): ...new function.
12246 (pass_mudflap_2): Convert from a global struct to a subclass of
12247 gimple_opt_pass along with...
12248 (pass_data_mudflap_2): ...new pass_data instance and...
12249 (make_pass_mudflap_2): ...new function.
12250 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
12251 gimple_opt_pass along with...
12252 (pass_data_nrv): ...new pass_data instance and...
12253 (make_pass_nrv): ...new function.
12254 (pass_return_slot): Convert from a global struct to a subclass of
12255 gimple_opt_pass along with...
12256 (pass_data_return_slot): ...new pass_data instance and...
12257 (make_pass_return_slot): ...new function.
12258 * tree-object-size.c (pass_object_sizes): Convert from a global struct
12259 to a subclass of gimple_opt_pass along with...
12260 (pass_data_object_sizes): ...new pass_data instance and...
12261 (make_pass_object_sizes): ...new function.
12262 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
12263 global struct to a subclass of gimple_opt_pass along with...
12264 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
12266 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12267 (pass_fixup_cfg): Convert from a global struct to a subclass of
12268 gimple_opt_pass along with...
12269 (pass_data_fixup_cfg): ...new pass_data instance and...
12270 (make_pass_fixup_cfg): ...new function.
12271 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
12272 (make_pass_mudflap_1): ...new function.
12273 (pass_mudflap_2): Replace declaration with that of...
12274 (make_pass_mudflap_2): ...new function.
12275 (pass_asan): Replace declaration with that of...
12276 (make_pass_asan): ...new function.
12277 (pass_asan_O0): Replace declaration with that of...
12278 (make_pass_asan_O0): ...new function.
12279 (pass_tsan): Replace declaration with that of...
12280 (make_pass_tsan): ...new function.
12281 (pass_tsan_O0): Replace declaration with that of...
12282 (make_pass_tsan_O0): ...new function.
12283 (pass_lower_cf): Replace declaration with that of...
12284 (make_pass_lower_cf): ...new function.
12285 (pass_refactor_eh): Replace declaration with that of...
12286 (make_pass_refactor_eh): ...new function.
12287 (pass_lower_eh): Replace declaration with that of...
12288 (make_pass_lower_eh): ...new function.
12289 (pass_lower_eh_dispatch): Replace declaration with that of...
12290 (make_pass_lower_eh_dispatch): ...new function.
12291 (pass_lower_resx): Replace declaration with that of...
12292 (make_pass_lower_resx): ...new function.
12293 (pass_build_cfg): Replace declaration with that of...
12294 (make_pass_build_cfg): ...new function.
12295 (pass_early_tree_profile): Replace declaration with that of...
12296 (make_pass_early_tree_profile): ...new function.
12297 (pass_cleanup_eh): Replace declaration with that of...
12298 (make_pass_cleanup_eh): ...new function.
12299 (pass_sra): Replace declaration with that of...
12300 (make_pass_sra): ...new function.
12301 (pass_sra_early): Replace declaration with that of...
12302 (make_pass_sra_early): ...new function.
12303 (pass_early_ipa_sra): Replace declaration with that of...
12304 (make_pass_early_ipa_sra): ...new function.
12305 (pass_tail_recursion): Replace declaration with that of...
12306 (make_pass_tail_recursion): ...new function.
12307 (pass_tail_calls): Replace declaration with that of...
12308 (make_pass_tail_calls): ...new function.
12309 (pass_tree_loop): Replace declaration with that of...
12310 (make_pass_tree_loop): ...new function.
12311 (pass_tree_loop_init): Replace declaration with that of...
12312 (make_pass_tree_loop_init): ...new function.
12313 (pass_lim): Replace declaration with that of...
12314 (make_pass_lim): ...new function.
12315 (pass_tree_unswitch): Replace declaration with that of...
12316 (make_pass_tree_unswitch): ...new function.
12317 (pass_predcom): Replace declaration with that of...
12318 (make_pass_predcom): ...new function.
12319 (pass_iv_canon): Replace declaration with that of...
12320 (make_pass_iv_canon): ...new function.
12321 (pass_scev_cprop): Replace declaration with that of...
12322 (make_pass_scev_cprop): ...new function.
12323 (pass_empty_loop): Replace declaration with that of...
12324 (make_pass_empty_loop): ...new function.
12325 (pass_record_bounds): Replace declaration with that of...
12326 (make_pass_record_bounds): ...new function.
12327 (pass_graphite): Replace declaration with that of...
12328 (make_pass_graphite): ...new function.
12329 (pass_graphite_transforms): Replace declaration with that of...
12330 (make_pass_graphite_transforms): ...new function.
12331 (pass_if_conversion): Replace declaration with that of...
12332 (make_pass_if_conversion): ...new function.
12333 (pass_loop_distribution): Replace declaration with that of...
12334 (make_pass_loop_distribution): ...new function.
12335 (pass_vectorize): Replace declaration with that of...
12336 (make_pass_vectorize): ...new function.
12337 (pass_slp_vectorize): Replace declaration with that of...
12338 (make_pass_slp_vectorize): ...new function.
12339 (pass_complete_unroll): Replace declaration with that of...
12340 (make_pass_complete_unroll): ...new function.
12341 (pass_complete_unrolli): Replace declaration with that of...
12342 (make_pass_complete_unrolli): ...new function.
12343 (pass_parallelize_loops): Replace declaration with that of...
12344 (make_pass_parallelize_loops): ...new function.
12345 (pass_loop_prefetch): Replace declaration with that of...
12346 (make_pass_loop_prefetch): ...new function.
12347 (pass_iv_optimize): Replace declaration with that of...
12348 (make_pass_iv_optimize): ...new function.
12349 (pass_tree_loop_done): Replace declaration with that of...
12350 (make_pass_tree_loop_done): ...new function.
12351 (pass_ch): Replace declaration with that of...
12352 (make_pass_ch): ...new function.
12353 (pass_ccp): Replace declaration with that of...
12354 (make_pass_ccp): ...new function.
12355 (pass_phi_only_cprop): Replace declaration with that of...
12356 (make_pass_phi_only_cprop): ...new function.
12357 (pass_build_ssa): Replace declaration with that of...
12358 (make_pass_build_ssa): ...new function.
12359 (pass_build_alias): Replace declaration with that of...
12360 (make_pass_build_alias): ...new function.
12361 (pass_build_ealias): Replace declaration with that of...
12362 (make_pass_build_ealias): ...new function.
12363 (pass_dominator): Replace declaration with that of...
12364 (make_pass_dominator): ...new function.
12365 (pass_dce): Replace declaration with that of...
12366 (make_pass_dce): ...new function.
12367 (pass_dce_loop): Replace declaration with that of...
12368 (make_pass_dce_loop): ...new function.
12369 (pass_cd_dce): Replace declaration with that of...
12370 (make_pass_cd_dce): ...new function.
12371 (pass_call_cdce): Replace declaration with that of...
12372 (make_pass_call_cdce): ...new function.
12373 (pass_merge_phi): Replace declaration with that of...
12374 (make_pass_merge_phi): ...new function.
12375 (pass_split_crit_edges): Replace declaration with that of...
12376 (make_pass_split_crit_edges): ...new function.
12377 (pass_pre): Replace declaration with that of...
12378 (make_pass_pre): ...new function.
12379 (pass_profile): Replace declaration with that of...
12380 (make_pass_profile): ...new function.
12381 (pass_strip_predict_hints): Replace declaration with that of...
12382 (make_pass_strip_predict_hints): ...new function.
12383 (pass_lower_complex_O0): Replace declaration with that of...
12384 (make_pass_lower_complex_O0): ...new function.
12385 (pass_lower_complex): Replace declaration with that of...
12386 (make_pass_lower_complex): ...new function.
12387 (pass_lower_vector): Replace declaration with that of...
12388 (make_pass_lower_vector): ...new function.
12389 (pass_lower_vector_ssa): Replace declaration with that of...
12390 (make_pass_lower_vector_ssa): ...new function.
12391 (pass_lower_omp): Replace declaration with that of...
12392 (make_pass_lower_omp): ...new function.
12393 (pass_diagnose_omp_blocks): Replace declaration with that of...
12394 (make_pass_diagnose_omp_blocks): ...new function.
12395 (pass_expand_omp): Replace declaration with that of...
12396 (make_pass_expand_omp): ...new function.
12397 (pass_expand_omp_ssa): Replace declaration with that of...
12398 (make_pass_expand_omp_ssa): ...new function.
12399 (pass_object_sizes): Replace declaration with that of...
12400 (make_pass_object_sizes): ...new function.
12401 (pass_strlen): Replace declaration with that of...
12402 (make_pass_strlen): ...new function.
12403 (pass_fold_builtins): Replace declaration with that of...
12404 (make_pass_fold_builtins): ...new function.
12405 (pass_stdarg): Replace declaration with that of...
12406 (make_pass_stdarg): ...new function.
12407 (pass_early_warn_uninitialized): Replace declaration with that of...
12408 (make_pass_early_warn_uninitialized): ...new function.
12409 (pass_late_warn_uninitialized): Replace declaration with that of...
12410 (make_pass_late_warn_uninitialized): ...new function.
12411 (pass_cse_reciprocals): Replace declaration with that of...
12412 (make_pass_cse_reciprocals): ...new function.
12413 (pass_cse_sincos): Replace declaration with that of...
12414 (make_pass_cse_sincos): ...new function.
12415 (pass_optimize_bswap): Replace declaration with that of...
12416 (make_pass_optimize_bswap): ...new function.
12417 (pass_optimize_widening_mul): Replace declaration with that of...
12418 (make_pass_optimize_widening_mul): ...new function.
12419 (pass_warn_function_return): Replace declaration with that of...
12420 (make_pass_warn_function_return): ...new function.
12421 (pass_warn_function_noreturn): Replace declaration with that of...
12422 (make_pass_warn_function_noreturn): ...new function.
12423 (pass_cselim): Replace declaration with that of...
12424 (make_pass_cselim): ...new function.
12425 (pass_phiopt): Replace declaration with that of...
12426 (make_pass_phiopt): ...new function.
12427 (pass_forwprop): Replace declaration with that of...
12428 (make_pass_forwprop): ...new function.
12429 (pass_phiprop): Replace declaration with that of...
12430 (make_pass_phiprop): ...new function.
12431 (pass_tree_ifcombine): Replace declaration with that of...
12432 (make_pass_tree_ifcombine): ...new function.
12433 (pass_dse): Replace declaration with that of...
12434 (make_pass_dse): ...new function.
12435 (pass_nrv): Replace declaration with that of...
12436 (make_pass_nrv): ...new function.
12437 (pass_rename_ssa_copies): Replace declaration with that of...
12438 (make_pass_rename_ssa_copies): ...new function.
12439 (pass_sink_code): Replace declaration with that of...
12440 (make_pass_sink_code): ...new function.
12441 (pass_fre): Replace declaration with that of...
12442 (make_pass_fre): ...new function.
12443 (pass_check_data_deps): Replace declaration with that of...
12444 (make_pass_check_data_deps): ...new function.
12445 (pass_copy_prop): Replace declaration with that of...
12446 (make_pass_copy_prop): ...new function.
12447 (pass_vrp): Replace declaration with that of...
12448 (make_pass_vrp): ...new function.
12449 (pass_uncprop): Replace declaration with that of...
12450 (make_pass_uncprop): ...new function.
12451 (pass_return_slot): Replace declaration with that of...
12452 (make_pass_return_slot): ...new function.
12453 (pass_reassoc): Replace declaration with that of...
12454 (make_pass_reassoc): ...new function.
12455 (pass_rebuild_cgraph_edges): Replace declaration with that of...
12456 (make_pass_rebuild_cgraph_edges): ...new function.
12457 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
12458 (make_pass_remove_cgraph_callee_edges): ...new function.
12459 (pass_build_cgraph_edges): Replace declaration with that of...
12460 (make_pass_build_cgraph_edges): ...new function.
12461 (pass_local_pure_const): Replace declaration with that of...
12462 (make_pass_local_pure_const): ...new function.
12463 (pass_tracer): Replace declaration with that of...
12464 (make_pass_tracer): ...new function.
12465 (pass_warn_unused_result): Replace declaration with that of...
12466 (make_pass_warn_unused_result): ...new function.
12467 (pass_diagnose_tm_blocks): Replace declaration with that of...
12468 (make_pass_diagnose_tm_blocks): ...new function.
12469 (pass_lower_tm): Replace declaration with that of...
12470 (make_pass_lower_tm): ...new function.
12471 (pass_tm_init): Replace declaration with that of...
12472 (make_pass_tm_init): ...new function.
12473 (pass_tm_mark): Replace declaration with that of...
12474 (make_pass_tm_mark): ...new function.
12475 (pass_tm_memopt): Replace declaration with that of...
12476 (make_pass_tm_memopt): ...new function.
12477 (pass_tm_edges): Replace declaration with that of...
12478 (make_pass_tm_edges): ...new function.
12479 (pass_split_functions): Replace declaration with that of...
12480 (make_pass_split_functions): ...new function.
12481 (pass_feedback_split_functions): Replace declaration with that of...
12482 (make_pass_feedback_split_functions): ...new function.
12483 (pass_strength_reduction): Replace declaration with that of...
12484 (make_pass_strength_reduction): ...new function.
12485 (pass_ipa_lower_emutls): Replace declaration with that of...
12486 (make_pass_ipa_lower_emutls): ...new function.
12487 (pass_ipa_function_and_variable_visibility): Replace declaration with
12489 (make_pass_ipa_function_and_variable_visibility): ...new function.
12490 (pass_ipa_tree_profile): Replace declaration with that of...
12491 (make_pass_ipa_tree_profile): ...new function.
12492 (pass_early_local_passes): Replace declaration with that of...
12493 (make_pass_early_local_passes): ...new function.
12494 (pass_ipa_whole_program_visibility): Replace declaration with that
12496 (make_pass_ipa_whole_program_visibility): ...new function.
12497 (pass_ipa_lto_gimple_out): Replace declaration with that of...
12498 (make_pass_ipa_lto_gimple_out): ...new function.
12499 (pass_ipa_increase_alignment): Replace declaration with that of...
12500 (make_pass_ipa_increase_alignment): ...new function.
12501 (pass_ipa_inline): Replace declaration with that of...
12502 (make_pass_ipa_inline): ...new function.
12503 (pass_ipa_free_lang_data): Replace declaration with that of...
12504 (make_pass_ipa_free_lang_data): ...new function.
12505 (pass_ipa_free_inline_summary): Replace declaration with that of...
12506 (make_pass_ipa_free_inline_summary): ...new function.
12507 (pass_ipa_cp): Replace declaration with that of...
12508 (make_pass_ipa_cp): ...new function.
12509 (pass_ipa_reference): Replace declaration with that of...
12510 (make_pass_ipa_reference): ...new function.
12511 (pass_ipa_pure_const): Replace declaration with that of...
12512 (make_pass_ipa_pure_const): ...new function.
12513 (pass_ipa_pta): Replace declaration with that of...
12514 (make_pass_ipa_pta): ...new function.
12515 (pass_ipa_lto_finish_out): Replace declaration with that of...
12516 (make_pass_ipa_lto_finish_out): ...new function.
12517 (pass_ipa_tm): Replace declaration with that of...
12518 (make_pass_ipa_tm): ...new function.
12519 (pass_ipa_profile): Replace declaration with that of...
12520 (make_pass_ipa_profile): ...new function.
12521 (pass_ipa_cdtor_merge): Replace declaration with that of...
12522 (make_pass_ipa_cdtor_merge): ...new function.
12523 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
12525 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12526 (pass_init_datastructures): Replace declaration with that of...
12527 (make_pass_init_datastructures): ...new function.
12528 (pass_fixup_cfg): Replace declaration with that of...
12529 (make_pass_fixup_cfg): ...new function.
12530 (pass_expand): Replace declaration with that of...
12531 (make_pass_expand): ...new function.
12532 (pass_instantiate_virtual_regs): Replace declaration with that of...
12533 (make_pass_instantiate_virtual_regs): ...new function.
12534 (pass_rtl_fwprop): Replace declaration with that of...
12535 (make_pass_rtl_fwprop): ...new function.
12536 (pass_rtl_fwprop_addr): Replace declaration with that of...
12537 (make_pass_rtl_fwprop_addr): ...new function.
12538 (pass_jump): Replace declaration with that of...
12539 (make_pass_jump): ...new function.
12540 (pass_jump2): Replace declaration with that of...
12541 (make_pass_jump2): ...new function.
12542 (pass_lower_subreg): Replace declaration with that of...
12543 (make_pass_lower_subreg): ...new function.
12544 (pass_cse): Replace declaration with that of...
12545 (make_pass_cse): ...new function.
12546 (pass_fast_rtl_dce): Replace declaration with that of...
12547 (make_pass_fast_rtl_dce): ...new function.
12548 (pass_ud_rtl_dce): Replace declaration with that of...
12549 (make_pass_ud_rtl_dce): ...new function.
12550 (pass_rtl_dce): Replace declaration with that of...
12551 (make_pass_rtl_dce): ...new function.
12552 (pass_rtl_dse1): Replace declaration with that of...
12553 (make_pass_rtl_dse1): ...new function.
12554 (pass_rtl_dse2): Replace declaration with that of...
12555 (make_pass_rtl_dse2): ...new function.
12556 (pass_rtl_dse3): Replace declaration with that of...
12557 (make_pass_rtl_dse3): ...new function.
12558 (pass_rtl_cprop): Replace declaration with that of...
12559 (make_pass_rtl_cprop): ...new function.
12560 (pass_rtl_pre): Replace declaration with that of...
12561 (make_pass_rtl_pre): ...new function.
12562 (pass_rtl_hoist): Replace declaration with that of...
12563 (make_pass_rtl_hoist): ...new function.
12564 (pass_rtl_store_motion): Replace declaration with that of...
12565 (make_pass_rtl_store_motion): ...new function.
12566 (pass_cse_after_global_opts): Replace declaration with that of...
12567 (make_pass_cse_after_global_opts): ...new function.
12568 (pass_rtl_ifcvt): Replace declaration with that of...
12569 (make_pass_rtl_ifcvt): ...new function.
12570 (pass_into_cfg_layout_mode): Replace declaration with that of...
12571 (make_pass_into_cfg_layout_mode): ...new function.
12572 (pass_outof_cfg_layout_mode): Replace declaration with that of...
12573 (make_pass_outof_cfg_layout_mode): ...new function.
12574 (pass_loop2): Replace declaration with that of...
12575 (make_pass_loop2): ...new function.
12576 (pass_rtl_loop_init): Replace declaration with that of...
12577 (make_pass_rtl_loop_init): ...new function.
12578 (pass_rtl_move_loop_invariants): Replace declaration with that of...
12579 (make_pass_rtl_move_loop_invariants): ...new function.
12580 (pass_rtl_unswitch): Replace declaration with that of...
12581 (make_pass_rtl_unswitch): ...new function.
12582 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
12583 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12584 (pass_rtl_doloop): Replace declaration with that of...
12585 (make_pass_rtl_doloop): ...new function.
12586 (pass_rtl_loop_done): Replace declaration with that of...
12587 (make_pass_rtl_loop_done): ...new function.
12588 (pass_web): Replace declaration with that of...
12589 (make_pass_web): ...new function.
12590 (pass_cse2): Replace declaration with that of...
12591 (make_pass_cse2): ...new function.
12592 (pass_df_initialize_opt): Replace declaration with that of...
12593 (make_pass_df_initialize_opt): ...new function.
12594 (pass_df_initialize_no_opt): Replace declaration with that of...
12595 (make_pass_df_initialize_no_opt): ...new function.
12596 (pass_reginfo_init): Replace declaration with that of...
12597 (make_pass_reginfo_init): ...new function.
12598 (pass_inc_dec): Replace declaration with that of...
12599 (make_pass_inc_dec): ...new function.
12600 (pass_stack_ptr_mod): Replace declaration with that of...
12601 (make_pass_stack_ptr_mod): ...new function.
12602 (pass_initialize_regs): Replace declaration with that of...
12603 (make_pass_initialize_regs): ...new function.
12604 (pass_combine): Replace declaration with that of...
12605 (make_pass_combine): ...new function.
12606 (pass_if_after_combine): Replace declaration with that of...
12607 (make_pass_if_after_combine): ...new function.
12608 (pass_ree): Replace declaration with that of...
12609 (make_pass_ree): ...new function.
12610 (pass_partition_blocks): Replace declaration with that of...
12611 (make_pass_partition_blocks): ...new function.
12612 (pass_match_asm_constraints): Replace declaration with that of...
12613 (make_pass_match_asm_constraints): ...new function.
12614 (pass_regmove): Replace declaration with that of...
12615 (make_pass_regmove): ...new function.
12616 (pass_split_all_insns): Replace declaration with that of...
12617 (make_pass_split_all_insns): ...new function.
12618 (pass_fast_rtl_byte_dce): Replace declaration with that of...
12619 (make_pass_fast_rtl_byte_dce): ...new function.
12620 (pass_lower_subreg2): Replace declaration with that of...
12621 (make_pass_lower_subreg2): ...new function.
12622 (pass_mode_switching): Replace declaration with that of...
12623 (make_pass_mode_switching): ...new function.
12624 (pass_sms): Replace declaration with that of...
12625 (make_pass_sms): ...new function.
12626 (pass_sched): Replace declaration with that of...
12627 (make_pass_sched): ...new function.
12628 (pass_ira): Replace declaration with that of...
12629 (make_pass_ira): ...new function.
12630 (pass_reload): Replace declaration with that of...
12631 (make_pass_reload): ...new function.
12632 (pass_clean_state): Replace declaration with that of...
12633 (make_pass_clean_state): ...new function.
12634 (pass_branch_prob): Replace declaration with that of...
12635 (make_pass_branch_prob): ...new function.
12636 (pass_value_profile_transformations): Replace declaration with that
12638 (make_pass_value_profile_transformations): ...new function.
12639 (pass_postreload_cse): Replace declaration with that of...
12640 (make_pass_postreload_cse): ...new function.
12641 (pass_gcse2): Replace declaration with that of...
12642 (make_pass_gcse2): ...new function.
12643 (pass_split_after_reload): Replace declaration with that of...
12644 (make_pass_split_after_reload): ...new function.
12645 (pass_branch_target_load_optimize1): Replace declaration with that
12647 (make_pass_branch_target_load_optimize1): ...new function.
12648 (pass_thread_prologue_and_epilogue): Replace declaration with that
12650 (make_pass_thread_prologue_and_epilogue): ...new function.
12651 (pass_stack_adjustments): Replace declaration with that of...
12652 (make_pass_stack_adjustments): ...new function.
12653 (pass_peephole2): Replace declaration with that of...
12654 (make_pass_peephole2): ...new function.
12655 (pass_if_after_reload): Replace declaration with that of...
12656 (make_pass_if_after_reload): ...new function.
12657 (pass_regrename): Replace declaration with that of...
12658 (make_pass_regrename): ...new function.
12659 (pass_cprop_hardreg): Replace declaration with that of...
12660 (make_pass_cprop_hardreg): ...new function.
12661 (pass_reorder_blocks): Replace declaration with that of...
12662 (make_pass_reorder_blocks): ...new function.
12663 (pass_branch_target_load_optimize2): Replace declaration with that
12665 (make_pass_branch_target_load_optimize2): ...new function.
12666 (pass_leaf_regs): Replace declaration with that of...
12667 (make_pass_leaf_regs): ...new function.
12668 (pass_split_before_sched2): Replace declaration with that of...
12669 (make_pass_split_before_sched2): ...new function.
12670 (pass_compare_elim_after_reload): Replace declaration with that of...
12671 (make_pass_compare_elim_after_reload): ...new function.
12672 (pass_sched2): Replace declaration with that of...
12673 (make_pass_sched2): ...new function.
12674 (pass_stack_regs): Replace declaration with that of...
12675 (make_pass_stack_regs): ...new function.
12676 (pass_stack_regs_run): Replace declaration with that of...
12677 (make_pass_stack_regs_run): ...new function.
12678 (pass_df_finish): Replace declaration with that of...
12679 (make_pass_df_finish): ...new function.
12680 (pass_compute_alignments): Replace declaration with that of...
12681 (make_pass_compute_alignments): ...new function.
12682 (pass_duplicate_computed_gotos): Replace declaration with that of...
12683 (make_pass_duplicate_computed_gotos): ...new function.
12684 (pass_variable_tracking): Replace declaration with that of...
12685 (make_pass_variable_tracking): ...new function.
12686 (pass_free_cfg): Replace declaration with that of...
12687 (make_pass_free_cfg): ...new function.
12688 (pass_machine_reorg): Replace declaration with that of...
12689 (make_pass_machine_reorg): ...new function.
12690 (pass_cleanup_barriers): Replace declaration with that of...
12691 (make_pass_cleanup_barriers): ...new function.
12692 (pass_delay_slots): Replace declaration with that of...
12693 (make_pass_delay_slots): ...new function.
12694 (pass_split_for_shorten_branches): Replace declaration with that of...
12695 (make_pass_split_for_shorten_branches): ...new function.
12696 (pass_split_before_regstack): Replace declaration with that of...
12697 (make_pass_split_before_regstack): ...new function.
12698 (pass_convert_to_eh_region_ranges): Replace declaration with that
12700 (make_pass_convert_to_eh_region_ranges): ...new function.
12701 (pass_shorten_branches): Replace declaration with that of...
12702 (make_pass_shorten_branches): ...new function.
12703 (pass_set_nothrow_function_flags): Replace declaration with that of...
12704 (make_pass_set_nothrow_function_flags): ...new function.
12705 (pass_dwarf2_frame): Replace declaration with that of...
12706 (make_pass_dwarf2_frame): ...new function.
12707 (pass_final): Replace declaration with that of...
12708 (make_pass_final): ...new function.
12709 (pass_rtl_seqabstr): Replace declaration with that of...
12710 (make_pass_rtl_seqabstr): ...new function.
12711 (pass_release_ssa_names): Replace declaration with that of...
12712 (make_pass_release_ssa_names): ...new function.
12713 (pass_early_inline): Replace declaration with that of...
12714 (make_pass_early_inline): ...new function.
12715 (pass_inline_parameters): Replace declaration with that of...
12716 (make_pass_inline_parameters): ...new function.
12717 (pass_update_address_taken): Replace declaration with that of...
12718 (make_pass_update_address_taken): ...new function.
12719 (pass_convert_switch): Replace declaration with that of...
12720 (make_pass_convert_switch): ...new function.
12721 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
12722 to a subclass of simple_ipa_opt_pass along with...
12723 (pass_data_ipa_tree_profile): ...new pass_data instance and...
12724 (make_pass_ipa_tree_profile): ...new function.
12725 * tree-sra.c (pass_sra_early): Convert from a global struct to a
12726 subclass of gimple_opt_pass along with...
12727 (pass_data_sra_early): ...new pass_data instance and...
12728 (make_pass_sra_early): ...new function.
12729 (pass_sra): Convert from a global struct to a subclass of
12730 gimple_opt_pass along with...
12731 (pass_data_sra): ...new pass_data instance and...
12732 (make_pass_sra): ...new function.
12733 (pass_early_ipa_sra): Convert from a global struct to a subclass of
12734 gimple_opt_pass along with...
12735 (pass_data_early_ipa_sra): ...new pass_data instance and...
12736 (make_pass_early_ipa_sra): ...new function.
12737 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
12738 subclass of gimple_opt_pass along with...
12739 (pass_data_ccp): ...new pass_data instance and...
12740 (make_pass_ccp): ...new function.
12741 (pass_fold_builtins): Convert from a global struct to a subclass of
12742 gimple_opt_pass along with...
12743 (pass_data_fold_builtins): ...new pass_data instance and...
12744 (make_pass_fold_builtins): ...new function.
12745 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
12746 subclass of gimple_opt_pass along with...
12747 (pass_data_copy_prop): ...new pass_data instance and...
12748 (make_pass_copy_prop): ...new function.
12749 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
12750 global struct to a subclass of gimple_opt_pass along with...
12751 (pass_data_rename_ssa_copies): ...new pass_data instance and...
12752 (make_pass_rename_ssa_copies): ...new function.
12753 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
12754 subclass of gimple_opt_pass along with...
12755 (pass_data_dce): ...new pass_data instance and...
12756 (make_pass_dce): ...new function.
12757 (pass_dce_loop): Convert from a global struct to a subclass of
12758 gimple_opt_pass along with...
12759 (pass_data_dce_loop): ...new pass_data instance and...
12760 (make_pass_dce_loop): ...new function.
12761 (pass_cd_dce): Convert from a global struct to a subclass of
12762 gimple_opt_pass along with...
12763 (pass_data_cd_dce): ...new pass_data instance and...
12764 (make_pass_cd_dce): ...new function.
12765 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
12766 subclass of gimple_opt_pass along with...
12767 (pass_data_dominator): ...new pass_data instance and...
12768 (make_pass_dominator): ...new function.
12769 (pass_phi_only_cprop): Convert from a global struct to a subclass of
12770 gimple_opt_pass along with...
12771 (pass_data_phi_only_cprop): ...new pass_data instance and...
12772 (make_pass_phi_only_cprop): ...new function.
12773 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
12774 subclass of gimple_opt_pass along with...
12775 (pass_data_dse): ...new pass_data instance and...
12776 (make_pass_dse): ...new function.
12777 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
12778 a subclass of gimple_opt_pass along with...
12779 (pass_data_forwprop): ...new pass_data instance and...
12780 (make_pass_forwprop): ...new function.
12781 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
12782 struct to a subclass of gimple_opt_pass along with...
12783 (pass_data_tree_ifcombine): ...new pass_data instance and...
12784 (make_pass_tree_ifcombine): ...new function.
12785 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
12786 subclass of gimple_opt_pass along with...
12787 (pass_data_ch): ...new pass_data instance and...
12788 (make_pass_ch): ...new function.
12789 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
12790 subclass of gimple_opt_pass along with...
12791 (pass_data_tree_loop): ...new pass_data instance and...
12792 (make_pass_tree_loop): ...new function.
12793 (pass_tree_loop_init): Convert from a global struct to a subclass of
12794 gimple_opt_pass along with...
12795 (pass_data_tree_loop_init): ...new pass_data instance and...
12796 (make_pass_tree_loop_init): ...new function.
12797 (pass_lim): Convert from a global struct to a subclass of
12798 gimple_opt_pass along with...
12799 (pass_data_lim): ...new pass_data instance and...
12800 (make_pass_lim): ...new function.
12801 (pass_tree_unswitch): Convert from a global struct to a subclass of
12802 gimple_opt_pass along with...
12803 (pass_data_tree_unswitch): ...new pass_data instance and...
12804 (make_pass_tree_unswitch): ...new function.
12805 (pass_predcom): Convert from a global struct to a subclass of
12806 gimple_opt_pass along with...
12807 (pass_data_predcom): ...new pass_data instance and...
12808 (make_pass_predcom): ...new function.
12809 (pass_vectorize): Convert from a global struct to a subclass of
12810 gimple_opt_pass along with...
12811 (pass_data_vectorize): ...new pass_data instance and...
12812 (make_pass_vectorize): ...new function.
12813 (pass_graphite): Convert from a global struct to a subclass of
12814 gimple_opt_pass along with...
12815 (pass_data_graphite): ...new pass_data instance and...
12816 (make_pass_graphite): ...new function.
12817 (pass_graphite_transforms): Convert from a global struct to a subclass
12818 of gimple_opt_pass along with...
12819 (pass_data_graphite_transforms): ...new pass_data instance and...
12820 (make_pass_graphite_transforms): ...new function.
12821 (pass_check_data_deps): Convert from a global struct to a subclass of
12822 gimple_opt_pass along with...
12823 (pass_data_check_data_deps): ...new pass_data instance and...
12824 (make_pass_check_data_deps): ...new function.
12825 (pass_iv_canon): Convert from a global struct to a subclass of
12826 gimple_opt_pass along with...
12827 (pass_data_iv_canon): ...new pass_data instance and...
12828 (make_pass_iv_canon): ...new function.
12829 (pass_scev_cprop): Convert from a global struct to a subclass of
12830 gimple_opt_pass along with...
12831 (pass_data_scev_cprop): ...new pass_data instance and...
12832 (make_pass_scev_cprop): ...new function.
12833 (pass_record_bounds): Convert from a global struct to a subclass of
12834 gimple_opt_pass along with...
12835 (pass_data_record_bounds): ...new pass_data instance and...
12836 (make_pass_record_bounds): ...new function.
12837 (pass_complete_unroll): Convert from a global struct to a subclass of
12838 gimple_opt_pass along with...
12839 (pass_data_complete_unroll): ...new pass_data instance and...
12840 (make_pass_complete_unroll): ...new function.
12841 (pass_complete_unrolli): Convert from a global struct to a subclass of
12842 gimple_opt_pass along with...
12843 (pass_data_complete_unrolli): ...new pass_data instance and...
12844 (make_pass_complete_unrolli): ...new function.
12845 (pass_parallelize_loops): Convert from a global struct to a subclass
12846 of gimple_opt_pass along with...
12847 (pass_data_parallelize_loops): ...new pass_data instance and...
12848 (make_pass_parallelize_loops): ...new function.
12849 (pass_loop_prefetch): Convert from a global struct to a subclass of
12850 gimple_opt_pass along with...
12851 (pass_data_loop_prefetch): ...new pass_data instance and...
12852 (make_pass_loop_prefetch): ...new function.
12853 (pass_iv_optimize): Convert from a global struct to a subclass of
12854 gimple_opt_pass along with...
12855 (pass_data_iv_optimize): ...new pass_data instance and...
12856 (make_pass_iv_optimize): ...new function.
12857 (pass_tree_loop_done): Convert from a global struct to a subclass of
12858 gimple_opt_pass along with...
12859 (pass_data_tree_loop_done): ...new pass_data instance and...
12860 (make_pass_tree_loop_done): ...new function.
12861 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
12862 struct to a subclass of gimple_opt_pass along with...
12863 (pass_data_cse_reciprocals): ...new pass_data instance and...
12864 (make_pass_cse_reciprocals): ...new function.
12865 (pass_cse_sincos): Convert from a global struct to a subclass of
12866 gimple_opt_pass along with...
12867 (pass_data_cse_sincos): ...new pass_data instance and...
12868 (make_pass_cse_sincos): ...new function.
12869 (pass_optimize_bswap): Convert from a global struct to a subclass of
12870 gimple_opt_pass along with...
12871 (pass_data_optimize_bswap): ...new pass_data instance and...
12872 (make_pass_optimize_bswap): ...new function.
12873 (pass_optimize_widening_mul): Convert from a global struct to a
12874 subclass of gimple_opt_pass along with...
12875 (pass_data_optimize_widening_mul): ...new pass_data instance and...
12876 (make_pass_optimize_widening_mul): ...new function.
12877 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
12878 subclass of gimple_opt_pass along with...
12879 (pass_data_phiopt): ...new pass_data instance and...
12880 (make_pass_phiopt): ...new function.
12881 (pass_cselim): Convert from a global struct to a subclass of
12882 gimple_opt_pass along with...
12883 (pass_data_cselim): ...new pass_data instance and...
12884 (make_pass_cselim): ...new function.
12885 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
12886 subclass of gimple_opt_pass along with...
12887 (pass_data_phiprop): ...new pass_data instance and...
12888 (make_pass_phiprop): ...new function.
12889 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
12890 subclass of gimple_opt_pass along with...
12891 (pass_data_pre): ...new pass_data instance and...
12892 (make_pass_pre): ...new function.
12893 (pass_fre): Convert from a global struct to a subclass of
12894 gimple_opt_pass along with...
12895 (pass_data_fre): ...new pass_data instance and...
12896 (make_pass_fre): ...new function.
12897 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
12898 subclass of gimple_opt_pass along with...
12899 (pass_data_reassoc): ...new pass_data instance and...
12900 (make_pass_reassoc): ...new function.
12901 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
12902 subclass of gimple_opt_pass along with...
12903 (pass_data_sink_code): ...new pass_data instance and...
12904 (make_pass_sink_code): ...new function.
12905 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
12906 subclass of gimple_opt_pass along with...
12907 (pass_data_strlen): ...new pass_data instance and...
12908 (make_pass_strlen): ...new function.
12909 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
12910 struct to a subclass of gimple_opt_pass along with...
12911 (pass_data_build_alias): ...new pass_data instance and...
12912 (make_pass_build_alias): ...new function.
12913 (pass_build_ealias): Convert from a global struct to a subclass of
12914 gimple_opt_pass along with...
12915 (pass_data_build_ealias): ...new pass_data instance and...
12916 (make_pass_build_ealias): ...new function.
12917 (pass_ipa_pta): Convert from a global struct to a subclass of
12918 simple_ipa_opt_pass along with...
12919 (pass_data_ipa_pta): ...new pass_data instance and...
12920 (make_pass_ipa_pta): ...new function.
12921 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
12922 subclass of gimple_opt_pass along with...
12923 (pass_data_uncprop): ...new pass_data instance and...
12924 (make_pass_uncprop): ...new function.
12925 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
12926 global struct to a subclass of gimple_opt_pass along with...
12927 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
12928 (make_pass_late_warn_uninitialized): ...new function.
12929 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
12930 to a subclass of gimple_opt_pass along with...
12931 (pass_data_init_datastructures): ...new pass_data instance and...
12932 (make_pass_init_datastructures): ...new function.
12933 (pass_early_warn_uninitialized): Convert from a global struct to a
12934 subclass of gimple_opt_pass along with...
12935 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
12936 (make_pass_early_warn_uninitialized): ...new function.
12937 (pass_update_address_taken): Convert from a global struct to a
12938 subclass of gimple_opt_pass along with...
12939 (pass_data_update_address_taken): ...new pass_data instance and...
12940 (make_pass_update_address_taken): ...new function.
12941 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
12942 struct to a subclass of gimple_opt_pass along with...
12943 (pass_data_release_ssa_names): ...new pass_data instance and...
12944 (make_pass_release_ssa_names): ...new function.
12945 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
12946 subclass of gimple_opt_pass along with...
12947 (pass_data_stdarg): ...new pass_data instance and...
12948 (make_pass_stdarg): ...new function.
12949 * tree-switch-conversion.c (pass_convert_switch): Convert from a
12950 global struct to a subclass of gimple_opt_pass along with...
12951 (pass_data_convert_switch): ...new pass_data instance and...
12952 (make_pass_convert_switch): ...new function.
12953 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
12954 to a subclass of gimple_opt_pass along with...
12955 (pass_data_tail_recursion): ...new pass_data instance and...
12956 (make_pass_tail_recursion): ...new function.
12957 (pass_tail_calls): Convert from a global struct to a subclass of
12958 gimple_opt_pass along with...
12959 (pass_data_tail_calls): ...new pass_data instance and...
12960 (make_pass_tail_calls): ...new function.
12961 * tree-vect-generic.c (pass_lower_vector): Convert from a global
12962 struct to a subclass of gimple_opt_pass along with...
12963 (pass_data_lower_vector): ...new pass_data instance and...
12964 (make_pass_lower_vector): ...new function.
12965 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
12966 gimple_opt_pass along with...
12967 (pass_data_lower_vector_ssa): ...new pass_data instance and...
12968 (make_pass_lower_vector_ssa): ...new function.
12969 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
12970 to a subclass of gimple_opt_pass along with...
12971 (pass_data_slp_vectorize): ...new pass_data instance and...
12972 (make_pass_slp_vectorize): ...new function.
12973 (pass_ipa_increase_alignment): Convert from a global struct to a
12974 subclass of simple_ipa_opt_pass along with...
12975 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
12976 (make_pass_ipa_increase_alignment): ...new function.
12977 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
12978 gimple_opt_pass along with...
12979 (pass_data_vrp): ...new pass_data instance and...
12980 (make_pass_vrp): ...new function.
12981 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
12982 subclass of simple_ipa_opt_pass along with...
12983 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
12984 (make_pass_ipa_free_lang_data): ...new function.
12985 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
12986 gimple_opt_pass along with...
12987 (pass_data_tsan): ...new pass_data instance and...
12988 (make_pass_tsan): ...new function.
12989 (pass_tsan_O0): Convert from a global struct to a subclass of
12990 gimple_opt_pass along with...
12991 (pass_data_tsan_O0): ...new pass_data instance and...
12992 (make_pass_tsan_O0): ...new function.
12993 * var-tracking.c (pass_variable_tracking): Convert from a global
12994 struct to a subclass of rtl_opt_pass along with...
12995 (pass_data_variable_tracking): ...new pass_data instance and...
12996 (make_pass_variable_tracking): ...new function.
12997 * web.c (pass_web): Convert from a global struct to a subclass of
12998 rtl_opt_pass along with...
12999 (pass_data_web): ...new pass_data instance and...
13000 (make_pass_web): ...new function.
13001 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13002 declaration with that of...
13003 (make_pass_mode_switch_use): ...new function.
13004 (pass_resolve_sw_modes): Replace declaration with that of...
13005 (make_pass_resolve_sw_modes): ...new function.
13006 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13007 from a global struct to a subclass of rtl_opt_pass along with...
13008 (pass_data_mode_switch_use): ...new pass_data instance and...
13009 (make_pass_mode_switch_use): ...new function.
13010 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13011 from a global struct to a subclass of rtl_opt_pass along with...
13012 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13013 (make_pass_resolve_sw_modes): ...new function.
13014 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13015 struct to a subclass of rtl_opt_pass along with...
13016 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13017 (make_pass_insert_vzeroupper): ...new function.
13018 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13019 global struct to a subclass of rtl_opt_pass along with...
13020 (pass_data_work_around_errata): ...new pass_data instance and...
13021 (make_pass_work_around_errata): ...new function.
13022 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13023 struct to a subclass of rtl_opt_pass along with...
13024 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13025 (make_pass_mips_machine_reorg2): ...new function.
13027 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13029 * passes.c (pass_manager::operator new): New.
13031 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13033 Handwritten part of conversion of passes to C++ classes.
13035 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13036 (toplev.o): Add dep on PASS_MANAGER_H.
13037 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13038 of early local pases to reflect this moving from a global to a
13039 member of gcc::pass_manager.
13040 (cgraph_add_new_function): Likewise.
13041 * lto-cgraph.c (lto_output_node): Update for conversion of
13042 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13043 * passes.c (opt_pass::clone): New.
13044 (opt_pass::gate): New.
13045 (opt_pass::execute): New.
13046 (opt_pass::opt_pass): New.
13047 (pass_manager::execute_early_local_passes): New.
13048 (pass_manager::execute_pass_mode_switching): new.
13049 (finish_optimization_passes): Convert to...
13050 (pass_manager::finish_optimization_passes): ...this.
13051 (finish_optimization_passes): Update for conversion of passes to
13053 (register_dump_files_1): Use has_gate since we cannot portably
13054 check a vtable entry against NULL.
13055 (dump_one_pass): Likewise.
13056 (ipa_write_summaries_2): Likewise.
13057 (ipa_write_optimization_summaries_1): Likewise.
13058 (ipa_read_summaries_1): Likewise.
13059 (ipa_read_optimization_summaries_1): Likewise.
13060 (execute_ipa_stmt_fixups): Likewise.
13061 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13062 pass-creation functions rather than wiring up globals, and
13063 storing the results in fields of pass_manager generated using
13064 pass-instances.def.
13065 (pass_manager::dump_profile_report): Update for conversion of
13066 passes to C++ classes.
13067 (pass_manager::execute_ipa_summary_passes): Likewise.
13068 (execute_one_ipa_transform_pass): Likewise.
13069 (execute_one_pass): Use has_gate and has_execute since we cannot
13070 portably check a vtable entry against NULL.
13071 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13072 (pass_manager): Use pass-instances.def to add fields for the
13073 various pass instances.
13074 * toplev.c (finalize): Update for move of
13075 finish_optimization_passes to a method of gcc::pass_manager.
13076 * toplev.h (finish_optimization_passes): Move to method of class
13078 * tree-pass.h (struct pass_data): New.
13079 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13080 (opt_pass::gate): Convert to virtual function.
13081 (opt_pass::~opt_pass): New.
13082 (opt_pass::clone): New.
13083 (opt_pass::execute): Convert to virtual function.
13084 (opt_pass::opt_pass): New.
13085 (opt_pass::ctxt_): new.
13086 (gimple_opt_pass): Convert to subclass of opt_pass.
13087 (gimple_opt_pass::gimple_opt_pass): New.
13088 (rtl_opt_pass): Convert to subclass of opt_pass.
13089 (rtl_opt_pass::rtl_opt_pass): New.
13090 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13091 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13092 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13093 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13094 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13095 invocation of pass_mode_switching to reflect this moving from a
13096 global to a member of gcc::pass_manager.
13097 (ix86_option_override): Rework how pass_insert_vzeroupper is
13098 added to the pass_manager to reflect autogenerated changes.
13099 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13101 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13103 PR rtl-optimization/57708
13104 * recog.c (peep2_find_free_register): Validate all regs in a
13107 2013-08-05 Jan Hubicka <jh@suse.cz>
13110 * cgraph.c (verify_cgraph_node): Accept local flags from other
13112 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13113 (function_and_variable_visibility): Likewise.
13114 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13116 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13118 * graph.c (init_graph_slim_pretty_print): Remove.
13119 (print_graph_cfg): Do not call it. Use local pretty printer.
13120 (start_graph_dump): Likewise.
13122 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13124 * gimple-pretty-print.c (buffer): Remove.
13125 (initialized): Likewise.
13126 (maybe_init_pretty_print): Likewise.
13127 (print_gimple_stmt): Do not call it. Use non-static local
13128 pretty_printer variable.
13129 (print_gimple_expr): Likewise.
13130 (print_gimple_seq): Likewise.
13131 (gimple_dump_bb): Likewise.
13133 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13135 * asan.c (asan_pp): Remove.
13136 (asan_pp_initialized): Likewise.
13137 (asan_pp_initialize): Likewise.
13138 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13139 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13140 (asan_add_global): Likewise.
13142 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13144 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13145 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13146 * pretty-print.c (pp_formatted_text_data): Likewise.
13147 (pp_indent): Rename from pp_base_indent.
13148 (pp_format): Rename from pp_base_format.
13149 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13150 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13151 (pp_flush): Rename from pp_base_flush.
13152 (pp_set_line_maximum_length): Rename from
13153 pp_base_set_line_maximum_length.
13154 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13155 (pp_set_prefix): Rename from pp_base_set_prefix.
13156 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13157 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13158 (pp_append_text): Rename from pp_base_append_text.
13159 (pp_formatted_text): Rename from pp_base_formatted_text.
13160 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13161 (pp_remaining_character_count_for_line): Rename from
13162 pp_base_remaining_character_count_for_line.
13163 (pp_newline): Rename from pp_base_newline.
13164 (pp_character): Rename from pp_base_character.
13165 (pp_string): Rename from pp_base_string.
13166 (pp_maybe_space): Rename from pp_base_maybe_space.
13167 * asan.c (asan_pp_string): Adjust.
13168 (asan_emit_stack_protection): Likewise.
13169 (asan_add_global): Likewise.
13170 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13171 * tree-mudflap.c (mf_varname_tree): Likewise.
13172 * tree-pretty-print.c (pp_tree_identifier): Rename from
13173 pp_base_tree_identifier.
13174 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13175 Declare as function.
13177 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13179 * pretty-print.h (pp_bar_bar): New.
13180 (pp_ampersand_ampersand): Likewise.
13181 (pp_less_equal): Likewise.
13182 (pp_greater_equal): Likewise.
13183 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13184 printer functions instead of pp_string or operators and punctuators.
13185 (dump_gimple_call): Likewise.
13186 (dump_gimple_omp_for): Likewise.
13187 (dump_gimple_transaction): Likewise.
13188 (dump_gimple_phi): Likewise.
13189 (pp_gimple_stmt_1): Likewise.
13190 * sched-vis.c (print_insn): Likewise.
13191 * tree-mudflap.c (mf_varname_tree): Likewise.
13192 * tree-pretty-print.c (dump_block_node): Likewise.
13193 (dump_generic_node): Likewise.
13195 2013-08-02 Jan Hubicka <jh@suse.cz>
13197 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13199 * lto-streamer-out.c (tree_is_indexable): Results decls and
13200 parm decls are not indexable.
13201 (DFS_write_tree_body): Do not follow args and results.
13202 (hash_tree): Likewise.
13203 (output_functions): Rearrange so struct function is needed
13204 only when real body is output; be able to also ouptut abstract
13205 functions; output DECL_ARGUMENTS and DECL_RESULT.
13206 (lto_output): When not in WPA, ale store abstract functions.
13207 (write_symbol): Do not care about RESULT_DECL.
13208 (output_symbol_p): Handle correctly sbtract decls.
13209 * lto-streamer-in.c (input_function): Rearrange so struct
13210 function can be NULL at entry; allow streaming of
13211 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13212 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13213 sanity check during LTO.
13214 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13215 RESULT_DECl and DECL_ARGUMENTS.
13216 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
13219 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13221 * pretty-print.h (pp_underscore): New.
13223 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
13224 printer functions instead of pp_character.
13225 (dump_binary_rhs): Likewise.
13226 (dump_ternary_rhs): Likewise.
13227 (dump_gimple_call_args): Likewise.
13228 (pp_points_to_solution): Likewise.
13229 (dump_gimple_call): Likewise.
13230 (dump_gimple_switch): Likewise.
13231 (dump_gimple_cond): Likewise.
13232 (dump_gimple_bind): Likewise.
13233 (dump_gimple_try): Likewise.
13234 (dump_gimple_omp_for): Likewise.
13235 (dump_gimple_omp_continue): Likewise.
13236 (dump_gimple_omp_single): Likewise.
13237 (dump_gimple_omp_sections): Likewise.
13238 (dump_gimple_omp_block): Likewise.
13239 (dump_gimple_omp_critical): Likewise.
13240 (dump_gimple_transaction): Likewise.
13241 (dump_gimple_asm): Likewise.
13242 (dump_gimple_phi): Likewise.
13243 (dump_gimple_omp_parallel): Likewise.
13244 (dump_gimple_omp_task): Likewise.
13245 (dump_gimple_omp_atomic_load): Likewise.
13246 (dump_gimple_omp_atomic_store): Likewise.
13247 (dump_gimple_mem_ops): Likewise.
13248 (pp_gimple_stmt_1): Likewise.
13249 (pp_cfg_jump): Likewise.
13250 (dump_implicit_edges): Likewise.
13251 (gimple_dump_bb_for_graph): Likewise.
13252 * graph.c (draw_cfg_node): Likewise.
13253 * langhooks.c (lhd_print_error_function): Likewise.
13254 * sched-vis.c (print_exp): Likewise.
13255 (print_value): Likewise.
13256 (print_pattern): Likewise.
13257 (print_insn): Likewise.
13258 (rtl_dump_bb_for_graph): Likewise.
13259 * tree-pretty-print.c (dump_function_declaration): Likewise.
13260 (dump_array_domain): Likewise.
13261 (dump_omp_clause): Likewise.
13262 (dump_location): Likewise.
13263 (dump_generic_node): Likewise.
13264 (print_struct_decl): Likewise.
13265 * diagnostic.c (diagnostic_show_locus): Use pp_space.
13267 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13269 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
13270 candidate table when replacing a candidate statement.
13271 (replace_rhs_if_not_dup): Likewise.
13272 (replace_one_candidate): Likewise.
13274 2013-08-02 Jan Hubicka <jh@suse.cz>
13275 Martin Liska <marxin.liska@gmail.com>
13277 * cgraphunit.c (add_new_function): Fix logic when adding from
13279 (assemble_thunk): Rename to ...
13280 (expand_thunk); .. this one; export; get it working with
13281 general functions; make produced gimple valid.
13282 * cgraph.h (expand_thunk): Declare.
13284 2013-08-02 Jan Hubicka <jh@suse.cz>
13286 * ipa-cp.c (gather_context_independent_values): Use
13287 ipa_get_param_move_cost.
13288 (get_replacement_map): Remove PARAM; move parameter folding
13290 (create_specialized_node): Update.
13291 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
13292 assert that we have gimple body; update move_cost.
13293 (count_formal_params): Assert that we have gimple body.
13294 (ipa_dump_param): New function.
13295 (ipa_alloc_node_params): Break out from ...
13296 (ipa_initialize_node_params): ... here.
13297 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
13298 (ipa_write_node_info): Stream move costs.
13299 (ipa_read_node_info): Read move costs.
13300 (ipa_update_after_lto_read): Do not recompute node params.
13301 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
13302 (ipa_get_param): Check we are not in WPA.
13303 (ipa_get_param_move_cost): New.
13304 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
13305 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
13306 parm numbers to be present.
13308 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13310 PR rtl-optimization/58048
13311 * lra-constraints.c (process_alt_operands): Don't check asm
13312 operand on register.
13314 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
13316 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
13317 the implied StoreLoad barrier for atomic operations if before.
13319 2013-08-02 Jan Hubicka <jh@suse.cz>
13320 Martin Liska <marxin.liska@gmail.com>
13322 * cgraph.c (cgraph_function_body_availability): Do not check
13324 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
13325 symtab_node_availability): Declare.
13326 * ipa.c (can_replace_by_local_alias): New.
13327 (function_and_variable_visibility): Use it.
13328 * symtab.c (symtab_for_node_and_aliases,
13329 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
13331 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13333 PR rtl-optimization/57963
13334 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
13335 (lra_constraints): Use them.
13337 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
13339 * config/arm/types.md (define_attr "type"): Add "load_acq"
13341 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
13343 (cortex_a53_store1): Likewise.
13345 2013-08-01 Jan Hubicka <jh@suse.cz>
13347 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
13348 partitions are not needed.
13350 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
13352 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
13353 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
13354 MAYBE_NON_Q_CLASS_P where appropriate.
13356 2013-08-01 Jan Hubicka <jh@suse.cz>
13358 * cgraph.h (release_function_body): Declare.
13359 * tree.c (free_lang_data_in_decl): Free, parameters and return values
13360 of unused delcarations.
13362 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13364 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
13365 RTL form when subtracting a constant.
13367 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13369 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
13370 Generate canonical plus rtx with negated immediate instead of minus
13372 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
13374 2013-08-01 Jan Hubicka <jh@suse.cz>
13376 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
13377 (cgraph_release_function_body): Likewise.
13378 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
13379 * cgraph.h (cgrpah_node): Rename abstract_and_needed
13380 to used_as_abstract_origin.
13381 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
13382 symbols used as abstract origins.
13383 * cgraphunit.c (analyze_functions): Update.
13384 * ipa.c (symtab_remove_unreachable_nodes): Recompute
13385 used_as_abstract_origin.
13386 * tree-inline.c (tree_function_versioning): Update
13387 used_as_abstract_origin; be ready for DECL_RESULT and
13388 DECL_ARGUMENTS to be NULL.
13390 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
13391 for abstract functions.
13392 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
13395 2013-08-01 Jan Hubicka <jh@suse.cz>
13397 * profile.c (compute_value_histograms): Fix thinko.
13399 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13401 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
13402 aarch-common-protos.h to extra_headers.
13403 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
13404 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
13405 * config/aarch64/t-aarch64 (aarch-common.o): Define.
13407 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13409 * config/aarch64/aarch64.md (define_attr "type"): Delete.
13410 Include "../arm/types.md". Define "type" attribute for all patterns.
13411 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
13414 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
13416 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
13417 to support power8 load fusion.
13418 (fusion_gpr_mem_load): Likewise.
13420 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
13422 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
13423 declarations for power8 load fusion.
13424 (emit_fusion_gpr_load): Likewise.
13426 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
13427 tuning for power8, turn on fusion mode by default. Turn on sign
13428 extending fusion mode if normal fusion mode is on, and we are at
13430 (fusion_gpr_load_p): New function, return true if we can fuse an
13431 addis instruction with a dependent load to a GPR.
13432 (emit_fusion_gpr_load): Emit the instructions for power8 load
13435 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
13436 (VSX load fusion peepholes): New peepholes to fuse together an
13437 addi instruction with a VSX load instruction.
13439 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
13440 peepholes to fuse an addis instruction with a load to a GPR base
13441 register. If we are supporting sign extending fusions, convert
13442 sign extending loads to zero extending loads and add an explicit
13445 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
13447 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
13448 aarch-common-protos.h to extra_headers.
13449 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
13450 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
13451 (arm_early_store_addr_dep): Likewise.
13452 (arm_no_early_alu_shift_dep): Likewise.
13453 (arm_no_early_alu_shift_value_dep): Likewise.
13454 (arm_no_early_mul_dep): Likewise.
13455 (arm_no_early_store_addr_dep): Likewise.
13456 (arm_mac_accumulator_is_mul_result): Likewise.
13457 (arm_mac_accumulator_is_result): Likewise.
13458 * config/arm/aarch-common.c: ... here. New file.
13459 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
13461 (arm_early_store_addr_dep): Likewise.
13462 (arm_no_early_alu_shift_dep): Likewise.
13463 (arm_no_early_alu_shift_value_dep): Likewise.
13464 (arm_no_early_mul_dep): Likewise.
13465 (arm_no_early_store_addr_dep): Likewise.
13466 (arm_mac_accumulator_is_mul_result): Likewise.
13467 (arm_mac_accumulator_is_result): Likewise.
13468 * config/arm/aarch-common-protos.h: ... here. New file.
13469 * config/arm/t-arm (aarch-common.o): Define.
13471 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
13473 * config/arm/arm.md: Include new file "types.md".
13474 (define_attr "type"): Move from here to ...
13475 (define_attr "mul32"): Likewise.
13476 (define_attr "mul64"): Likewise.
13477 * config/arm/types.md: ... here. New file.
13479 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
13481 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
13482 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
13484 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13486 * gen-pass-instances.awk: Fix offset of substr().
13488 2013-07-31 David Malcolm <dmalcolm@redhat.com>
13490 * Makefile.in (pass-instances.def): New.
13491 (passes.o): Replace dependency on passes.def with one on
13494 * gen-pass-instances.awk: New.
13496 * passes.c (pass_manager::pass_manager): Use pass-instances.def
13497 rather than passes.def, updating local definition of NEXT_PASS
13498 macro to add an extra NUM parameter (currently unused).
13500 2013-07-30 David Malcolm <dmalcolm@redhat.com>
13502 * Makefile.in (PASS_MANAGER_H): New.
13503 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
13504 (passes.o): Likewise.
13505 (statistics.o): Likewise.
13506 (cgraphunit.o): Likewise.
13507 (context.o): Depend on PASS_MANAGER_H.
13509 * pass_manager.h: New.
13511 * cgraphunit.c (cgraph_add_new_function): Update for moves
13512 of globals to fields of pass_manager.
13513 (analyze_function): Likewise.
13514 (expand_function): Likewise.
13515 (ipa_passes): Likewise.
13516 (compile): Likewise.
13518 * context.c (context::context): New.
13519 * context.h (context::context): New.
13520 (context::get_passes): New.
13521 (context::passes_): New.
13523 * lto-cgraph.c (input_node): Update for moves of globals to
13524 fields of pass_manager.
13526 * passes.c (all_passes): Remove, in favor of a field of the
13527 same name within the new class pass_manager.
13528 (all_small_ipa_passes): Likewise.
13529 (all_lowering_passes): Likewise.
13530 (all_regular_ipa_passes): Likewise.
13531 (all_late_ipa_passes): Likewise.
13532 (all_lto_gen_passes): Likewise.
13533 (passes_by_id): Likewise.
13534 (passes_by_id_size): Likewise.
13535 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
13536 the new class pass_manager.
13537 (set_pass_for_id): Convert to...
13538 (pass_manager::set_pass_for_id): ...method.
13539 (get_pass_for_id): Convert to...
13540 (pass_manager::get_pass_for_id): ...method.
13541 (register_one_dump_file): Move body of implementation into...
13542 (pass_manager::register_one_dump_file): ...here.
13543 (register_dump_files_1): Convert to...
13544 (pass_manager::register_dump_files_1): ...method.
13545 (register_dump_files): Convert to...
13546 (pass_manager::register_dump_files): ...method.
13547 (create_pass_tab): Update for moves of globals to fields of
13549 (dump_passes): Move body of implementation into...
13550 (pass_manager::dump_passes): ...here.
13551 (register_pass): Move body of implementation into...
13552 (pass_manager::register_pass): ...here.
13553 (init_optimization_passes): Convert into...
13554 (pass_manager::pass_manager): ...constructor for new
13555 pass_manager class, and initialize the pass_lists array.
13556 (check_profile_consistency): Update for moves of globals to
13557 fields of pass_manager.
13558 (dump_profile_report): Move body of implementation into...
13559 (pass_manager::dump_profile_report): ...here.
13560 (ipa_write_summaries_1): Update for moves of pass lists from
13561 being globals to fields of pass_manager.
13562 (ipa_write_optimization_summaries): Likewise.
13563 (ipa_read_summaries): Likewise.
13564 (ipa_read_optimization_summaries): Likewise.
13565 (execute_all_ipa_stmt_fixups): Likewise.
13567 * statistics.c (statistics_fini): Update for moves of globals to
13568 fields of pass_manager.
13570 * toplev.c (general_init): Replace call to
13571 init_optimization_passes with construction of the pass_manager
13574 * tree-pass.h (all_passes): Remove, in favor of a field of the
13575 same name within the new class pass_manager.
13576 (all_small_ipa_passes): Likewise.
13577 (all_lowering_passes): Likewise.
13578 (all_regular_ipa_passes): Likewise.
13579 (all_lto_gen_passes): Likewise.
13580 (all_late_ipa_passes): Likewise.
13581 (passes_by_id): Likewise.
13582 (passes_by_id_size): Likewise.
13583 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
13584 the new class pass_manager.
13585 (get_pass_for_id): Remove.
13587 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
13589 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
13592 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
13594 * arm.md (mulhi3): New expand pattern.
13596 2013-07-30 Jan Hubicka <jh@suse.cz>
13597 Martin Liska <marxin.liska@gmail.com>
13599 * profile.c (compute_value_histograms): Do not ICE when
13600 there is mismatch only on some counters.
13602 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13604 PR rtl-optimization/57637
13605 * function.c (move_insn_for_shrink_wrap): Also check the
13606 GEN set of the LIVE problem for the liveness analysis
13607 if it exists, otherwise give up.
13609 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13611 PR tree-optimization/57993
13612 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
13613 replaced statement in the candidate table.
13614 (phi_add_costs): Return infinite cost when the hidden basis does
13615 not dominate all phis on which the candidate is dependent.
13616 (replace_one_candidate): Record replaced statement in the
13619 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
13621 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
13622 (ashlv2si3): New expander.
13623 (*ashlv2si3_i): New define_insn_and_split.
13624 * predicates.md (float_operation): Allow patterns with three
13625 basic sub-patterns.
13627 PR rtl-optimization/58021
13628 * mode-switching.c (create_pre_exit): Always split off preceding
13629 insns if we are not at the basic block head.
13631 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
13633 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
13634 (UCLIBC_DYNAMIC_LINKER): New macro.
13635 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
13637 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
13638 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
13640 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
13641 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
13642 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
13643 for SF and DF modes. Use ieee_quad_format for TF mode.
13644 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
13645 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
13646 (mips_option_override): Handle `-mnan=legacy'.
13647 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
13648 `-mabs=2008' and `-mnan=2008'.
13649 (OPTION_DEFAULT_SPECS): Add "nan" default.
13650 (ASM_SPEC): Handle `-mnan='.
13651 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
13652 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
13653 comment accordingly.
13654 (neg<mode>2): Likewise.
13655 * config/mips/mips.opt (mabs, mnan): New options.
13656 * doc/install.texi (Configuration): Document `--with-nan=' option.
13657 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
13659 (MIPS Options): Document them.
13660 * config.gcc <mips*-*-*>: Handle `--with-nan='.
13661 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
13662 * configure: Regenerate.
13663 * config.in: Regenerate.
13665 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
13667 * config/i386/i386.md (float post-reload splitters): Do not check
13668 for subregs of SSE registers.
13670 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
13671 H.J. Lu <hongjiu.lu@intel.com>
13675 * config/i386/i386.md (post-reload splitter
13676 to avoid partial SSE reg dependency stalls): New pattern.
13678 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
13680 * config/s390/s390.md ("movcc"): Swap load and store instructions.
13682 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
13684 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
13685 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
13687 2013-07-26 Cary Coutant <ccoutant@google.com>
13689 * dwarf2out.c (die_checksum_ordered): Don't include template
13690 instantiations in signature.
13691 (is_template_parameter): New function.
13692 (is_template_instantiation): New function.
13693 (generate_skeleton_bottom_up): Don't include template instantiations
13695 (generate_skeleton): Likewise.
13696 (break_out_comdat_types): Move recursive call to break out nested
13698 (prune_unused_types_mark_generic_parms_dies): Call
13699 is_template_parameter.
13701 2013-07-26 Ian Bolton <ian.bolton@arm.com>
13703 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
13704 uses vector registers.
13705 * config/aarch64/iterators.md: Add attributes rtn and vas.
13707 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13708 Richard Earnshaw <richard.earnshaw@arm.com>
13710 * combine.c (simplify_comparison): Re-canonicalize operands
13712 * config/arm/arm.md (movcond_addsi): New splitter.
13714 2013-07-25 Sterling Augustine <saugustine@google.com>
13716 * dwarf2out.c (size_of_pubnames): Move code to...
13717 (include_pubname_in_output): ...here. New.
13718 (want_pubnames): Rearrange.
13719 (output_pubnames): Call include_pubname_in_output. Move assertion.
13721 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
13723 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
13725 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
13728 * doc/extend.texi: Remove obsolete builtins. Fix
13729 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
13731 2013-07-25 Jan Hubicka <jh@suse.cz>
13733 * cgraph.c (release_function_body): Break out from ...
13734 (cgraph_release_function_body): ... this one; also release DECL_RESULT
13735 and DECL_ARGUMENTS.
13736 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
13737 old_tree in the map.
13738 (create_specialized_node): Update.
13739 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
13741 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
13742 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
13743 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
13744 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
13747 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13749 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
13750 addsi3_carryin_alt2_<optab>): Correct output template.
13752 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13754 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
13755 Adjust for arm_restrict_it.
13756 Remove trailing whitespace.
13758 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
13760 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
13761 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
13763 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
13765 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
13767 PR rtl-optimization/57960
13768 * lra-constraints.c (process_alt_operands): Use the right mode
13769 when checking strict_low.
13771 2013-07-25 Jan Hubicka <jh@suse.cz>
13773 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
13774 * cgraph.c (cgraph_remove_node): Do not release function body
13775 when in cgraph streaming.
13776 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
13777 in other partitions are not considered reachable; fix handling of
13780 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13782 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
13784 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13789 * config/arm/arm.md ("*sibcall_insn): Replace use of
13790 Ss with US. Adjust output for v5 and v4t.
13791 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
13793 * config/arm/constraints.md ("Ss"): Rename to US.
13795 2013-07-25 Terry Guo <terry.guo@arm.com>
13797 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
13798 shift_add/shift_sub0/shift_sub1 RTXs.
13800 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
13801 Anton Blanchard <anton@au1.ibm.com>
13803 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
13804 (altivec_vpks<VI_char>ss): Likewise.
13805 (altivec_vpks<VI_char>us): Likewise.
13806 (altivec_vpku<VI_char>us): Likewise.
13807 (altivec_vpku<VI_char>um): Likewise.
13809 2013-07-24 David Malcolm <dmalcolm@redhat.com>
13811 Introduce context class.
13813 * Makefile.in (CONTEXT_H): New.
13814 (OBJS): Add context.o.
13815 (toplev.o): Add CONTEXT_H to dependencies.
13818 * toplev.c (general_init): Create the singleton gcc::context instance.
13824 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
13826 PR rtl-optimization/57968
13827 * mode-switching.c (create_pre_exit): Allow instructions that
13828 don't set a return register to need a non-exit mode.
13830 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13831 Anton Blanchard <anton@au1.ibm.com>
13833 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
13834 operands to vperm for little endian.
13835 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
13836 of lvsl to create the control mask for a vperm for little endian.
13838 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13839 Anton Blanchard <anton@au1.ibm.com>
13841 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
13842 two operands for little-endian.
13844 2013-07-23 Steve Ellcey <sellcey@mips.com>
13846 * config/mips/mips.c (mips_case_values_threshold): New.
13847 (TARGET_CASE_VALUES_THRESHOLD): Define.
13849 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13850 Anton Blanchard <anton@au1.ibm.com>
13852 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
13853 selection of field for vector splat in little endian mode.
13855 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
13857 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
13858 expanders to rs6000.md.
13859 (ior<mode>3): Likewise.
13860 (and<mode>3): Likewise.
13861 (one_cmpl<mode>2): Likewise.
13862 (nor<mode>3): Likewise.
13863 (andc<mode>3): Likewise.
13864 (eqv<mode>3): Likewise.
13865 (nand<mode>3): Likewise.
13866 (orc<mode>3): Likewise.
13868 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
13871 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
13872 to split multi-word logical operations.
13873 (rs6000_split_logical_di): Likewise.
13874 (rs6000_split_logical): Likewise.
13876 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
13877 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
13878 and allow TImode operations in 32-bit.
13879 (vsx_and<mode>3_64bit): Likewise.
13880 (vsx_ior<mode>3_32bit): Likewise.
13881 (vsx_ior<mode>3_64bit): Likewise.
13882 (vsx_xor<mode>3_32bit): Likewise.
13883 (vsx_xor<mode>3_64bit): Likewise.
13884 (vsx_one_cmpl<mode>2_32bit): Likewise.
13885 (vsx_one_cmpl<mode>2_64bit): Likewise.
13886 (vsx_nor<mode>3_32bit): Likewise.
13887 (vsx_nor<mode>3_64bit): Likewise.
13888 (vsx_andc<mode>3_32bit): Likewise.
13889 (vsx_andc<mode>3_64bit): Likewise.
13890 (vsx_eqv<mode>3_32bit): Likewise.
13891 (vsx_eqv<mode>3_64bit): Likewise.
13892 (vsx_nand<mode>3_32bit): Likewise.
13893 (vsx_nand<mode>3_64bit): Likewise.
13894 (vsx_orc<mode>3_32bit): Likewise.
13895 (vsx_orc<mode>3_64bit): Likewise.
13897 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
13898 logical types in GPRs.
13900 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
13901 logical insns to rs6000.md, and allow TImode operations in
13903 (altivec_ior<mode>3): Likewise.
13904 (altivec_xor<mode>3): Likewise.
13905 (altivec_one_cmpl<mode>2): Likewise.
13906 (altivec_nor<mode>3): Likewise.
13907 (altivec_andc<mode>3): Likewise.
13909 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
13910 attributes for moving the 128-bit logical operations into
13912 (BOOL_REGS_OUTPUT): Likewise.
13913 (BOOL_REGS_OP1): Likewise.
13914 (BOOL_REGS_OP2): Likewise.
13915 (BOOL_REGS_UNARY): Likewise.
13916 (BOOL_REGS_AND_CR0): Likewise.
13917 (one_cmpl<mode>2): Add support for DI logical operations on
13918 32-bit, splitting the operations to 32-bit.
13919 (anddi3): Likewise.
13920 (iordi3): Likewise.
13921 (xordi3): Likewise.
13922 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
13923 changes to combine the 32/64-bit code, allow logical operations on
13924 TI mode in 32-bit, and to use similar match_operator patterns like
13925 scalar mode uses. Combine the Altivec and VSX code for logical
13926 operations, and move it here.
13927 (ior<mode>3, 128-bit types): Likewise.
13928 (xor<mode>3, 128-bit types): Likewise.
13929 (one_cmpl<mode>3, 128-bit types): Likewise.
13930 (nor<mode>3, 128-bit types): Likewise.
13931 (andc<mode>3, 128-bit types): Likewise.
13932 (eqv<mode>3, 128-bit types): Likewise.
13933 (nand<mode>3, 128-bit types): Likewise.
13934 (orc<mode>3, 128-bit types): Likewise.
13935 (and<mode>3_internal): Likewise.
13936 (bool<mode>3_internal): Likewise.
13937 (boolc<mode>3_internal1): Likewise.
13938 (boolc<mode>3_internal2): Likewise.
13939 (boolcc<mode>3_internal1): Likewise.
13940 (boolcc<mode>3_internal2): Likewise.
13941 (eqv<mode>3_internal1): Likewise.
13942 (eqv<mode>3_internal2): Likewise.
13943 (one_cmpl1<mode>3_internal): Likewise.
13945 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
13947 * config/microblaze/microblaze.c (microblaze_expand_prologue):
13948 Rename flag_stack_usage to flag_stack_usage_info.
13950 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
13952 * config/microblaze/sync.md: New file.
13953 * config/microblaze/microblaze.md: Include sync.md
13954 * config/microblaze/microblaze.c: Add print_operand 'y'.
13955 * config/microblaze/constraints.md: Add memory_contraint
13956 'Q' which is a single register.
13958 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
13960 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
13962 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
13964 * reload.c (find_reloads): Exit loop once we find this operand
13965 cannot be reloaded somehow for this alternative.
13967 * reload.c (find_reloads): Exit loop once we find a hard register.
13969 * rtlanal.c (computed_jump_p): Exit loop once we find label
13972 * i386.c (ix86_pad_returns): Exit loop after setting replace.
13974 * cfgloopmanip.c (remove_path): Exit loop after setting
13977 * gensupport.c (subst_dup): Avoid loop if code is not
13978 MATCH_DUP nor MATCH_OP_DUP.
13980 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
13982 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
13984 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13986 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
13987 true for SP_REGNUM if mode == ptr_mode.
13988 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
13989 with value R0_REGNUM + 31.
13991 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13993 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
13994 pad pointer-typed argument downward.
13996 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13998 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
13999 and __ILP32__ when the ILP32 model is in use.
14001 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14003 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14004 (aarch64_load_symref_appropriately): In the case of
14005 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14006 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14007 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14008 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14009 if 'mode' doesn't equal to 'ptr_mode'.
14010 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14011 'vcall_offset'; change to call aarch64_emit_move differently depending
14012 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14013 to calculate the upper bound of 'vcall_offset'.
14014 (aarch64_cannot_force_const_mem): Change to also return true if
14016 (aarch64_legitimize_reload_address): In the case of large
14017 displacements, add new local variable 'xmode' and an assertion
14018 based on it; change to use 'xmode' to generate the new rtx and
14020 (aarch64_asm_trampoline_template): Change to generate the template
14021 differently depending on TARGET_ILP32 or not; change to use
14022 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14023 (aarch64_trampoline_size): Removed.
14024 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14025 and replace immediate literals with it. Change to use 'ptr_mode'
14026 instead of 'DImode' and call convert_memory_address if the mode
14027 of 'fnaddr' doesn't equal to 'ptr_mode'.
14028 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14030 (aarch64_elf_asm_destructor): Likewise.
14031 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14032 on TARGET_ILP32 instead of aarch64_trampoline_size.
14033 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14034 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14035 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14036 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14037 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14038 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14039 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14040 depending on the value of 'mode'.
14041 (add_losym_<mode>): New.
14042 (ldr_got_small_<mode>): New, based on ldr_got_small.
14043 (ldr_got_small): Remove.
14044 (ldr_got_small_sidi): New.
14045 * config/aarch64/iterators.md (P): New.
14046 (PTR): Change to 'ptr_mode' in the condition.
14048 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14050 * config.gcc (aarch64*-*-*): Support --with-abi.
14051 (aarch64*-*-elf): Support --with-multilib-list.
14052 (aarch64*-*-linux*): Likewise.
14053 (supported_defaults): Add abi to aarch64*-*-*.
14054 * configure.ac: Mention AArch64 for --with-multilib-list.
14055 * configure: Re-generated.
14056 * config/aarch64/biarchilp32.h: New file.
14057 * config/aarch64/biarchlp64.h: New file.
14058 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14060 (MULTILIB_DEFAULTS): Ditto.
14061 (DRIVER_SELF_SPECS): Ditto.
14062 (ASM_SPEC): Update to also substitute -mabi.
14063 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14064 file whose name depends on -mabi= and -mbig-endian.
14065 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14067 (POINTER_SIZE): New define.
14068 (POINTERS_EXTEND_UNSIGNED): Ditto.
14069 (enum aarch64_abi_type): New enumeration tag.
14070 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14071 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14072 (TARGET_ILP32): New define.
14073 * config/aarch64/aarch64.opt (mabi): New.
14074 (aarch64_abi): New.
14075 (ilp32, lp64): New values for -mabi.
14076 * config/aarch64/t-aarch64 (comma): New define.
14077 (MULTILIB_OPTIONS): Ditto.
14078 (MULTILIB_DIRNAMES): Ditto.
14079 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14080 * doc/invoke.texi: Document -mabi for AArch64.
14082 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14084 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14086 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14087 Anton Blanchard <anton@au1.ibm.com>
14089 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14090 endianness when selecting field to splat.
14092 2013-07-22 Eric Christopher <echristo@gmail.com>
14094 * dwarf2out.c (die_odr_checksum): New function to use
14095 CHECKSUM_ macros and ULEB128 for DIE tag.
14096 (generate_type_signature): Use.
14098 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14100 * config.gcc (sparc*-*-*): Accept leon3 processor.
14101 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14102 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14103 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14104 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14105 (mfix-ut699): Adjust comment.
14106 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14107 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14108 (CPP_CPU_SPEC): Likewise.
14109 (ASM_CPU_SPEC): Likewise.
14110 * config/sparc/sparc.c (leon3_cost): New constant.
14111 (sparc_option_override): Add leon3 support.
14112 (mem_ref): New function.
14113 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14114 (sparc_do_work_around_errata): Look into the instruction in the delay
14115 slot and adjust accordingly. Add fix for the data cache nullify issues
14116 of the UT699. Change insertion position for the NOP.
14117 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14118 (leon3_load): New reservation.
14119 (leon_store): Bump latency to 2.
14120 (grfpu): New automaton.
14121 (grfpu_alu): New unit.
14122 (grfpu_ds): Likewise.
14123 (leon_fp_alu): Adjust.
14124 (leon_fp_mult): Delete.
14125 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14126 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14127 * config/sparc/sparc.md (cpu): Add leon3.
14128 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14129 (swapsi): Likewise.
14130 (atomic_test_and_set): Likewise.
14131 (ldstub): Likewise.
14133 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14135 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14136 default for R5900 targets.
14137 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14138 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14139 * config/mips/mips.c (mips_option_override): Report an error for
14140 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14141 for -march=r5900 -mhard-float.
14143 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14145 * df-problems.c (can_move_insns_across): Exit loop once we
14146 find a non-fixed, non-global register.
14148 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14151 * omega.c (omega_eliminate_red): Break after setting red_found.
14152 (omega_problem_has_red_equations): Similarly after setting found.
14153 (omega_query_variable): Similarly after setting coupled.
14155 2013-07-22 Marek Polacek <polacek@redhat.com>
14157 * gimplify.c: Don't include gimple.h twice.
14159 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14161 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14162 instead of TARGET_THUMB1.
14163 (Pz): New constraint.
14164 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14166 (compare_negsi_si): Likewise.
14167 (compare_addsi2_op0): Likewise.
14168 (compare_addsi2_op1): Likewise.
14169 (addsi3_carryin_<optab>): Likewise.
14170 (addsi3_carryin_alt2_<optab>): Likewise.
14171 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14172 for arm_restrict_it.
14173 (subsi3_carryin): Likewise.
14174 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14175 (minmax_arithsi): Disable for arm_restrict_it.
14176 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14177 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14178 (satsi_<SAT:code>_shift): Likewise.
14179 (arm_shiftsi3): Add alternative for 16-bit encoding.
14180 (arm32_movhf): Disable for arm_restrict_it.
14181 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14182 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14184 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14186 * config/arm/arm.md (attribute "insn"): Delete.
14187 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14188 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14189 (not_shiftsi): Update for attribute change.
14190 (not_shiftsi_compare0): Likewise.
14191 (not_shiftsi_compare0_scratch): Likewise.
14192 (arm_one_cmplsi2): Likewise.
14193 (thumb1_one_cmplsi2): Likewise.
14194 (notsi_compare0): Likewise.
14195 (notsi_compare0_scratch): Likewise.
14196 (thumb1_movdi_insn): Likewise.
14197 (arm_movsi_insn): Likewise.
14198 (movhi_insn_arch4): Likewise.
14199 (movhi_bytes): Likewise.
14200 (arm_movqi_insn): Likewise.
14201 (thumb1_movqi_insn): Likewise.
14202 (arm32_movhf): Likewise.
14203 (thumb1_movhf): Likewise.
14204 (arm_movsf_soft_insn): Likewise.
14205 (thumb1_movsf_insn): Likewise.
14206 (thumb_movdf_insn): Likewise.
14207 (movsicc_insn): Likewise.
14208 (movsfcc_soft_insn): Likewise.
14209 (and_scc): Likewise.
14210 (cond_move): Likewise.
14211 (if_move_not): Likewise.
14212 (if_not_move): Likewise.
14213 (if_shift_move): Likewise.
14214 (if_move_shift): Likewise.
14215 (if_shift_shift): Likewise.
14216 (if_not_arith): Likewise.
14217 (if_arith_not): Likewise.
14218 (cond_move_not): Likewise.
14219 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
14220 (neon_mov<mode>): Likewise.
14221 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
14222 (thumb2_movsi_vfp): Likewise.
14223 (movsf_vfp): Likewise.
14224 (thumb2_movsf_vfp): Likewise.
14225 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14227 (cortexa7_older_only): Likewise.
14228 (cortexa7_younger): Likewise.
14229 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
14230 (1020alu_shift_op): Likewise.
14231 (1020alu_shift_reg_op): Likewise.
14232 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
14233 (alu_shift_op): Likewise.
14234 (alu_shift_reg_op): Likewise.
14235 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
14236 (11_alu_shift_op): Likewise.
14237 (11_alu_shift_reg_op): Likewise.
14238 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
14239 (9_alu_shift_reg_op): Likewise.
14240 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14242 (cortex_a15_alu_shift): Likewise.
14243 (cortex_a15_alu_shift_reg): Likewise.
14244 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
14245 (cortex_a5_alu_shift): Likewise.
14246 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
14248 (cortex_a53_alu_shift): Likewise.
14249 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14251 (cortex_a7_alu_reg): Likewise.
14252 (cortex_a7_alu_shift): Likewise.
14253 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14254 (cortex_a8_alu_shift): Likewise.
14255 (cortex_a8_alu_shift_reg): Likewise.
14256 (cortex_a8_mov): Likewise.
14257 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
14258 (cortex_a9_dp_shift): Likewise.
14259 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
14260 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
14261 (cortex_r4_mov): Likewise.
14262 (cortex_r4_alu_shift): Likewise.
14263 (cortex_r4_alu_shift_reg): Likewise.
14264 * config/arm/fa526.md (526_alu_op): Update for attribute change.
14265 (526_alu_shift_op): Likewise.
14266 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
14267 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
14268 (626te_alu_shift_op): Likewise.
14269 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
14270 (726te_alu_op): Likewise.
14271 (726te_alu_shift_op): Likewise.
14272 (726te_alu_shift_reg_op): Likewise.
14273 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
14274 (mp626_alu_shift_op): Likewise.
14275 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
14276 (pj4_alu_e1_conds): Likewise.
14277 (pj4_alu): Likewise.
14278 (pj4_alu_conds): Likewise.
14279 (pj4_shift): Likewise.
14280 (pj4_shift_conds): Likewise.
14281 (pj4_alu_shift): Likewise.
14282 (pj4_alu_shift_conds): Likewise.
14284 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14286 * config/arm/predicates.md (shiftable_operator_strict_it):
14288 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
14289 Disable cond_exec version for arm_restrict_it.
14290 (thumb2_smaxsi3): Convert to generate cond_exec.
14291 (thumb2_sminsi3): Likewise.
14292 (thumb32_umaxsi3): Likewise.
14293 (thumb2_uminsi3): Likewise.
14294 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
14295 (thumb2_neg_abssi2): Likewise.
14296 (thumb2_mov_scc): Add alternative for 16-bit encoding.
14297 (thumb2_movsicc_insn): Adjust alternatives.
14298 (thumb2_mov_negscc): Disable for arm_restrict_it.
14299 (thumb2_mov_negscc_strict_it): New pattern.
14300 (thumb2_mov_notscc_strict_it): New pattern.
14301 (thumb2_mov_notscc): Disable for arm_restrict_it.
14302 (thumb2_ior_scc): Likewise.
14303 (thumb2_ior_scc_strict_it): New pattern.
14304 (thumb2_cond_move): Adjust for arm_restrict_it.
14305 (thumb2_cond_arith): Disable for arm_restrict_it.
14306 (thumb2_cond_arith_strict_it): New pattern.
14307 (thumb2_cond_sub): Adjust for arm_restrict_it.
14308 (thumb2_movcond): Likewise.
14309 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
14310 (thumb2_zero_extendhisi2_v6): Likewise.
14311 (thumb2_zero_extendqisi2_v6): Likewise.
14312 (orsi_notsi_si): Likewise.
14313 (orsi_not_shiftsi_si): Likewise.
14315 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
14317 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
14318 instruction sequence is 1 byte shorter.
14320 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
14322 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
14323 it is not needed after split.
14325 2013-07-20 Iain Sandoe <iain@codesourcery.com>
14328 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
14329 second label for nonlocal goto receivers. Don't output pic base labels
14330 unless we're producing PIC; mark that action unreachable().
14331 (ix86_save_reg): If the function contains a nonlocal label, save the
14333 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
14334 * config/darwin.c (emitted_pic_label_num): New GTY.
14335 (update_pic_label_number_if_needed): New.
14336 (machopic_output_function_base_name): Adjust for nonlocal receiver
14338 (machopic_should_output_picbase_label): New.
14339 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
14340 (nonlocal_goto_receiver): New insn and split.
14342 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
14344 * config/aarch64/aarch64-builtins.c
14345 (aarch64_fold_builtin): Fold abs in all modes.
14346 * config/aarch64/aarch64-simd-builtins.def
14347 (abs): Enable for all modes.
14348 * config/aarch64/arm_neon.h
14349 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
14350 (vabs_f64): Add missing intrinsic.
14352 2013-07-19 Ian Bolton <ian.bolton@arm.com>
14354 * config/aarch64/arm_neon.h (vabs_s64): New function
14356 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
14359 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
14360 * config/avr/avr.md (adjust_len): Add `round'.
14361 * config/avr/avr-protos.h (avr_out_round): New prototype.
14362 (avr_out_plus): Add `out_label' argument.
14363 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
14364 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
14365 Handle the case where `insn' is just a pattern.
14366 (avr_out_bitop): Handle the case where `insn' is just a pattern.
14367 (avr_out_round): New function.
14368 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
14370 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
14372 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14373 Add check for flag_stack_usage to handle -fstack-usage support
14375 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
14377 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
14378 interaction for new Power8 flags and VSX.
14380 2013-07-18 Sriraman Tallam <tmsriram@google.com>
14382 PR middle-end/57698
14383 * tree-inline.c (expand_call_inline): Emit errors during
14384 early_inlining only if optimization is not turned on.
14386 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14390 * passes.c (init_optimization_passes): Move the construction of
14391 the pass hierarchy into a new passes.def file.
14393 * Makefile.in (passes.o): Add dependency on passes.def.
14395 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14397 * passes.c (init_optimization_passes): Introduce macros for
14398 constructing the tree of passes (INSERT_PASSES_AFTER,
14399 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
14400 TERMINATE_PASS_LIST).
14402 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
14403 Wei Mi <wmi@google.com>
14405 PR rtl-optimization/57878
14406 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
14408 (reload_pseudo_compare_func): Check nregs first for reload
14411 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14413 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
14415 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
14417 * read-rtl.c (validate_const_int): Once an invalid character is
14418 seen, quit the loop.
14420 * gengtype.c (write_roots): Similarly once we find the "deletable"
14421 or "if_marked" option.
14423 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14425 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
14426 "xtab" and "sat". Move value "clz" from here to ...
14427 (attriubte "type"): ... here.
14428 (satsi_<SAT:code>): Delete "insn" attribute.
14429 (satsi_<SAT:code>_shift): Likewise.
14430 (arm_zero_extendqisi2addsi): Likewise.
14431 (arm_extendqisi2addsi): Likewise.
14432 (clzsi2): Update for attribute changes.
14433 (rbitsi2): Likewise.
14434 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
14436 (arm_usatsihi): Likewise.
14437 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14439 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14441 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
14442 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
14443 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
14444 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
14445 in alphabetical order.
14446 (attribute "core_cycles"): Update for attribute changes.
14447 (arm_addsi3): Likewise.
14448 (addsi3_compare0): Likewise.
14449 (addsi3_compare0_scratch): Likewise.
14450 (addsi3_compare_op1): Likewise.
14451 (addsi3_compare_op2): Likewise.
14452 (compare_addsi2_op0): Likewise.
14453 (compare_addsi2_op1): Likewise.
14454 (addsi3_carryin_shift_<optab>): Likewise.
14455 (subsi3_carryin_shift): Likewise.
14456 (rsbsi3_carryin_shift): Likewise.
14457 (arm_subsi3_insn): Likewise.
14458 (subsi3_compare0): Likewise.
14459 (subsi3_compare): Likewise.
14460 (arm_andsi3_insn): Likewise.
14461 (thumb1_andsi3_insn): Likewise.
14462 (andsi3_compare0): Likewise.
14463 (andsi3_compare0_scratch): Likewise.
14464 (zeroextractsi_compare0_scratch
14465 (andsi_not_shiftsi_si): Likewise.
14466 (iorsi3_insn): Likewise.
14467 (iorsi3_compare0): Likewise.
14468 (iorsi3_compare0_scratch): Likewise.
14469 (arm_xorsi3): Likewise.
14470 (thumb1_xorsi3_insn): Likewise.
14471 (xorsi3_compare0): Likewise.
14472 (xorsi3_compare0_scratch): Likewise.
14473 (satsi_<SAT:code>_shift): Likewise.
14475 (arm_shiftsi3): Likewise.
14476 (shiftsi3_compare0): Likewise.
14477 (not_shiftsi): Likewise.
14478 (not_shiftsi_compare0): Likewise.
14479 (not_shiftsi_compare0_scratch): Likewise.
14480 (arm_one_cmplsi2): Likewise.
14481 (thumb_one_complsi2): Likewise.
14482 (notsi_compare0): Likewise.
14483 (notsi_compare0_scratch): Likewise.
14484 (thumb1_zero_extendhisi2): Likewise.
14485 (arm_zero_extendhisi2): Likewise.
14486 (arm_zero_extendhisi2_v6): Likewise.
14487 (arm_zero_extendhisi2addsi): Likewise.
14488 (thumb1_zero_extendqisi2): Likewise.
14489 (thumb1_zero_extendqisi2_v6): Likewise.
14490 (arm_zero_extendqisi2): Likewise.
14491 (arm_zero_extendqisi2_v6): Likewise.
14492 (arm_zero_extendqisi2addsi): Likewise.
14493 (thumb1_extendhisi2): Likewise.
14494 (arm_extendhisi2): Likewise.
14495 (arm_extendhisi2_v6): Likewise.
14496 (arm_extendqisi): Likewise.
14497 (arm_extendqisi_v6): Likewise.
14498 (arm_extendqisi2addsi): Likewise.
14499 (thumb1_extendqisi2): Likewise.
14500 (thumb1_movdi_insn): Likewise.
14501 (arm_movsi_insn): Likewise.
14502 (movsi_compare0): Likewise.
14503 (movhi_insn_arch4): Likewise.
14504 (movhi_bytes): Likewise.
14505 (arm_movqi_insn): Likewise.
14506 (thumb1_movqi_insn): Likewise.
14507 (arm32_movhf): Likewise.
14508 (thumb1_movhf): Likewise.
14509 (arm_movsf_soft_insn): Likewise.
14510 (thumb1_movsf_insn): Likewise.
14511 (movdf_soft_insn): Likewise.
14512 (thumb_movdf_insn): Likewise.
14513 (arm_cmpsi_insn): Likewise.
14514 (cmpsi_shiftsi): Likewise.
14515 (cmpsi_shiftsi_swp): Likewise.
14516 (arm_cmpsi_negshiftsi_si): Likewise.
14517 (movsicc_insn): Likewise.
14518 (movsfcc_soft_insn): Likewise.
14519 (arith_shiftsi): Likewise.
14520 (arith_shiftsi_compare0
14521 (arith_shiftsi_compare0_scratch
14522 (sub_shiftsi): Likewise.
14523 (sub_shiftsi_compare0
14524 (sub_shiftsi_compare0_scratch
14525 (and_scc): Likewise.
14526 (cond_move): Likewise.
14527 (if_plus_move): Likewise.
14528 (if_move_plus): Likewise.
14529 (if_move_not): Likewise.
14530 (if_not_move): Likewise.
14531 (if_shift_move): Likewise.
14532 (if_move_shift): Likewise.
14533 (if_shift_shift): Likewise.
14534 (if_not_arith): Likewise.
14535 (if_arith_not): Likewise.
14536 (cond_move_not): Likewise.
14537 (thumb1_ashlsi3): Set type attribute.
14538 (thumb1_ashrsi3): Likewise.
14539 (thumb1_lshrsi3): Likewise.
14540 (thumb1_rotrsi3): Likewise.
14541 (shiftsi3_compare0_scratch): Likewise.
14542 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
14543 (neon_mov<mode>): Likewise.
14544 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
14546 (thumb2_movsi_insn): Likewise.
14547 (thumb2_cmpsi_neg_shiftsi): Likewise.
14548 (thumb2_extendqisi_v6): Likewise.
14549 (thumb2_zero_extendhisi2_v6): Likewise.
14550 (thumb2_zero_extendqisi2_v6): Likewise.
14551 (thumb2_shiftsi3_short): Likewise.
14552 (thumb2_addsi3_compare0_scratch): Likewise.
14553 (orsi_not_shiftsi_si): Likewise.
14554 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
14555 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
14557 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
14558 (1020alu_shift_op): Likewise.
14559 (1020alu_shift_reg_op): Likewise.
14560 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
14561 (alu_shift_op): Likewise.
14562 (alu_shift_reg_op): Likewise.
14563 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
14564 (11_alu_shift_op): Likewise.
14565 (11_alu_shift_reg_op): Likewise.
14566 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
14567 (9_alu_shift_reg_op): Likewise.
14568 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14570 (cortex_a15_alu_shift): Likewise.
14571 (cortex_a15_alu_shift_reg): Likewise.
14572 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
14574 (cortex_a5_alu_shift): Likewise.
14575 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
14577 (cortex_a53_alu_shift): Likewise.
14578 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14580 (cortex_a7_alu_reg): Likewise.
14581 (cortex_a7_alu_shift): Likewise.
14582 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
14584 (cortex_a8_alu_shift): Likewise.
14585 (cortex_a8_alu_shift_reg): Likewise.
14586 (cortex_a8_mov): Likewise.
14587 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
14588 (cortex_a9_dp_shift): Likewise.
14589 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
14591 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
14593 (cortex_r4_mov): Likewise.
14594 (cortex_r4_alu_shift): Likewise.
14595 (cortex_r4_alu_shift_reg): Likewise.
14596 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
14597 (526_alu_shift_op): Likewise.
14598 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
14599 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
14600 (626te_alu_shift_op): Likewise.
14601 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
14602 (726te_alu_op): Likewise.
14603 (726te_alu_shift_op): Likewise.
14604 (726te_alu_shift_reg_op): Likewise.
14605 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
14606 (mp626_alu_shift_op): Likewise.
14607 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
14608 (pj4_alu_e1_conds): Likewise.
14609 (pj4_alu): Likewise.
14610 (pj4_alu_conds): Likewise.
14611 (pj4_shift): Likewise.
14612 (pj4_shift_conds): Likewise.
14613 (pj4_alu_shift): Likewise.
14614 (pj4_alu_shift_conds): Likewise.
14615 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14617 (cortexa7_older_only): Likewise.
14618 (cortexa7_younger): Likewise.
14620 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14622 * ipa-pure-const.c (generate_summary): Rename to...
14623 (pure_const_generate_summary): ... this.
14625 2013-07-17 Iain Sandoe <iain@codesourcery.com>
14627 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
14629 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
14632 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
14635 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14637 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
14638 enabled without -march=zEC12.
14639 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
14642 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
14644 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
14645 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
14646 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
14647 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
14648 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
14649 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
14650 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
14652 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
14653 and ISA_HAS_NMADD3_NMSUB3.
14654 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
14655 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
14656 (nmsub4<mode>, nmsub3<mode>): Likewise.
14657 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
14659 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
14661 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
14662 TARGET_MIPS5400 checking.
14664 2013-07-16 Jakub Jelinek <jakub@redhat.com>
14665 Peter Bergner <bergner@vnet.ibm.com>
14667 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
14668 registers in the comment.
14669 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
14670 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
14671 rather than FIRST_PSEUDO_REGISTERS.
14673 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
14675 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
14676 enable extra ISA flags with TARGET_HTM.
14678 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
14680 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
14683 2013-07-15 Cong Hou <congh@google.com>
14685 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
14686 in compare function for sorting.
14688 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
14690 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
14691 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
14692 * config/rs6000/rs6000.opt: Add -mhtm option.
14693 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
14694 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
14695 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
14696 __HTM__ if the HTM instructions are available.
14697 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
14698 htm_spr_reg_operand): New define_predicates.
14699 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
14700 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
14702 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
14703 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
14704 HTM builtin functions.
14705 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
14706 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
14707 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
14708 (rs6000_builtin_mask_calculate): Likewise.
14709 (rs6000_option_override_internal): Likewise.
14710 (bdesc_htm): Add new HTM builtin support.
14711 (htm_spr_num): New function.
14712 (htm_spr_regno): Likewise.
14713 (rs6000_htm_spr_icode): Likewise.
14714 (htm_expand_builtin): Likewise.
14715 (htm_init_builtins): Likewise.
14716 (rs6000_expand_builtin): Add support for HTM builtin functions.
14717 (rs6000_init_builtins): Likewise.
14718 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
14720 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
14721 (TARGET_HTM, MASK_HTM): Define macros.
14722 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
14723 (FIXED_REGISTERS): Likewise.
14724 (CALL_USED_REGISTERS): Likewise.
14725 (CALL_REALLY_USED_REGISTERS): Likewise.
14726 (REG_ALLOC_ORDER): Likewise.
14727 (enum reg_class): Likewise.
14728 (REG_CLASS_NAMES): Likewise.
14729 (REG_CLASS_CONTENTS): Likewise.
14730 (REGISTER_NAMES): Likewise.
14731 (ADDITIONAL_REGISTER_NAMES): Likewise.
14732 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
14733 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
14734 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
14735 * config/rs6000/htm.md: New file.
14736 * config/rs6000/htmintrin.h: New file.
14737 * config/rs6000/htmxlintrin.h: New file.
14739 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
14741 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
14742 Define SYMBOL_TINY_GOT, update comment.
14743 * config/aarch64/aarch64.c
14744 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
14745 (aarch64_expand_mov_immediate): Likewise.
14746 (aarch64_print_operand): Likewise.
14747 (aarch64_classify_symbol): Likewise.
14748 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
14749 (ldr_got_tiny): Define.
14751 2013-07-13 Tobias Grosser <tobias@grosser.es>
14753 PR tree-optimization/54094
14754 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
14755 scheduling dimension for the parallelism check from the polyhedral
14756 information in the AST.
14757 * graphite-dependences.c (carries_deps): Do not assume the schedule is
14760 2013-07-13 Jason Merrill <jason@redhat.com>
14762 * print-tree.c (debug_vec_tree): Use debug_raw.
14763 (debug_raw (vec<tree, va_gc> &)): New.
14764 (debug_raw (vec<tree, va_gc> *)): New.
14765 * tree.h: Declare them.
14767 2013-07-13 Bin Cheng <bin.cheng@arm.com>
14769 * ifcvt.c (ifcvt_after_combine): New static variable.
14770 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
14772 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
14773 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
14774 rest_of_handle_if_after_reload): Pass new argument for if_convert.
14776 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
14778 * config/mips/mips.c (mips_expand_call): Remove empty statement.
14780 2013-07-12 Michael Matz <matz@suse.de>
14782 PR middle-end/55771
14783 * convert.c (convert_to_real): Reject non-float inner types.
14785 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
14787 * config/aarch64/aarch64-protos.h
14788 (aarch64_simd_immediate_valid_for_move): Remove.
14789 * config/aarch64/aarch64.c (simd_immediate_info): New member.
14790 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
14792 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
14794 2013-07-11 Steve Ellcey <sellcey@mips.com>
14796 * config/mips/mips.c (mips_conditional_register_usage): Do not
14797 use t[0-7] registers in MIPS16 mode when optimizing for size.
14799 2013-07-11 Sriraman Tallam <tmsriram@google.com>
14801 * config/i386/i386.c (dispatch_function_versions): Fix array
14802 indexing of function_version_info to match actual_versions.
14804 2013-07-11 Teresa Johnson <tejohnson@google.com>
14806 * vec.h (struct va_gc): Move release out-of-line.
14807 (va_gc::release): Call ggc_free on released vec.
14809 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
14811 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
14812 Require GOT register as additional operand in UNSPEC.
14813 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
14814 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
14815 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
14816 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
14817 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
14818 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
14819 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
14821 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
14824 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
14825 name seen by assembler/linker rather if available.
14827 2013-07-11 Andreas Schwab <schwab@suse.de>
14829 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
14831 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
14833 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
14835 2013-07-10 Joseph Myers <joseph@codesourcery.com>
14837 * doc/tm.texi.in: Move hook documentation to ....
14838 * target.def: ... here.
14840 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
14841 text on @hook line.
14842 * doc/tm.texi: Regenerate.
14844 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
14847 * doc/invoke.texi: Document -Wconditionally-supported.
14849 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
14852 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
14855 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
14858 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
14859 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
14860 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
14861 Remove duplicate devices.
14862 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
14863 * config/avr/t-multilib: Regenerate.
14864 * config/avr/avr-tables.opt: Regenerate.
14865 * doc/avr-mmcu.texi: Regenerate.
14867 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
14870 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
14872 2013-07-10 Graham Stott <graham.stott@btinternet.com>
14874 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
14875 the cost of MULT when optimizing for size.
14877 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14879 * config/cr16/cr16-protos.h: Don't include target.h.
14881 2013-07-09 Joseph Myers <joseph@codesourcery.com>
14883 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
14884 adjust register size for TDmode and TFmode for VSX registers.
14886 2013-07-08 Kai Tietz <ktietz@redhat.com>
14889 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
14890 hook_bool_const_tree_true.
14892 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14894 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
14895 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
14896 * config/s390/s390.md: Define FPR*_REGNUM constants.
14897 Fix FPR2_REGNUM constant (18 -> 17).
14898 ("*trunc<BFP:mode><DFP_ALL:mode>2")
14899 ("*trunc<DFP_ALL:mode><BFP:mode>2")
14900 ("trunc<BFP:mode><DFP_ALL:mode>2")
14901 ("trunc<DFP_ALL:mode><BFP:mode>2")
14902 ("*extend<BFP:mode><DFP_ALL:mode>2")
14903 ("*extend<DFP_ALL:mode><BFP:mode>2")
14904 ("extend<BFP:mode><DFP_ALL:mode>2")
14905 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
14908 2013-07-08 Graham Stott <graham.stott@btinternet.com>
14910 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
14912 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14914 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
14915 and cfun_fpr_bit_p to cfun_fpr_save_p.
14916 (s390_frame_area, s390_register_info, s390_frame_info)
14917 (s390_emit_prologue, s390_emit_epilogue)
14918 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
14920 * config/s390/s390.h: Define *_REGNUM macros for floating point
14923 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
14925 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
14927 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
14929 PR rtl-optimization/57786
14930 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
14931 and break out of the loop when it is set to false.
14933 2013-07-08 Jakub Jelinek <jakub@redhat.com>
14936 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
14937 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
14938 (const_int 63)) 0)).
14939 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
14940 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
14941 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
14943 PR rtl-optimization/57829
14944 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
14945 mask bits outside of mode are just sign-extension from mode to HWI.
14947 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
14949 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
14950 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
14951 adjust_address instead of change_address to keep info about alignment.
14952 (emit_strmov): Remove.
14953 (emit_memmov): New function.
14954 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
14955 (expand_movmem_epilogue): Likewise and return updated rtx for
14957 (expand_constant_movmem_prologue): Likewise and return updated rtx for
14958 destination and source.
14959 (decide_alignment): Refactor, handle vector_loop.
14960 (ix86_expand_movmem): Likewise.
14961 (ix86_expand_setmem): Likewise.
14962 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
14964 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
14966 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
14967 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
14969 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
14971 * config/i386/sse.md (sse_movlhps): Change alternative 3
14972 of operand 2 to "m".
14974 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
14977 * config/i386/sse.md (iptr): New mode attribute.
14978 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
14979 (<sse>_vm<plusminus_insn><mode>3): Ditto.
14980 (<sse>_vmmul<mode>3): Ditto.
14981 (<sse>_vmdiv<mode>3): Ditto.
14982 (sse_vmrcpv4sf2): Ditto.
14983 (<sse>_vmsqrt<mode>2): Ditto.
14984 (sse_vmrsqrtv4sf2): Ditto.
14985 (<sse>_vm<code><mode>3): Ditto.
14986 (avx_vmcmp<mode>3): Ditto.
14987 (<sse>_vmmaskcmp<mode>3): Ditto.
14988 (<sse>_comi): Ditto.
14989 (<sse>_ucomi): Ditto.
14990 (*xop_vmfrcz_<mode>): Ditto.
14991 (*fmai_fmadd_<mode>): Ditto.
14992 (*fmai_fmsub_<mode>): Ditto.
14993 (*fmai_fnmadd_<mode>): Ditto.
14994 (*fmai_fnmsub_<mode>): Ditto.
14995 (*fma4i_vmfmadd_<mode>): Ditto.
14996 (*fma4i_vmfmsub_<mode>): Ditto.
14997 (*fma4i_vmfnmadd_<mode>): Ditto.
14998 (*fma4i_vmfnmsub_<mode>): Ditto.
14999 (*xop_vmfrcz_<mode>): Ditto.
15000 (sse_cvtps2pi): Ditto.
15001 (sse_cvttps2pi): Ditto.
15002 (sse_cvtss2si): Ditto.
15003 (sse_cvtss2si_2): Ditto.
15004 (sse_cvtss2siq_2): Ditto.
15005 (sse_cvttss2si): Ditto.
15006 (sse_cvttss2siq): Ditto.
15007 (sse_cvtsd2si): Ditto.
15008 (sse_cvtsd2si_2): Ditto.
15009 (sse_cvtsd2siq_2): Ditto.
15010 (sse_cvttsd2si): Ditto.
15011 (sse_cvttsd2siq): Ditto.
15012 (sse_cvtsd2ss): Ditto.
15013 (sse_cvtss2sd): Ditto.
15014 (avx2_pbroadcast<mode>): Ditto.
15015 (avx2_pbroadcast<mode>_1): Ditto.
15016 (*avx_vperm_broadcast_v4sf): Ditto.
15018 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15019 (sse_movlhps): Ditto.
15020 (sse_storehps): Ditto.
15021 (sse_loadhps): Ditto.
15022 (sse_storelps): Ditto.
15023 (sse_loadlps): Ditto.
15024 (*vec_concatv4sf): Ditto.
15025 (*vec_interleave_highv2df): Ditto.
15026 (*vec_interleave_lowv2df): Ditto.
15027 (*vec_extractv2df_1_sse): Ditto.
15028 (*vec_extractv2df_0_sse): Ditto.
15029 (sse2_storelpd): Ditto.
15030 (sse2_loadlpd): Ditto.
15031 (sse2_movsd): Ditto.
15032 (*vec_concatv4si): Ditto.
15033 (vec_concatv2di): Ditto.
15035 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15036 for Intel asm dialect.
15037 (mmx_punpcklwd): Ditto.
15038 (mmx_punpckldq): Ditto.
15040 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15041 for intel assembler dialect.
15043 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15046 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15047 for BUILT_IN_C{LZ,LRSB}*.
15048 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15049 * tree-vrp.c (extract_range_basic): Handle
15050 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15051 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15052 fall thru to code calling set_value*.
15053 * builtins.c (expand_builtin): Remove *IMAX cases.
15054 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15055 if width is bigger than 2*HWI.
15057 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15059 PR rtl-optimization/55342
15060 * lra-int.h (lra_subreg_reload_pseudos): New.
15061 * lra.c: Add undoing optional reloads to the block diagram.
15062 (lra_subreg_reload_pseudos): New.
15063 (lra_optional_reload_pseudos): Change comments.
15064 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15065 lra_optional_reload_pseudos after undo transformations.
15066 * lra-assigns.c (pseudo_prefix_title): New.
15067 (lra_setup_reg_renumber): Use it.
15068 (spill_for): Ditto. Check subreg reload pseudos too.
15069 (assign_by_spills): Consider subreg reload pseudos too.
15070 * lra-constraints.c (simplify_operand_subreg): Use
15071 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15072 (curr_insn_transform): Recognize and do optional reloads.
15073 (undo_optional_reloads): New.
15074 (lra_undo_inheritance): Call undo_optional_reloads.
15076 2013-07-05 Thomas Quinot <quinot@adacore.com>
15078 * tree-complex.c (expand_complex_operations_1): Fix typo.
15080 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15082 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15083 (tune_params): New member 'const vec_costs'.
15084 * config/aarch64/aarch64.c (generic_vector_cost): New.
15085 (generic_tunings): New member 'generic_vector_cost'.
15086 (aarch64_builtin_vectorization_cost): New.
15087 (aarch64_add_stmt_cost): New.
15088 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15089 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15091 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15094 * config/i386/predicates.md (vsib_address_operand): Disallow
15095 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15097 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15099 PR middle-end/55030
15100 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15101 expand_builtin_setjmp_receiver.
15102 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15103 with NULL for the label parameter.
15104 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15105 the frame-pointer. Adjust comments.
15106 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15107 only if LABEL is non-NULL.
15109 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15111 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15112 (ARM_ABI_AAPCS64): Ditto.
15114 (ARM_DEFAULT_ABI): Ditto.
15116 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15118 * config/aarch64/aarch64-builtins.c
15119 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15120 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15122 * config/aarch64/aarch64-simd.md
15123 (aarch64_ld1<VALL:mode>): New.
15124 (aarch64_st1<VALL:mode>): Likewise.
15125 * config/aarch64/arm_neon.h
15126 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15128 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15130 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15132 (ix86_option_override_internal):Turn on all -mavx target flags by
15133 default as they are dependent on AVX anyway.
15135 2013-07-02 Cary Coutant <ccoutant@google.com>
15137 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15138 deterministic hash.
15139 (loc_checksum_ordered): Likewise.
15140 (hash_loc_operands): Remove inline keyword.
15142 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15144 PR tree-optimization/57741
15145 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15146 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15147 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15148 Allow REAL_CST step_exprs if flag_associative_math.
15149 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15151 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15153 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15155 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15157 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15159 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15161 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15163 (iorsi3_insn): Likewise.
15164 (arm_xorsi3): Likewise.
15166 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15168 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15170 (attribute "type"): ... here, and prefix with "wmmx_".
15171 (attribute "core_cycles"): Update for attribute changes.
15172 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15173 (tbcstv4hi): Likewise.
15174 (tbcstv2si): Likewise.
15175 (iwmmxt_iordi3): Likewise.
15176 (iwmmxt_xordi3): Likewise.
15177 (iwmmxt_anddi3): Likewise.
15178 (iwmmxt_nanddi3): Likewise.
15179 (iwmmxt_arm_movdi): Likewise.
15180 (iwmmxt_movsi_insn): Likewise.
15181 (mov<mode>_internal): Likewise.
15182 (and<mode>3_iwmmxt): Likewise.
15183 (ior<mode>3_iwmmxt): Likewise.
15184 (xor<mode>3_iwmmxt): Likewise.
15185 (add<mode>3_iwmmxt): Likewise.
15186 (ssaddv8qi3): Likewise.
15187 (ssaddv4hi3): Likewise.
15188 (ssaddv2si3): Likewise.
15189 (usaddv8qi3): Likewise.
15190 (usaddv4hi3): Likewise.
15191 (usaddv2si3): Likewise.
15192 (sub<mode>3_iwmmxt): Likewise.
15193 (sssubv8qi3): Likewise.
15194 (sssubv4hi3): Likewise.
15195 (sssubv2si3): Likewise.
15196 (ussubv8qi3): Likewise.
15197 (ussubv4hi3): Likewise.
15198 (ussubv2si3): Likewise.
15199 (mulv4hi3_iwmmxt): Likewise.
15200 (smulv4hi3_highpart): Likewise.
15201 (umulv4hi3_highpart): Likewise.
15202 (iwmmxt_wmacs): Likewise.
15203 (iwmmxt_wmacsz): Likewise.
15204 (iwmmxt_wmacu): Likewise.
15205 (iwmmxt_wmacuz): Likewise.
15206 (iwmmxt_clrdi): Likewise.
15207 (iwmmxt_clrv8qi): Likewise.
15208 (iwmmxt_clr4hi): Likewise.
15209 (iwmmxt_clr2si): Likewise.
15210 (iwmmxt_uavgrndv8qi3): Likewise.
15211 (iwmmxt_uavgrndv4hi3): Likewise.
15212 (iwmmxt_uavgv8qi3): Likewise.
15213 (iwmmxt_uavgv4hi3): Likewise.
15214 (iwmmxt_tinsrb): Likewise.
15215 (iwmmxt_tinsrh): Likewise.
15216 (iwmmxt_tinsrw): Likewise.
15217 (iwmmxt_textrmub): Likewise.
15218 (iwmmxt_textrmsb): Likewise.
15219 (iwmmxt_textrmuh): Likewise.
15220 (iwmmxt_textrmsh): Likewise.
15221 (iwmmxt_textrmw): Likewise.
15222 (iwmxxt_wshufh): Likewise.
15223 (eqv8qi3): Likewise.
15224 (eqv4hi3): Likewise.
15225 (eqv2si3): Likewise.
15226 (gtuv8qi3): Likewise.
15227 (gtuv4hi3): Likewise.
15228 (gtuv2si3): Likewise.
15229 (gtv8qi3): Likewise.
15230 (gtv4hi3): Likewise.
15231 (gtv2si3): Likewise.
15232 (smax<mode>3_iwmmxt): Likewise.
15233 (umax<mode>3_iwmmxt): Likewise.
15234 (smin<mode>3_iwmmxt): Likewise.
15235 (umin<mode>3_iwmmxt): Likewise.
15236 (iwmmxt_wpackhss): Likewise.
15237 (iwmmxt_wpackwss): Likewise.
15238 (iwmmxt_wpackdss): Likewise.
15239 (iwmmxt_wpackhus): Likewise.
15240 (iwmmxt_wpackwus): Likewise.
15241 (iwmmxt_wpackdus): Likewise.
15242 (iwmmxt_wunpckihb): Likewise.
15243 (iwmmxt_wunpckihh): Likewise.
15244 (iwmmxt_wunpckihw): Likewise.
15245 (iwmmxt_wunpckilb): Likewise.
15246 (iwmmxt_wunpckilh): Likewise.
15247 (iwmmxt_wunpckilw): Likewise.
15248 (iwmmxt_wunpckehub): Likewise.
15249 (iwmmxt_wunpckehuh): Likewise.
15250 (iwmmxt_wunpckehuw): Likewise.
15251 (iwmmxt_wunpckehsb): Likewise.
15252 (iwmmxt_wunpckehsh): Likewise.
15253 (iwmmxt_wunpckehsw): Likewise.
15254 (iwmmxt_wunpckelub): Likewise.
15255 (iwmmxt_wunpckeluh): Likewise.
15256 (iwmmxt_wunpckeluw): Likewise.
15257 (iwmmxt_wunpckelsb): Likewise.
15258 (iwmmxt_wunpckelsh): Likewise.
15259 (iwmmxt_wunpckelsw): Likewise.
15260 (ror<mode>3): Likewise.
15261 (ashr<mode>3_iwmmxt): Likewise.
15262 (lshr<mode>3_iwmmxt): Likewise.
15263 (ashl<mode>3_iwmmxt): Likewise.
15264 (ror<mode>3_di): Likewise.
15265 (ashr<mode>3_di): Likewise.
15266 (lshr<mode>3_di): Likewise.
15267 (ashl<mode>3_di): Likewise.
15268 (iwmmxt_wmadds): Likewise.
15269 (iwmmxt_wmaddu): Likewise.
15270 (iwmmxt_tmia): Likewise.
15271 (iwmmxt_tmiaph): Likewise.
15272 (iwmmxt_tmiabb): Likewise.
15273 (iwmmxt_tmiatb): Likewise.
15274 (iwmmxt_tmiabt): Likewise.
15275 (iwmmxt_tmiatt): Likewise.
15276 (iwmmxt_tmovmskb): Likewise.
15277 (iwmmxt_tmovmskh): Likewise.
15278 (iwmmxt_tmovmskw): Likewise.
15279 (iwmmxt_waccb): Likewise.
15280 (iwmmxt_wacch): Likewise.
15281 (iwmmxt_waccw): Likewise.
15282 (iwmmxt_waligni): Likewise.
15283 (iwmmxt_walignr): Likewise.
15284 (iwmmxt_walignr0): Likewise.
15285 (iwmmxt_walignr1): Likewise.
15286 (iwmmxt_walignr2): Likewise.
15287 (iwmmxt_walignr3): Likewise.
15288 (iwmmxt_wsadb): Likewise.
15289 (iwmmxt_wsadh): Likewise.
15290 (iwmmxt_wsadbz): Likewise.
15291 (iwmmxt_wsadhz): Likewise.
15292 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
15293 (iwmmxt_wabsdiffb): Likewise.
15294 (iwmmxt_wabsdiffh): Likewise.
15295 (iwmmxt_wabsdiffw): Likewise.
15296 (iwmmxt_waddsubhx): Likewise
15297 (iwmmxt_wsubaddhx): Likewise.
15298 (addc<mode>3): Likewise.
15299 (iwmmxt_avg4): Likewise.
15300 (iwmmxt_avg4r): Likewise.
15301 (iwmmxt_wmaddsx): Likewise.
15302 (iwmmxt_wmaddux): Likewise.
15303 (iwmmxt_wmaddsn): Likewise.
15304 (iwmmxt_wmaddun): Likewise.
15305 (iwmmxt_wmulwsm): Likewise.
15306 (iwmmxt_wmulwum): Likewise.
15307 (iwmmxt_wmulsmr): Likewise.
15308 (iwmmxt_wmulumr): Likewise.
15309 (iwmmxt_wmulwsmr): Likewise.
15310 (iwmmxt_wmulwumr): Likewise.
15311 (iwmmxt_wmulwl): Likewise.
15312 (iwmmxt_wqmulm): Likewise.
15313 (iwmmxt_wqmulwm): Likewise.
15314 (iwmmxt_wqmulmr): Likewise.
15315 (iwmmxt_wqmulwmr): Likewise.
15316 (iwmmxt_waddbhusm): Likewise.
15317 (iwmmxt_waddbhusl): Likewise.
15318 (iwmmxt_wqmiabb): Likewise.
15319 (iwmmxt_wqmiabt): Likewise.
15320 (iwmmxt_wqmiatb): Likewise.
15321 (iwmmxt_wqmiatt): Likewise.
15322 (iwmmxt_wqmiabbn): Likewise.
15323 (iwmmxt_wqmiabtn): Likewise.
15324 (iwmmxt_wqmiatbn): Likewise.
15325 (iwmmxt_wqmiattn): Likewise.
15326 (iwmmxt_wmiabb): Likewise.
15327 (iwmmxt_wmiabt): Likewise.
15328 (iwmmxt_wmiatb): Likewise.
15329 (iwmmxt_wmiatt): Likewise.
15330 (iwmmxt_wmiabbn): Likewise.
15331 (iwmmxt_wmiabtn): Likewise.
15332 (iwmmxt_wmiatbn): Likewise.
15333 (iwmmxt_wmiattn): Likewise.
15334 (iwmmxt_wmiawbb): Likewise.
15335 (iwmmxt_wmiawbt): Likewise.
15336 (iwmmxt_wmiawtb): Likewise.
15337 (iwmmxt_wmiawtt): Likewise.
15338 (iwmmxt_wmiawbbn): Likewise.
15339 (iwmmxt_wmiawbtn): Likewise.
15340 (iwmmxt_wmiawtbn): Likewise.
15341 (iwmmxt_wmiawttn): Likewise.
15342 (iwmmxt_wmerge): Likewise.
15343 (iwmmxt_tandc<mode>3): Likewise.
15344 (iwmmxt_torc<mode>3): Likewise.
15345 (iwmmxt_torvsc<mode>3): Likewise.
15346 (iwmmxt_textrc<mode>3): Likewise.
15347 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
15348 (wmmxt_pack): Likewise.
15349 (wmmxt_mult_c1): Likewise.
15350 (wmmxt_mult_c2): Likewise.
15351 (wmmxt_alu_c1): Likewise.
15352 (wmmxt_alu_c2): Likewise.
15353 (wmmxt_alu_c3): Likewise.
15354 (wmmxt_transfer_c1): Likewise.
15355 (wmmxt_transfer_c2): Likewise.
15356 (wmmxt_transfer_c3): Likewise.
15357 (marvell_f_iwmmxt_wstr): Likewise.
15358 (marvell_f_iwmmxt_wldr): Likewise.
15360 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
15362 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
15364 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15367 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15368 * lra-constraints.c (need_for_split_p): Check call used hard regs
15369 living through calls.
15371 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
15372 call used regs for call insn.
15374 2013-06-28 Jakub Jelinek <jakub@redhat.com>
15377 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
15378 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
15380 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15382 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
15384 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15386 * lra-constraints.c (need_for_split_p): Check call used hard regs
15387 living through calls.
15389 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
15392 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
15393 to tie with any other modes. Eliminate Altivec vector mode tests,
15394 since these are a subset of ALTIVEC or VSX vector modes. Simplify
15395 code, to return 0 if testing MODE2 for a condition, if we've
15396 already tested MODE1 for the same condition.
15398 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15400 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
15403 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15405 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15406 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
15408 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15410 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
15412 (aarch64_symbolic_constant_p): Remove.
15413 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
15414 static. Fix line length and white space.
15415 (aarch64_symbolic_constant_p): Remove.
15416 * config/aarch64/predicates.md (aarch64_valid_symref):
15417 Use aarch64_classify_symbol_expression.
15419 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15421 * config/arm/constraints.md (Ts): New constraint.
15422 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
15424 (compare_scc): Use "Ts" constraint for operand 0.
15425 (ior_scc_scc): Likewise.
15426 (and_scc_scc): Likewise.
15427 (and_scc_scc_nodom): Likewise.
15428 (ior_scc_scc_cmp): Likewise for operand 7.
15429 (and_scc_scc_cmp): Likewise.
15430 * config/arm/thumb2.md (thumb2_movsi_insn):
15431 Add alternatives for 16-bit encodings.
15432 (thumb2_movhi_insn): Likewise.
15433 (thumb2_movsicc_insn): Likewise.
15434 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
15435 (thumb2_negscc): Use "Ts" constraint.
15436 Move mvn instruction outside cond_exec block.
15437 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
15438 for 16-bit encodings.
15440 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15442 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
15444 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
15445 (mulsi3subsi): Likewise.
15446 (mulsidi3adddi): Likewise.
15447 (mulsidi3_v6): Likewise.
15448 (umulsidi3_v6): Likewise.
15449 (umulsidi3adddi_v6): Likewise.
15450 (smulsi3_highpart_v6): Likewise.
15451 (umulsi3_highpart_v6): Likewise.
15452 (mulhisi3tb): Likewise.
15453 (mulhisi3bt): Likewise.
15454 (mulhisi3tt): Likewise.
15455 (maddhisi4): Likewise.
15456 (maddhisi4tb): Likewise.
15457 (maddhisi4tt): Likewise.
15458 (maddhidi4): Likewise.
15459 (maddhidi4tb): Likewise.
15460 (maddhidi4tt): Likewise.
15461 (zeroextractsi_compare0_scratch): Likewise.
15462 (insv_zero): Likewise.
15463 (insv_t2): Likewise.
15464 (anddi_notzesidi_di): Likewise.
15465 (anddi_notsesidi_di): Likewise.
15466 (andsi_notsi_si): Likewise.
15467 (iordi_zesidi_di): Likewise.
15468 (xordi_zesidi_di): Likewise.
15469 (andsi_iorsi3_notsi): Likewise.
15470 (smax_0): Likewise.
15471 (smax_m1): Likewise.
15472 (smin_0): Likewise.
15473 (not_shiftsi): Likewise.
15474 (unaligned_loadsi): Likewise.
15475 (unaligned_loadhis): Likewise.
15476 (unaligned_loadhiu): Likewise.
15477 (unaligned_storesi): Likewise.
15478 (unaligned_storehi): Likewise.
15479 (extv_reg): Likewise.
15480 (extzv_t2): Likewise.
15481 (divsi3): Likewise.
15482 (udivsi3): Likewise.
15483 (arm_zero_extendhisi2addsi): Likewise.
15484 (arm_zero_extendqisi2addsi): Likewise.
15485 (compareqi_eq0): Likewise.
15486 (arm_extendhisi2_v6): Likewise.
15487 (arm_extendqisi2addsi): Likewise.
15488 (arm_movt): Likewise.
15489 (thumb2_ldrd): Likewise.
15490 (thumb2_ldrd_base): Likewise.
15491 (thumb2_ldrd_base_neg): Likewise.
15492 (thumb2_strd): Likewise.
15493 (thumb2_strd_base): Likewise.
15494 (thumb2_strd_base_neg): Likewise.
15495 (arm_negsi2): Add alternative for 16-bit encoding.
15496 (arm_one_cmplsi2): Likewise.
15498 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15500 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
15501 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
15502 (movdfcc): Likewise.
15503 * config/arm/vfp.md (*thumb2_movsf_vfp):
15504 Disable predication for arm_restrict_it.
15505 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
15506 (*thumb2_movdfcc_vfp): Likewise.
15507 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
15508 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
15509 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
15510 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
15511 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
15512 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
15513 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
15514 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
15515 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
15516 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
15517 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
15518 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
15519 Disable predication for arm_restrict_it.
15521 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
15523 * config/i386/bmiintrin.h (_bextr_u32): New.
15524 (_bextr_u64): Ditto.
15526 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
15528 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
15529 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
15530 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
15531 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
15532 * config/mips/n32-elf.h: ...this new file.
15534 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
15537 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
15538 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
15540 2013-06-27 Catherine Moore <clm@codesourcery.com>
15542 * config/mips/mips-tables.opt: Regenerate.
15543 * config/mips/mips-cpus.def: Add m14ke and m14kec.
15544 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
15545 * doc/invoke.texi: Add -m14kc.
15547 2013-06-27 Jakub Jelinek <jakub@redhat.com>
15550 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
15551 constraints of operand 1 and 2.
15554 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
15555 to match RTL canonicalization. Swap predicates and
15556 constraints of operand 1 and 2.
15558 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
15560 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
15561 Process OP_INOUT regs for splitting too.
15563 2013-06-27 Jakub Jelinek <jakub@redhat.com>
15565 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
15566 decl before the loop, initialize to NULL.
15567 (vectorizable_load): Initialize ptr_incr to NULL.
15569 2013-06-27 Martin Jambor <mjambor@suse.cz>
15572 * ipa-ref.h (ipa_maybe_record_reference): Declare.
15573 * ipa-ref.c (ipa_maybe_record_reference): New function.
15574 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
15575 * ipa-cp.c (create_specialized_node): Record potential references from
15577 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
15579 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
15581 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
15582 parameter 'mode' of type 'enum machine_mode mode'; change to pass
15583 'mode' to force_reg.
15584 (aarch64_add_offset): Update calls to aarch64_force_temporary.
15585 (aarch64_expand_mov_immediate): Likewise.
15587 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
15589 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
15590 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
15592 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15594 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
15595 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
15596 (struct machine_function): Add tbegin_p.
15597 (s390_canonicalize_comparison): Fold CC mode compares to
15598 conditional jump if possible.
15599 (s390_emit_jump): Return the emitted jump.
15600 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
15601 Handle CCRAWmode compares.
15602 (s390_option_override): Default to -mhtm if available.
15603 (s390_reg_clobbered_rtx): Handle floating point regs as well.
15604 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
15605 FPRs instead of df_regs_ever_live_p.
15606 (s390_optimize_nonescaping_tx): New function.
15607 (s390_init_frame_layout): Extend clobbered_regs array to cover
15609 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
15610 (s390_expand_tbegin): New function.
15611 (enum s390_builtin): New enum definition.
15612 (code_for_builtin): New array definition.
15613 (s390_init_builtins): New function.
15614 (s390_expand_builtin): New function.
15615 (TARGET_INIT_BUILTINS): Define.
15616 (TARGET_EXPAND_BUILTIN): Define.
15617 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
15618 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
15619 (s390_alc_comparison): Likewise.
15620 * config/s390/s390-modes.def: Add CCRAWmode.
15621 * config/s390/s390.h (processor_flags): Add PF_TX.
15622 (TARGET_CPU_HTM): Define macro.
15623 (TARGET_HTM): Define macro.
15624 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
15625 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
15626 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
15627 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
15628 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
15629 (TBEGIN_MASK, TBEGINC_MASK): New constants.
15630 ("*cc_to_int"): Move up.
15631 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
15632 constants other than 0.
15633 ("*ccraw_to_int"): New insn and splitter definition.
15634 ("tbegin", "tbegin_nofloat", "tbegin_retry")
15635 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
15636 ("tx_assist"): New expander.
15637 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
15638 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
15639 * config/s390/s390.opt: Add -mhtm option.
15640 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
15641 * config/s390/htmxlintrin.h: New file.
15642 * config/s390/htmintrin.h: New file.
15643 * config/s390/s390intrin.h: New file.
15644 * doc/extend.texi: Document htm builtins.
15645 * config.gcc: Add the new header files to extra_headers.
15647 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
15649 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
15650 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
15652 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
15653 Pat Haugen <pthaugen@us.ibm.com>
15654 Peter Bergner <bergner@vnet.ibm.com>
15656 * config/rs6000/power8.md: New.
15657 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
15658 setting for power8 entry.
15659 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
15660 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
15661 test for Power4/Power5 only.
15662 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
15664 (force_new_group): Adjust comment.
15665 * config/rs6000/rs6000.md: Include power8.md.
15667 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
15669 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
15670 * config/arm/arm-protos.h (arm_max_conditional_execute): New
15672 (tune_params): Update comment.
15673 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
15674 (arm_max_conditional_execute): New function.
15675 (thumb2_final_prescan_insn): Use max_insn_skipped and
15676 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
15678 2013-06-25 Jakub Jelinek <jakub@redhat.com>
15680 PR tree-optimization/57705
15681 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
15682 SSA_NAME step, provided that it is not defined inside the loop.
15683 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
15684 (get_initial_def_for_induction): Handle SSA_NAME IV step.
15686 2013-06-25 Martin Jambor <mjambor@suse.cz>
15688 PR middle-end/57670
15689 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
15690 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
15692 (ipa_note_param_call): Initialize member_ptr flag.
15693 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
15694 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
15695 (ipa_write_indirect_edge_info): Stream member_ptr flag.
15696 (ipa_read_indirect_edge_info): Likewise.
15698 2013-06-25 Richard Biener <rguenther@suse.de>
15700 PR middle-end/56977
15701 * passes.c (init_optimization_passes): Move pass_fold_builtins
15702 and pass_dce earlier with -Og.
15704 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
15706 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
15707 <BIT_FIELD_REF>: Remove trailing TAB.
15708 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
15711 2013-06-24 Martin Jambor <mjambor@suse.cz>
15713 PR tree-optimization/57358
15714 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
15715 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
15716 (ipa_analyze_params_uses): Generate pessimistic info when true.
15718 2013-06-24 Martin Jambor <mjambor@suse.cz>
15720 PR tree-optimization/57539
15721 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
15722 global.inlined_to of the new node to it. All callers changed.
15723 * ipa-inline-transform.c (clone_inlined_nodes): New variable
15724 inlining_into, pass it to cgraph_clone_node.
15725 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
15726 ipa_free_edge_args_substructures.
15727 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
15728 rdesc linked list. Do not assert rdesc edges have inlined caller.
15729 Assert we have found an rdesc in the rdesc list.
15731 2013-06-24 Richard Biener <rguenther@suse.de>
15733 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
15734 (pointer_set_lookup): Declare.
15735 (class pointer_map): New template class implementing a
15736 generic pointer to T map.
15737 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
15738 pointer_map<T>::contains, pointer_map<T>::insert,
15739 pointer_map<T>::traverse): New functions.
15740 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
15741 (pointer_set_lookup): New function.
15742 (pointer_set_contains): Use pointer_set_lookup.
15743 (pointer_set_insert): Likewise.
15744 (insert_aux): Remove.
15745 (struct pointer_map_t): Embed a pointer_set_t.
15746 (pointer_map_create): Adjust.
15747 (pointer_map_destroy): Likewise.
15748 (pointer_map_contains): Likewise.
15749 (pointer_map_insert): Likewise.
15750 (pointer_map_traverse): Likewise.
15751 * tree-streamer.h (struct streamer_tree_cache_d): Use a
15752 pointer_map<unsigned> instead of a pointer_map_t.
15753 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
15754 (streamer_tree_cache_lookup): Likewise.
15755 (streamer_tree_cache_create): Likewise.
15756 (streamer_tree_cache_delete): Likewise.
15757 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
15758 pointer_map<unsigned> instead of a pointer_map_t.
15759 (lto_init_tree_ref_encoder): Adjust.
15760 (lto_destroy_tree_ref_encoder): Likewise.
15761 * lto-section-out.c (lto_output_decl_index): Likewise.
15762 (lto_record_function_out_decl_state): Likewise.
15763 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
15765 2013-06-24 Richard Biener <rguenther@suse.de>
15767 PR tree-optimization/57488
15768 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
15770 2013-06-24 Alan Modra <amodra@gmail.com>
15772 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
15773 (gen_easy_altivec_constant): Likewise.
15774 * config/rs6000/predicates.md (easy_vector_constant_add_self,
15775 easy_vector_constant_msb): Likewise.
15777 2013-06-23 Jakub Jelinek <jakub@redhat.com>
15780 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
15781 add missing return true.
15783 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
15786 * config/sh/predicates.md (general_extend_operand): Invoke
15787 general_movsrc_operand for memory operands.
15788 (general_movsrc_operand): Allow reg+reg addressing, do not use
15789 general_operand for memory operands.
15791 2013-06-23 Sriraman Tallam <tmsriram@google.com>
15793 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
15794 when current target options does not apply.
15795 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
15796 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
15797 * config/i386/bmiintrin.h: Pass appropriate target
15798 attributes to header.
15799 * config/i386/mmintrin.h: Ditto.
15800 * config/i386/nmmintrin.h: Ditto.
15801 * config/i386/avx2intrin.h: Ditto.
15802 * config/i386/fxsrintrin.h: Ditto.
15803 * config/i386/tbmintrin.h: Ditto.
15804 * config/i386/xsaveintrin.h: Ditto.
15805 * config/i386/f16cintrin.h: Ditto.
15806 * config/i386/xtestintrin.h: Ditto.
15807 * config/i386/xsaveoptintrin.h: Ditto.
15808 * config/i386/bmi2intrin.h: Ditto.
15809 * config/i386/lzcntintrin.h: Ditto.
15810 * config/i386/smmintrin.h: Ditto.
15811 * config/i386/wmmintrin.h: Ditto.
15812 * config/i386/x86intrin.h: Remove all header include guards.
15813 * config/i386/prfchwintrin.h: Ditto.
15814 * config/i386/pmmintrin.h: Ditto.
15815 * config/i386/tmmintrin.h: Ditto.
15816 * config/i386/xmmintrin.h: Ditto.
15817 * config/i386/popcntintrin.h: Ditto.
15818 * config/i386/rdseedintrin.h: Ditto.
15819 * config/i386/ammintrin.h: Ditto.
15820 * config/i386/emmintrin.h: Ditto.
15821 * config/i386/immintrin.h: Remove all header include guards.
15822 * config/i386/fma4intrin.h: Ditto.
15823 * config/i386/lwpintrin.h: Ditto.
15824 * config/i386/xopintrin.h: Ditto.
15825 * config/i386/ia32intrin.h: Ditto.
15826 * config/i386/avxintrin.h: Ditto.
15827 * config/i386/rtmintrin.h: Ditto.
15828 * config/i386/fmaintrin.h: Ditto.
15829 * config/i386/mm3dnow.h: Ditto.
15831 2013-06-22 Sriraman Tallam <tmsriram@google.com>
15833 * common/config/i386/i386-common.c: Handle LZCNT.
15835 2013-06-22 Andi Kleen <ak@linux.intel.com>
15837 * doc/extend.texi: Use __atomic_store_n instead of
15838 __atomic_store in HLE example.
15840 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
15842 * config/sh/sh.c: Remove <cstdlib> workaround.
15844 2013-06-21 Andi Kleen <ak@linux.intel.com>
15846 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
15848 2013-06-21 Andi Kleen <ak@linux.intel.com>
15850 * doc/extend.texi: Document that __atomic_clear and
15851 __atomic_test_and_set should only be used with bool.
15853 2013-06-20 Jan Hubicka <jh@suse.cz>
15855 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
15856 types_same_for_odr.
15857 * tree.c (decls_same_for_odr): New function.
15858 (same_for_edr): New function.
15859 (types_same_for_odr): New function.
15860 (get_binfo_at_offset): Use it.
15861 * tree.h (types_same_for_odr): Declare.
15863 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
15864 Jason Merrill <jason@redhat.com>
15866 * system.h: Include <cstdlib> as well as <stdlib.h>.
15868 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
15871 * config/i386/i386.c (construct_container): Report error if
15872 long double is used with disabled x87 float returns.
15874 2013-06-20 Jan Hubicka <jh@suse.cz>
15876 * lto-cgraph.c (input_symtab): Do not set cgraph state.
15878 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
15880 PR rtl-optimization/57425
15881 PR rtl-optimization/57569
15882 * alias.c (write_dependence_p): Remove parameters mem_mode and
15883 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
15884 Changed all callers.
15885 (canon_anti_dependence): Get comments and semantics in sync.
15886 Add parameter mem_canonicalized. Changed all callers.
15887 * rtl.h (canon_anti_dependence): Update prototype.
15889 2013-06-20 Richard Biener <rguenther@suse.de>
15891 * data-streamer-in.c (streamer_read_uhwi): Optimize single
15892 byte case, inline streamer_read_uchar and defer section
15895 2013-06-20 Richard Biener <rguenther@suse.de>
15897 PR tree-optimization/57584
15898 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
15899 SSA names into the expanded expression that take part in
15900 abnormal coalescing.
15902 2013-06-19 Sharad Singhai <singhai@google.com>
15904 * gcov.c (print_usage): Handle new option.
15905 (process_args): Ditto.
15906 (get_gcov_intermediate_filename): New function.
15907 (output_intermediate_file): New function.
15908 (output_gcov_file): New function
15909 (generate_results): Handle new option.
15910 (release_function): Relase demangled name.
15911 (read_graph_file): Handle demangled name.
15912 (output_lines): Ditto.
15913 * doc/gcov.texi: Document gcov intermediate format.
15915 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
15918 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
15919 (lra_emit_add): Use the functions. Add comment about Y as an
15922 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
15925 * collect2.c (collect_atexit): New.
15926 (collect_exit): Delete.
15927 (main): Register collect_atexit with atexit.
15928 (collect_wait): Change collect_exit to exit.
15930 * collect2.h (collect_exit): Delete.
15931 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
15933 2013-06-19 Wei Mi <wmi@google.com>
15935 PR rtl-optimization/57518
15936 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
15937 if regno is used in paradoxical subreg.
15938 (update_equiv_regs): Check pdx_subregs[regno] before
15939 set a reg to be equivalent with a mem.
15941 2013-06-19 Matthias Klose <doko@ubuntu.com>
15944 * file-find.h (find_a_file): Add a mode parameter.
15945 * file-find.c (find_a_file): Likewise.
15946 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
15947 with X_OK for the executables.
15948 * collect2.c (main): Call find_a_file with X_OK.
15950 2013-06-19 Steve Ellcey <sellcey@mips.com>
15953 * config/mips/mips.md (casesi_internal_mips16_<mode>):
15954 Use NEXT_INSN instead of next_real_insn.
15956 2013-06-19 Jan Hubicka <jh@suse.cz>
15958 * cgraph.h (const_value_known_p): Replace by ...
15959 (ctor_for_folding): .. this one.
15960 * cgraphunit.c (process_function_and_variable_attributes): Use it.
15961 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
15962 * expr.c (expand_expr_real_1): Likewise.
15963 (string_constant): Likewise.
15964 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
15965 * ipa.c (process_references): Likewise.
15966 (symtab_remove_unreachable_nodes): Likewise.
15967 * ipa-inline-analysis.c (param_change_prob): Likewise.
15968 * gimple-fold.c (canonicalize_constructor_val): Likewise.
15969 (get_base_constructor): Likwise.
15970 * varpool.c (varpool_remove_node): Likewise.
15971 (varpool_remove_initializer): LIkewise.
15972 (dump_varpool_node): LIkwise.
15973 (const_value_known_p): Rewrite to ...
15974 (ctor_for_folding): ... this one.
15976 2013-06-19 Jakub Jelinek <jakub@redhat.com>
15979 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
15980 PERSONALITY in $PATH derived prefixes.
15982 2013-06-19 Jeff Law <law@redhat.com>
15984 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
15987 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
15988 (simplify_bitwise_binary): Use it to simpify certain binary ops on
15991 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
15993 * config/arm/vfp.md: Move VFP instruction classification documentation
15995 * config/arm/arm.md: ... here. Update instruction classification
15998 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16000 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16001 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16002 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16004 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16007 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16008 with NEXT_INSN. Use tablejump_p to check for jump table data
16011 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16014 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16015 that now in C++ the value is correct per the C++ standards.
16017 2013-06-19 Richard Biener <rguenther@suse.de>
16019 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16020 for global context.
16022 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16025 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16028 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16029 with next_active_insn.
16031 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16033 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16034 functions are inlined during failures to flag an error.
16035 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16036 in early inline pass.
16038 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16040 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16043 2013-06-18 Julian Brown <julian@codesourcery.com>
16045 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16046 Permit virtual register pre-reload if !strict.
16047 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16049 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16051 * config/arm/neon.md (movmisalign<mode>): Use
16052 neon_perm_struct_or_reg_operand instead of
16053 neon_struct_or_register_operand.
16054 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16055 neon_permissive_struct_operand instead of neon_struct_operand.
16056 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16057 neon_vector_mem_operand.
16058 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16059 neon_vector_mem_operand.
16060 (neon_permissive_struct_operand): New.
16061 (neon_struct_or_register_operand): Rename to...
16062 (neon_perm_struct_or_reg_operand): This. Adjust call to
16063 neon_vector_mem_operand.
16065 2013-06-18 Richard Biener <rguenther@suse.de>
16067 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16068 * lto-streamer.h: Include pointer-set.h.
16069 (struct lto_decl_slot): Remove.
16070 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16071 Remove next_index entry.
16072 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16073 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16074 (lto_init_tree_ref_encoder): Adjust.
16075 (lto_destroy_tree_ref_encoder): Likewise.
16076 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16077 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16078 (lto_output_decl_index): Adjust.
16079 (lto_new_out_decl_state): Likewise.
16080 (lto_record_function_out_decl_state): Likewise.
16081 * lto-streamer-out.c (copy_function): Likewise.
16083 2013-06-18 Richard Biener <rguenther@suse.de>
16085 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16086 * cgraphunit.c: Include cfgloop.h.
16087 (init_lowered_empty_function): Initialize the loop tree.
16088 (assemble_thunk): Insert new BBs into loops.
16090 2013-06-18 Richard Biener <rguenther@suse.de>
16092 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16093 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16094 the map from cache entry to cache index optional.
16095 (streamer_tree_cache_replace_tree): Adjust accordingly.
16096 (streamer_tree_cache_append): Likewise.
16097 (streamer_tree_cache_delete): Likewise.
16098 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16099 streamer cache map from cache entry to cache index.
16100 * lto-streamer-out.c (create_output_block): Adjust.
16102 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16104 * config/arm/arm.md (attribute "insn"): Move multiplication and
16105 division attributes to...
16106 (attribute "type"): ... here. Remove mult.
16107 (attribute "mul32"): New attribute.
16108 (attribute "mul64"): Add umaal.
16109 (*arm_mulsi3): Update attributes.
16110 (*arm_mulsi3_v6): Likewise.
16111 (*thumb_mulsi3): Likewise.
16112 (*thumb_mulsi3_v6): Likewise.
16113 (*mulsi3_compare0): Likewise.
16114 (*mulsi3_compare0_v6): Likewise.
16115 (*mulsi_compare0_scratch): Likewise.
16116 (*mulsi_compare0_scratch_v6): Likewise.
16117 (*mulsi3addsi): Likewise.
16118 (*mulsi3addsi_v6): Likewise.
16119 (*mulsi3addsi_compare0): Likewise.
16120 (*mulsi3addsi_compare0_v6): Likewise.
16121 (*mulsi3addsi_compare0_scratch): Likewise.
16122 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16123 (*mulsi3subsi): Likewise.
16124 (*mulsidi3adddi): Likewise.
16125 (*mulsi3addsi_v6): Likewise.
16126 (*mulsidi3adddi_v6): Likewise.
16127 (*mulsidi3_nov6): Likewise.
16128 (*mulsidi3_v6): Likewise.
16129 (*umulsidi3_nov6): Likewise.
16130 (*umulsidi3_v6): Likewise.
16131 (*umulsidi3adddi): Likewise.
16132 (*umulsidi3adddi_v6): Likewise.
16133 (*smulsi3_highpart_nov6): Likewise.
16134 (*smulsi3_highpart_v6): Likewise.
16135 (*umulsi3_highpart_nov6): Likewise.
16136 (*umulsi3_highpart_v6): Likewise.
16137 (mulhisi3): Likewise.
16138 (*mulhisi3tb): Likewise.
16139 (*mulhisi3bt): Likewise.
16140 (*mulhisi3tt): Likewise.
16141 (maddhisi4): Likewise.
16142 (*maddhisi4tb): Likewise.
16143 (*maddhisi4tt): Likewise.
16144 (maddhidi4): Likewise.
16145 (*maddhidi4tb): Likewise.
16146 (*maddhidi4tt): Likewise.
16147 (divsi3): Likewise.
16148 (udivsi3): Likewise.
16149 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16150 (thumb2_mulsi_short_compare0): Likewise.
16151 (thumb2_mulsi_short_compare0_scratch): Likewise.
16152 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16153 (1020mult2): Likewise.
16154 (1020mult3): Likewise.
16155 (1020mult4): Likewise.
16156 (1020mult5): Likewise.
16157 (1020mult6): Likewise.
16158 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16160 (cortex_a15_mult64): Likewise.
16161 (cortex_a15_sdiv): Likewise.
16162 (cortex_a15_udiv): Likewise.
16163 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16169 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16170 (pj4_ir_div): Likewise.
16171 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16172 (11_mult2): Likewise.
16173 (11_mult3): Likewise.
16174 (11_mult4): Likewise.
16175 (11_mult5): Likewise.
16176 (11_mult6): Likewise.
16177 (11_mult7): Likewise.
16178 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16179 (cortex_a8_mla): Likewise.
16180 (cortex_a8_mull): Likewise.
16181 (cortex_a8_smulwy): Likewise.
16182 (cortex_a8_smlald): Likewise.
16183 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16184 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16185 (cortex_r4_mul_3): Likewise.
16186 (cortex_r4_mla_4): Likewise.
16187 (cortex_r4_mla_3): Likewise.
16188 (cortex_r4_smlald): Likewise.
16189 (cortex_r4_mull): Likewise.
16190 (cortex_r4_sdiv): Likewise.
16191 (cortex_r4_udiv): Likewise.
16192 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16193 (cortex_a7_idiv): Likewise.
16194 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16195 (9_mult2): Likewise.
16196 (9_mult3): Likewise.
16197 (9_mult4): Likewise.
16198 (9_mult5): Likewise.
16199 (9_mult6): Likewise.
16200 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16201 (cortex_a53_sdiv): Likewise.
16202 (cortex_a53_udiv): Likewise.
16203 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16204 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16205 (mp626_mult2): Likewise.
16206 (mp626_mult3): Likewise.
16207 (mp626_mult4): Likewise.
16208 * config/arm/fa526.md (526_mult1): Update attribute change.
16209 (526_mult2): Likewise.
16210 * config/arm/arm-generic.md (mult): Update attribute change.
16211 (mult_ldsched_strongarm): Likewise.
16212 (mult_ldsched): Likewise.
16213 (multi_cycle): Likewise.
16214 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16215 * config/arm/fa606te.md (606te_mult1): Update attribute change.
16216 (606te_mult2): Likewise.
16217 (606te_mult3): Likewise.
16218 (606te_mult4): Likewise.
16219 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
16220 (cortex_a9_mac16): Likewise.
16221 (cortex_a9_multiply): Likewise.
16222 (cortex_a9_mac): Likewise.
16223 (cortex_a9_multiply_long): Likewise.
16224 * config/arm/fa626te.md (626te_mult1): Update attribute change.
16225 (626te_mult2): Likewise.
16226 (626te_mult3): Likewise.
16227 (626te_mult4): Likewise.
16229 2013-06-18 Richard Biener <rguenther@suse.de>
16232 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
16234 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16237 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16238 with next_active_insn.
16240 2013-06-18 Alan Modra <amodra@gmail.com>
16242 * config/rs6000/rs6000.h (enum data_align): New.
16243 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
16244 (DATA_ABI_ALIGNMENT): Define.
16245 (CONSTANT_ALIGNMENT): Correct comment.
16246 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
16247 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
16249 2013-06-17 David Malcolm <dmalcolm@redhat.com>
16251 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
16252 ATTRIBUTE_UNUSED marking.
16254 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
16256 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
16257 alternative and update.
16258 (aarch64_dup_lanedi): Delete.
16259 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
16260 * config/aarch64/aarch64-simd-builtins.def: Update.
16262 2013-06-17 Richard Biener <rguenther@suse.de>
16264 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
16265 (lto_input_scc): Declare.
16266 (lto_input_tree_1): Likewise.
16267 (struct lto_stats_d): Add num_tree_bodies_output and
16268 num_pickle_refs_output.
16269 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
16270 (lto_read_tree_1): Split out from ...
16271 (lto_read_tree): ... this.
16272 (lto_input_scc): New function.
16273 (lto_input_tree_1): Split out from ...
16274 (lto_input_tree): ... this. Handle LTO_tree_scc.
16275 (lto_data_in_create): Create the streamer cache without hashes.
16276 * lto-streamer-out.c (create_output_block): Create the streamer
16277 cache with hashes when not doing WPA.
16278 (lto_write_tree_1): Split out from ...
16279 (lto_write_tree): ... this.
16280 (get_symbol_initial_value): New function.
16281 (lto_output_tree_1): Split out from ...
16282 (lto_output_tree): ... this. Write trees as series of SCCs
16283 using a DFS walk via DFS_write_tree.
16284 (struct sccs, struct scc_entry): New types.
16285 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
16286 (DFS_write_tree_body): New function.
16287 (DFS_write_tree): Likewise.
16288 (hash_tree): Likewise.
16289 (scc_entry_compare): Likewise.
16290 (hash_scc): Likewise.
16291 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
16292 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
16293 TREE_CHAIN as regular reference.
16294 (streamer_read_integer_cst): Remove.
16295 (streamer_get_pickled_tree): Adjust.
16296 * tree-streamer-out.c (streamer_write_chain): Disable streaming
16297 of DECL_EXTERNALs in BLOCK_VARS for now.
16298 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
16300 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
16301 Add hash value argument and record that if hashes are recorded
16303 (streamer_tree_cache_insert_1): Adjust.
16304 (streamer_tree_cache_insert): Likewise.
16305 (streamer_tree_cache_insert_at): Rename to ...
16306 (streamer_tree_cache_replace_tree): ... this and adjust.
16307 (streamer_tree_cache_append): Adjust.
16308 (record_common_node): Likewise.
16309 (streamer_tree_cache_create): Add argument whether to
16310 record hash values together with trees.
16311 (streamer_tree_cache_delete): Adjust.
16312 * tree-streamer.h (struct streamer_tree_cache_d): Add
16314 (streamer_read_integer_cst): Remove.
16315 (streamer_tree_cache_insert): Adjust.
16316 (streamer_tree_cache_append): Likewise.
16317 (streamer_tree_cache_insert_at): Rename to ...
16318 (streamer_tree_cache_replace_tree): ... this and adjust.
16319 (streamer_tree_cache_create): Add argument whether to record hashes.
16320 (streamer_tree_cache_get): Rename to ...
16321 (streamer_tree_cache_get_tree): ... this.
16322 (streamer_tree_cache_get_hash): New function.
16323 * tree.c (cache_integer_cst): New function.
16324 * tree.h (cache_integer_cst): Declare.
16325 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
16326 * lto-symtab.c (lto_varpool_replace_node): Only release
16327 DECL_INITIAL of non-prevailing decls.
16328 * varpool.c (varpool_remove_initializer): Do not release
16329 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
16331 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16333 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
16334 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
16335 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
16336 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
16337 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
16338 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
16339 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
16340 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
16341 instead of TARGET_64BIT.
16342 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
16343 Require ISA_HAS_<D>DIV.
16345 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
16347 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
16348 (mips*-*-linux*): Move default with_llsc setting to where other
16350 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
16352 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
16353 Likewise. Remove default with_tune setting. Move default float
16354 setting to its own block. Handle with_llsc in the same block as above.
16356 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
16358 PR rtl-optimization/57425
16359 PR rtl-optimization/57569
16360 * alias.c (write_dependence_p): Add new parameters mem_mode,
16361 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
16362 Changed all callers.
16363 (canon_anti_dependence): New function.
16364 * cse.c (check_dependence): Use canon_anti_dependence.
16365 * cselib.c (cselib_invalidate_mem): Likewise.
16366 * rtl.h (canon_anti_dependence): Declare.
16368 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16370 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
16371 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
16372 ".set mips3" for 64-bit targets.
16374 2013-06-15 Dehao Chen <dehao@google.com>
16376 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
16377 * gimple-low.c (gimple_check_call_matching_types): Likewise.
16378 (gimple_check_call_args): Likewise.
16379 * value-prof.c (check_ic_target): Likewise.
16380 * ipa-inline.c (early_inliner): Likewise.
16381 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
16382 * cgraph.c (cgraph_create_edge_1): Likewise.
16383 (cgraph_make_edge_direct): Likewise.
16385 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
16388 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
16389 rs6000_output_move_128bit to handle emitting quad memory
16390 operations. Set attribute length to 8 bytes.
16392 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
16394 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
16396 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
16397 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
16398 (aarch64_<su>mlsl<mode>): Likewise.
16400 2013-06-14 Mike Stump <mikestump@comcast.net>
16402 * Makefile.in (TARGET_H): Add insn-codes.h.
16404 2013-06-14 Alan Modra <amodra@gmail.com>
16406 PR middle-end/57134
16407 PR middle-end/57586
16408 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
16409 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
16410 bitfield expansion when EXPAND_MEMORY.
16411 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
16413 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
16415 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
16416 test for clearing quad memory on 32-bit later.
16418 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
16420 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
16421 (fold_negate_expr): Likewise.
16422 (fold_real_zero_addition_p): Handle vectors.
16423 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
16425 2013-06-14 Alan Modra <amodra@gmail.com>
16427 * varasm.c (force_const_mem): Revert 2013-06-07 change.
16429 2013-06-13 Jan Hubicka <jh@suse.cz>
16431 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16432 Local comdats are not externally visible.
16433 * symtab.c (dump_symtab_base): Dump externally visible.
16434 (verify_symtab_base): Verify back links in the symtab hash.
16436 2013-06-13 Bin Cheng <bin.cheng@arm.com>
16438 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
16439 CONVERT_EXPR as equal nodes.
16441 2013-06-13 Bin Cheng <bin.cheng@arm.com>
16443 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
16445 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
16447 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
16448 Generalize to complex and vector.
16449 * tree.c (build_all_ones_cst): New function.
16450 * tree.h (build_all_ones_cst): Declare it.
16452 2013-06-13 Alan Modra <amodra@gmail.com>
16454 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
16455 * config/rs6000/rs6000.md (signbittf2): New insn.
16456 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
16457 (abstf2_internal, cmptf_internal2): Likewise.
16458 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
16460 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
16461 Pat Haugen <pthaugen@us.ibm.com>
16462 Peter Bergner <bergner@vnet.ibm.com>
16464 * config/rs6000/rs6000.c (emit_load_locked): Add support for
16465 power8 byte, half-word, and quad-word atomic instructions.
16466 (emit_store_conditional): Likewise.
16467 (rs6000_expand_atomic_compare_and_swap): Likewise.
16468 (rs6000_expand_atomic_op): Likewise.
16470 * config/rs6000/sync.md (larx): Add new modes for power8.
16472 (AINT): New mode iterator to include TImode as well as normal
16473 integer modes on power8.
16474 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
16475 that VSX registers are not considered. Use AINT mode iterator
16476 instead of INT1 to allow inclusion of quad word atomic operations
16478 (load_locked<mode>): Likewise.
16479 (store_conditional<mode>): Likewise.
16480 (atomic_compare_and_swap<mode>): Likewise.
16481 (atomic_exchange<mode>): Likewise.
16482 (atomic_nand<mode>): Likewise.
16483 (atomic_fetch_<fetchop_name><mode>): Likewise.
16484 (atomic_nand_fetch<mode>): Likewise.
16485 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
16487 (ATOMIC): On power8, add QImode, HImode modes.
16488 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
16489 modes that promote to SImode.
16490 (load_lockedti): Convert TImode arguments to PTImode, so that we
16491 get a guaranteed even/odd register pair.
16492 (load_lockedpti): Likewise.
16493 (store_conditionalti): Likewise.
16494 (store_conditionalpti): Likewise.
16496 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
16497 atomic load/store instructions.
16500 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
16502 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
16504 (insn_count): New attribute, with most cases extracted from...
16505 (length): ...here. Redefine most cases in terms of insn_count.
16506 (single_insn): Delete.
16507 (can_delay): Use insn_count to check for single instructions.
16508 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
16509 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
16510 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
16511 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
16512 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
16513 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
16514 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
16515 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
16516 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
16517 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
16518 rather than "length".
16519 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
16520 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
16521 Use "insn_count" rather than "length".
16522 * config/mips/mips-dsp.md
16523 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
16524 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
16527 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
16529 PR tree-optimization/57361
16530 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
16532 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
16534 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
16536 (aarch64_simd_combine<mode>): New instruction expansion.
16537 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
16538 function prototype.
16539 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
16540 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
16542 2013-06-12 Jan Hubicka <jh@suse.cz>
16544 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
16545 decl has when in streaming stage.
16546 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
16547 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
16549 2013-06-12 Roland Stigge <stigge@antcom.de>
16552 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
16554 2013-06-12 Jakub Jelinek <jakub@redhat.com>
16556 PR tree-optimization/57537
16557 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
16558 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
16560 2013-06-12 Richard Biener <rguenther@suse.de>
16562 * data-streamer.h (streamer_write_char_stream): CSE
16563 obs->current_pointer.
16564 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
16565 streamer_write_char_stream manually and optimize the resulting loop.
16566 (streamer_write_hwi_stream): Likewise.
16568 2013-06-12 Jan Hubicka <jh@suse.cz>
16570 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
16571 * cgraph.h (varpool_create_empty_node): Declare.
16572 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
16574 * symtab.c (symtab_unregister_node): Be lax about missin entries
16576 (symtab_get_node): Update comment.
16577 * varpool.c (varpool_create_empty_node): Break out from ...
16578 (varpool_node_for_decl): ... here.
16579 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
16581 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
16583 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
16584 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
16585 part. Use straight-line flow at the end.
16586 <COMPONENT_REF>: Remove superfluous else.
16587 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
16589 2013-06-12 Jakub Jelinek <jakub@redhat.com>
16592 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
16593 target hook even for !TREE_PUBLIC decls. If no resolution info
16594 is available, return false for common and external decls.
16596 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16598 * config/rl78/constraints.md (U): New constraint.
16599 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
16602 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
16605 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
16606 to allow returning address to AT_PLATFORM name.
16608 2013-06-11 Jan Hubicka <jh@suse.cz>
16610 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
16611 * cgraph.h (symtab_node_base): Add weakref flag.
16612 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
16613 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
16614 (output_weakrefs): Use weakref flag.
16615 * fold-const.c (simple_operand_p): Handle WEAK.
16616 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
16617 * ipa.c (varpool_externally_visible_p): Drop weakref.
16618 (function_and_variable_visibility): Update comment; fix weakref
16619 sanity checks; do not clear DECL_WEAK on them.
16620 * lto-cgraph.c (lto_output_node): update.
16621 (lto_output_varpool_node): Update.
16622 (input_overwrite_node): Update.
16623 (input_node): Update.
16624 (input_varpool_node): Update.
16625 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
16626 (lto_symtab_merge_symbols): Add sanity check.
16627 (lto_symtab_prevailing_decl): Do not special case weakrefs.
16628 * passes.c (rest_of_decl_compilation): Set static flag, too.
16629 * symtab.c (dump_symtab_base): Dump weakref.
16630 (verify_symtab_base): Sanity check weakrefs.
16631 (symtab_make_decl_local): Remove duplicated code.
16632 (symtab_alias_ultimate_target): Simplify.
16633 * varpool.c (varpool_create_variable_alias): Set weakref flag.
16635 2013-06-11 Tom de Vries <tom@codesourcery.com>
16637 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
16638 sequence_vect == NULL.
16640 2013-06-11 DJ Delorie <dj@redhat.com>
16642 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
16643 (rl78_unwind_word_mode): New.
16645 2013-06-11 David Malcolm <dmalcolm@redhat.com>
16647 * final.c (debug_prefix_maps): Make static.
16649 2013-06-11 David Malcolm <dmalcolm@redhat.com>
16651 * function.c (initial_trampoline): Remove stray copy.
16653 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
16655 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
16657 2013-06-11 Martin Jambor <mjambor@suse.cz>
16659 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
16660 within bounds at the beginning of the function.
16662 2013-06-11 Alan Modra <amodra@gmail.com>
16664 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
16666 (get_named_section): Don't NULL !DECL_P decl.
16668 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
16670 * doc/invoke.texi (core-avx2): Document.
16672 (atom): Updated with MOVBE.
16674 2013-06-11 Richard Biener <rguenther@suse.de>
16676 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
16678 2013-06-11 Anton Blanchard <anton@samba.org>
16680 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
16681 correct shift value in little-endian mode.
16683 2013-06-11 Jakub Jelinek <jakub@redhat.com>
16686 * varasm.c (get_variable_align): Move #endif to the right place.
16688 2013-06-10 Cary Coutant <ccoutant@google.com>
16690 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
16691 for hash so that hash table traversal order is deterministic.
16693 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
16694 Pat Haugen <pthaugen@us.ibm.com>
16695 Peter Bergner <bergner@vnet.ibm.com>
16697 * config/rs6000/vector.md (GPR move splitter): Do not split moves
16698 of vectors in GPRS if they are direct moves or quad word load or
16701 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
16703 (direct_move_p): Likewise.
16704 (quad_load_store_p): Likewise.
16706 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
16707 classes into bins based on the physical register type.
16708 (reg_class_to_reg_type): Likewise.
16709 (IS_STD_REG_TYPE): Likewise.
16710 (IS_FP_VECT_REG_TYPE): Likewise.
16711 (reload_fpr_gpr): Arrays to determine what insn to use if we can
16712 use direct move instructions.
16713 (reload_gpr_vsx): Likewise.
16714 (reload_vsx_gpr): Likewise.
16715 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
16716 information that is a simplification of register classes. Also
16717 precalculate direct move reload helpers.
16718 (direct_move_p): New function to return true if the operation can
16719 be done as a direct move instruciton.
16720 (quad_load_store_p): New function to return true if the operation
16721 is a quad memory operation.
16722 (rs6000_legitimize_address): If quad memory, only allow register
16723 indirect for TImode addresses.
16724 (rs6000_legitimate_address_p): Likewise.
16725 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
16726 (rs6000_reload_register_type): Likewise.
16727 (register_to_reg_type): Return register type.
16728 (rs6000_secondary_reload_simple_move): New helper function for
16729 secondary reload and secondary memory needed to identify anything
16730 that is a simple move, and does not need reloading.
16731 (rs6000_secondary_reload_direct_move): New helper function for
16732 secondary reload to identify cases that can be done with several
16733 instructions via the direct move instructions.
16734 (rs6000_secondary_reload_move): New helper function for secondary
16735 reload to identify moves between register types that can be done.
16736 (rs6000_secondary_reload): Add support for quad memory operations
16737 and for direct move.
16738 (rs6000_secondary_memory_needed): Likewise.
16739 (rs6000_debug_secondary_memory_needed): Change argument names.
16740 (rs6000_output_move_128bit): New function to return the move to
16741 use for 128-bit moves, including knowing about the various
16742 limitations of quad memory operations.
16744 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
16745 memory operations. call rs6000_output_move_128bit for the actual
16746 instruciton(s) to generate.
16747 (vsx_movti_64bit): Likewise.
16749 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
16750 (UNSPEC_P8V_MTVSRWZ): Likewise.
16751 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
16752 (UNSPEC_P8V_MTVSRD): Likewise.
16753 (UNSPEC_P8V_XXPERMDI): Likewise.
16754 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
16755 (UNSPEC_FUSION_GPR): Likewise.
16756 (FMOVE128_GPR): New iterator for direct move.
16757 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
16758 (f32_sv): Likewise.
16759 (f32_dm): Likewise.
16760 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
16761 loads and direct move instructions.
16762 (zero_extendsidi2_lfiwzx): Likewise.
16763 (extendsidi2_lfiwax): Likewise.
16764 (extendsidi2_nocell): Likewise.
16765 (floatsi<mode>2_lfiwax): Likewise.
16766 (lfiwax): Likewise.
16767 (floatunssi<mode>2_lfiwzx): Likewise.
16768 (lfiwzx): Likewise.
16769 (fix_trunc<mode>_stfiwx): Likewise.
16770 (fixuns_trunc<mode>_stfiwx): Likewise.
16771 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
16772 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
16773 (parity<mode>2_cmpb): Set length/type attr.
16774 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
16775 for 'mr.' to fast_compare.
16776 (bpermd_<mode>): Change type attr to popcnt.
16777 (p8_fmrgow_<mode>): New insns for power8 direct move support.
16778 (p8_mtvsrwz_1): Likewise.
16779 (p8_mtvsrwz_2): Likewise.
16780 (reload_fpr_from_gpr<mode>): Likewise.
16781 (p8_mtvsrd_1): Likewise.
16782 (p8_mtvsrd_2): Likewise.
16783 (p8_xxpermdi_<mode>): Likewise.
16784 (reload_vsx_from_gpr<mode>): Likewise.
16785 (reload_vsx_from_gprsf): Likewise.
16786 (p8_mfvsrd_3_<mode>): LIkewise.
16787 (reload_gpr_from_vsx<mode>): Likewise.
16788 (reload_gpr_from_vsxsf): Likewise.
16789 (p8_mfvsrd_4_disf): Likewise.
16790 (multi-word GPR splits): Do not split direct moves or quad memory
16793 2013-06-10 David Malcolm <dmalcolm@redhat.com>
16795 * tree-into-ssa.c (interesting_blocks): Make static.
16797 2013-06-10 Jakub Jelinek <jakub@redhat.com>
16800 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
16801 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
16802 Use DATA_ABI_ALIGNMENT for that case instead if defined.
16803 (get_variable_align): New function.
16804 (get_variable_section, emit_bss, emit_common,
16805 assemble_variable_contents, place_block_symbol): Use
16806 get_variable_align instead of DECL_ALIGN.
16807 (assemble_noswitch_variable): Add align argument, use it
16808 instead of DECL_ALIGN.
16809 (assemble_variable): Adjust caller. Use get_variable_align
16810 instead of DECL_ALIGN.
16811 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
16813 (DATA_ABI_ALIGNMENT): Define.
16814 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
16815 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
16816 opt is false, only return the psABI mandated alignment increase.
16817 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
16818 (DATA_ABI_ALIGNMENT): ... this.
16819 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
16820 (DATA_ABI_ALIGNMENT): ... this.
16821 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
16822 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
16823 (DATA_ABI_ALIGNMENT): ... this.
16824 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
16825 * doc/tm.texi: Regenerated.
16827 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
16829 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
16830 cmp_code to construct REG_EQUAL note.
16832 2013-06-09 Jakub Jelinek <jakub@redhat.com>
16835 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
16836 that operands[2] doesn't overlap with operands[0].
16838 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
16839 Jan Hubicka <jh@suse.cz>
16841 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
16842 hack to mark symbols as used.
16844 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
16846 PR rtl-optimization/57559
16847 * lra-constraints.c (process_alt_operands): Don't discourage
16848 memory with known offset for offsetable memory constraint.
16849 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
16851 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
16853 * varasm.c (struct oc_local_state): Reorder fields.
16854 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
16855 and adjust accordingly.
16856 (output_constructor): Reorder initialization code and adjust call to
16857 output_constructor_bitfield.
16859 2013-06-07 Jan Hubicka <jh@suse.cz>
16861 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
16863 2013-06-07 David Malcolm <dmalcolm@redhat.com>
16865 * tree-object-size.c (unknown): Make const.
16867 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16869 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
16870 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
16871 for last alternative in the cpu_facility attribute.
16873 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16876 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
16877 (xordi3): Change operand 2 constraint to arm_xordi_operand.
16878 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
16879 * config/arm/constraints.md (Dg): New constraint.
16880 * config/arm/neon.md (xordi3_neon): Remove.
16881 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
16882 * config/arm/predicates.md (arm_xordi_operand): New predicate.
16884 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16886 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
16887 Clean up alternatives.
16889 2013-06-07 Alan Modra <amodra@gmail.com>
16891 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
16894 2013-06-07 Alan Modra <amodra@gmail.com>
16896 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
16898 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16900 * config/arm/constraints.md (Df): New constraint.
16901 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
16902 Correct length attribute for last two alternatives.
16904 2013-06-07 Alan Modra <amodra@gmail.com>
16906 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
16907 override user -mfp-in-toc.
16908 (offsettable_ok_by_alignment): Consider just the current access
16909 rather than the whole object, unless BLKmode. Handle
16910 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
16911 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
16912 for -mcmodel=medium.
16913 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
16914 override user -mfp-in-toc or -msum-in-toc. Default to
16915 -mno-fp-in-toc for -mcmodel=medium.
16917 2013-06-06 DJ Delorie <dj@redhat.com>
16919 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
16920 TARGET_VALID_POINTER_MODE.
16922 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
16923 Pat Haugen <pthaugen@us.ibm.com>
16924 Peter Bergner <bergner@vnet.ibm.com>
16926 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
16927 Document new power8 builtins.
16929 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
16930 condition code register, to allow 128-bit logical operations to be
16931 done in the VSX or GPR registers.
16932 (nor<mode>3): Use the canonical form for nor.
16933 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
16934 vclz*, and vpopcnt* vector instructions.
16935 (nand<mode>3): Likewise.
16936 (orc<mode>3): Likewise.
16937 (clz<mode>2): LIkewise.
16938 (popcount<mode>2): Likewise.
16940 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
16941 that only the GPRs are recognized.
16943 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16944 support for new power8 builtins.
16946 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
16948 (xscvdpspn): Likewise.
16954 (vpopcnt): Likewise.
16955 (vpopcntb): Likewise.
16956 (vpopcnth): Likewise.
16957 (vpopcntw): Likewise.
16958 (vpopcntd): Likewise.
16960 (vmrgew): Likewise.
16961 (vmrgow): Likewise.
16963 (eqv_v16qi3): Likewise.
16964 (eqv_v8hi3): Likewise.
16965 (eqv_v4si3): Likewise.
16966 (eqv_v2di3): Likewise.
16967 (eqv_v4sf3): Likewise.
16968 (eqv_v2df3): Likewise.
16970 (nand_v16qi3): Likewise.
16971 (nand_v8hi3): Likewise.
16972 (nand_v4si3): Likewise.
16973 (nand_v2di3): Likewise.
16974 (nand_v4sf3): Likewise.
16975 (nand_v2df3): Likewise.
16977 (orc_v16qi3): Likewise.
16978 (orc_v8hi3): Likewise.
16979 (orc_v4si3): Likewise.
16980 (orc_v2di3): Likewise.
16981 (orc_v4sf3): Likewise.
16982 (orc_v2df3): Likewise.
16984 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
16985 allow power8 quad mode in 64-bit.
16986 (rs6000_builtin_vectorized_function): Add support to vectorize
16987 ISA 2.07 count leading zeros, population count builtins.
16988 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
16989 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
16990 (builtin_function_type): Add vgbbd builtin function which takes an
16992 (altivec_expand_vec_perm_const): Add support for new power8 merge
16995 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
16996 that does not include TImdoe for use with 32-bit.
16997 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
16999 (UNSPEC_VSX_CVDPSPN): Likewise.
17000 (vsx_xscvdpspn): Likewise.
17001 (vsx_xscvspdpn): Likewise.
17002 (vsx_xscvdpspn_scalar): Likewise.
17003 (vsx_xscvspdpn_directmove): Likewise.
17004 (vsx_and<mode>3): Split logical operations into 32-bit and
17005 64-bit. Add support to do logical operations on TImode as well as
17006 VSX vector types. Allow logical operations to be done in either
17007 VSX registers or in general purpose registers in 64-bit mode. Add
17008 splitters if GPRs were used. For AND, add clobber of CCmode to
17009 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17011 (vsx_and<mode>3_32bit): Likewise.
17012 (vsx_and<mode>3_64bit): Likewise.
17013 (vsx_ior<mode>3): Likewise.
17014 (vsx_ior<mode>3_32bit): Likewise.
17015 (vsx_ior<mode>3_64bit): Likewise.
17016 (vsx_xor<mode>3): Likewise.
17017 (vsx_xor<mode>3_32bit): Likewise.
17018 (vsx_xor<mode>3_64bit): Likewise.
17019 (vsx_one_cmpl<mode>2): Likewise.
17020 (vsx_one_cmpl<mode>2_32bit): Likewise.
17021 (vsx_one_cmpl<mode>2_64bit): Likewise.
17022 (vsx_nor<mode>3): Likewise.
17023 (vsx_nor<mode>3_32bit): Likewise.
17024 (vsx_nor<mode>3_64bit): Likewise.
17025 (vsx_andc<mode>3): Likewise.
17026 (vsx_andc<mode>3_32bit): Likewise.
17027 (vsx_andc<mode>3_64bit): Likewise.
17028 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17029 and xxlorc instructions.
17030 (vsx_eqv<mode>3_64bit): Likewise.
17031 (vsx_nand<mode>3_32bit): Likewise.
17032 (vsx_nand<mode>3_64bit): Likewise.
17033 (vsx_orc<mode>3_32bit): Likewise.
17034 (vsx_orc<mode>3_64bit): Likewise.
17036 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17038 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17040 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17041 (p8_vmrgow): Likewise.
17042 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17043 GPRs to be split under VSX.
17044 (p8v_clz<mode>2): Add power8 count leading zero support.
17045 (p8v_popcount<mode>2): Add power8 population count support.
17046 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17049 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17052 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17054 (vec_nand): Likewise.
17055 (vec_vclz): Likewise.
17056 (vec_vclzb): Likewise.
17057 (vec_vclzd): Likewise.
17058 (vec_vclzh): Likewise.
17059 (vec_vclzw): Likewise.
17060 (vec_vgbbd): Likewise.
17061 (vec_vmrgew): Likewise.
17062 (vec_vmrgow): Likewise.
17063 (vec_vpopcnt): Likewise.
17064 (vec_vpopcntb): Likewise.
17065 (vec_vpopcntd): Likewise.
17066 (vec_vpopcnth): Likewise.
17067 (vec_vpopcntw): Likewise.
17069 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17071 PR rtl-optimization/57468
17072 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17075 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17077 PR rtl-optimization/57459
17078 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17079 type when setting live regs.
17081 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17083 * config/s390/s390.opt (mlra): New option.
17084 * config/s390/s390.c (s390_decompose_address): Check displacement
17085 for all registers for LRA.
17086 (s390_secondary_reload): Don't used secondary reloads for LRA.
17087 (s390_lra_p): New function.
17088 (TARGET_LRA_P): Define.
17089 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17090 of attribute cpu_facility to zarch for the last alternative.
17091 (*cmpmem_short): Ditto.
17093 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17095 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17096 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17097 (arm_expand_prologue): Likewise.
17099 2013-06-06 Teresa Johnson <tejohnson@google.com>
17102 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17103 as this is now done by redirect_edge_and_branch_force.
17104 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17105 barriers, and fix interaction with splitting.
17106 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17107 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17108 reflect changes made in the routine.
17109 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17110 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17111 since this is called in cfglayout mode, and replace partition fixup
17112 with assert as that is now done by force_nonfallthru_and_redirect.
17113 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17114 already be marked with region crossing note.
17115 (insert_section_boundary_note): Make non-static, gate on flag
17116 has_bb_partition, rewrite to also check for multiple partitions.
17117 (rest_of_handle_reorder_blocks): Remove call to
17118 insert_section_boundary_note, now done later during free_cfg.
17119 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17120 * bb-reorder.h (insert_section_boundary_note): Declare.
17121 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17122 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17123 invoke insert_section_boundary_note.
17124 (try_redirect_by_replacing_jump): Remove unnecessary
17125 check for region crossing note.
17126 (fixup_partition_crossing): New function.
17127 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17128 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17129 in non-cfglayout mode.
17130 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17131 remove old code that tried to do this. Emit barrier correctly
17132 when we are in cfglayout mode.
17133 (last_bb_in_partition): New function.
17134 (rtl_split_edge): Correctly fixup partition boundaries.
17135 (commit_one_edge_insertion): Remove old code that tried to
17136 fixup region crossing edge since this is now handled in
17137 split_block, and set up insertion point correctly since
17138 block may now end in a jump.
17139 (verify_hot_cold_block_grouping): Guard against checking when not in
17140 linearized RTL mode.
17141 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17143 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17144 rtl_verify_flow_info, so not called in cfglayout mode.
17145 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17146 (fixup_reorder_chain): Remove old code that attempted to fixup region
17147 crossing note as this is now handled in force_nonfallthru_and_redirect.
17148 (duplicate_insn_chain): Don't duplicate switch section notes.
17149 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17151 * basic-block.h (emit_barrier_after_bb): Declare.
17153 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17155 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17156 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17157 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17159 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17161 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17163 * config/arm/ldmstm.md: Regenerate.
17165 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17167 * config/arm/sync.md (atomic_loaddi_1):
17168 Disable predication for arm_restrict_it.
17169 (arm_load_exclusive<mode>): Likewise.
17170 (arm_load_exclusivesi): Likewise.
17171 (arm_load_exclusivedi): Likewise.
17172 (arm_load_acquire_exclusive<mode>): Likewise.
17173 (arm_load_acquire_exclusivesi): Likewise.
17174 (arm_load_acquire_exclusivedi): Likewise.
17175 (arm_store_exclusive<mode>): Likewise.
17176 (arm_store_exclusive<mode>): Likewise.
17177 (arm_store_release_exclusivedi): Likewise.
17178 (arm_store_release_exclusive<mode>): Likewise.
17180 2013-06-06 Richard Biener <rguenther@suse.de>
17182 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17184 (lto_tag_is_tree_code_p): Adjust.
17185 (lto_tag_is_gimple_code_p): Likewise.
17186 (lto_gimple_code_to_tag): Likewise.
17187 (lto_tag_to_gimple_code): Likewise.
17188 (lto_tree_code_to_tag): Likewise.
17189 (lto_tag_to_tree_code): Likewise.
17190 * data-streamer.h (streamer_write_hwi_in_range): Use
17191 uhwi streaming to stream the normalized range.
17192 (streamer_read_hwi_in_range): Likewise.
17194 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17196 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17197 (predicable_short_it): Likewise.
17198 (predicated): Likewise.
17199 (enabled): Handle above.
17200 (define_cond_exec): Set predicated attribute to yes.
17202 2013-06-05 Mike Stump <mikestump@comcast.net>
17204 * gdbinit.in (__FUNCTION__): Add.
17206 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17208 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17209 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17211 2013-06-05 Jan Hubicka <jh@suse.cz>
17213 * varasm.c (mark_decl_referenced): Revert the removal until targets
17216 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
17218 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
17219 instead of mark_decl_referenced.
17221 2013-06-05 Jan Hubicka <jh@suse.cz>
17223 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
17224 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
17225 and symtab_used_from_object_file_p.
17226 (cgraph_make_node_local_1): Clear forced_by_abi.
17227 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
17228 * cgraph.h (symtab_node_base): Add forced_by_abi.
17229 (decide_is_variable_needed): Remove.
17230 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
17231 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
17232 (decide_is_symbol_needed): ... this one; handle symbols in general;
17233 always analyze virtuals; honnor forced_by_abi.
17234 (cgraph_finalize_function): Update.
17235 (varpool_finalize_decl): Update.
17236 (symbol_defined_and_needed): Remove.
17237 (analyze_functions): Update.
17238 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17239 output_refs, input_overwrite_node): Handle forced_by_abi.
17240 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
17241 (address_taken_from_non_vtable_p): ... this one.
17242 (comdat_can_be_unshared_p_1): New function.
17243 (cgraph_comdat_can_be_unshared_p): Rename to ...
17244 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
17245 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
17246 (function_and_variable_visibility): Clear forced_by_abi as needed.
17247 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
17248 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
17249 * symtab.c (dump_symtab_base): Dump forced_by_abi.
17250 * varpool.c (decide_is_variable_needed): Remove.
17252 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17254 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
17255 (arm_option_override): Override arm_restrict_it where appropriate.
17256 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
17257 * config/arm/arm.opt (mrestrict-it): New command-line option.
17258 * doc/invoke.texi: Document -mrestrict-it.
17260 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17262 * tsan.c (tsan_atomic_table): Make const.
17264 2013-06-05 Richard Biener <rguenther@suse.de>
17266 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
17267 index associated with the tree we are supposed to replace.
17268 * tree-streamer-out.c (pack_ts_base_value_fields): Output
17269 TREE_ASM_WRITTEN as zero for everything but SSA names.
17271 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17273 * tree-ssa-structalias.c (call_stmt_vars): Make static.
17275 2013-06-04 Jan Hubicka <jh@suse.cz>
17277 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
17278 (input_node, input_varpool_node): Handle correctly external same
17280 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
17281 nodes at ltrans stage.
17283 2013-06-04 Jan Hubicka <jh@suse.cz>
17285 * ipa-inline.c (update_caller_keys): Fix availability test.
17286 (update_callee_keys): Likewise.
17287 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
17288 to follow ELF standard.
17290 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
17292 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
17293 (mips64r5900el-*-elf*): New configurations.
17294 * config/mips/mips-cpus.def (r5900): New processor.
17295 * config/mips/mips-tables.opt: Regenerate.
17296 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
17297 (mips_issue_rate): Handle PROCESSOR_R5900.
17298 (mips_reorg_process_insns): Force reorder mode for the R5900.
17299 * config/mips/mips.h (TARGET_MIPS5900): Define.
17300 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
17302 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
17304 * config/mips/mips.md (processor): Add r5900.
17305 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
17307 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17309 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
17310 into function to generate MOVI instruction.
17311 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
17312 (aarch64_preferred_simd_mode): Turn into wrapper.
17313 (aarch64_output_scalar_simd_mov_immediate): New function.
17314 * config/aarch64/aarch64-protos.h: Add prototype for above.
17316 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17318 * config/aarch64/aarch64.c (simd_immediate_info): Remove
17319 element_char member.
17320 (sizetochar): Return signed char.
17321 (aarch64_simd_valid_immediate): Remove elchar and other
17322 unnecessary variables.
17323 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
17324 Calculate element_char as required.
17325 * config/aarch64/aarch64-protos.h: Update and move prototype
17326 for aarch64_output_simd_mov_immediate.
17327 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
17330 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17332 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
17333 information completed by aarch64_simd_valid_immediate.
17334 (aarch64_legitimate_constant_p): Update arguments.
17335 (aarch64_simd_valid_immediate): Work with struct rather than many
17337 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
17338 (aarch64_simd_make_constant): Update arguments.
17339 (aarch64_output_simd_mov_immediate): Work with struct rather than
17340 many pointers. Output immediate directly rather than as operand.
17341 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17343 * config/aarch64/constraints.md (Dn): Update arguments.
17345 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17347 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
17349 (aarch64_simd_immediate_valid_for_move): Remove.
17350 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
17351 (aarch64_simd_make_constant): Update call.
17352 (aarch64_output_simd_mov_immediate): Update call.
17353 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17355 * config/aarch64/constraints.md (Dn): Update call.
17357 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17359 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
17360 return type to bool for prototype.
17361 (aarch64_legitimate_constant_p): Check for true instead of not -1.
17362 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
17363 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
17365 2013-06-04 Catherine Moore <clm@codesourcery.com>
17367 * config/mips/mips.opt (meva): New.
17368 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
17369 (ASM_SPEC): Handle -meva.
17370 * doc/invoke.texi (meva): Document.
17372 2013-06-04 Alan Modra <amodra@gmail.com>
17374 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
17377 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17379 * rtl.def: Add extra fourth optional field to define_cond_exec.
17380 * gensupport.c (process_one_cond_exec): Process attributes from
17382 * doc/md.texi: Document fourth field in define_cond_exec.
17384 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
17386 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
17387 out the processing order as in store_bit_field_1.
17389 2013-06-04 Jan Hubicka <jh@suse.cz>
17391 PR middle-end/57500
17392 * cgraphunit.c (cgraph_process_same_body_aliases): Create
17393 non-VAR_DECL node if it does not exist yet.
17395 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
17397 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
17398 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
17399 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
17400 target_cpu_default setting.
17402 2013-06-03 Teresa Johnson <tejohnson@google.com>
17404 * dumpfile.c (opt_info_switch_p): Change -fopt-info
17405 default to -fopt-info=optimized instead of all.
17406 * doc/invoke.texi: Ditto.
17407 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
17408 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
17409 (execute_vect_slp): Emit BB vectorization success under
17410 MSG_OPTIMIZED_LOCATIONS.
17411 * tree-vect-slp.c (vect_slp_transform_bb): Change
17412 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
17413 * tree-vect-loop.c (vect_transform_loop): Ditto.
17415 2013-06-03 Jason Merrill <jason@redhat.com>
17418 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
17419 Use TARGET_EXPR for C++.
17421 2013-06-03 Jakub Jelinek <jakub@redhat.com>
17423 PR rtl-optimization/57268
17424 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
17425 if DEBUG_INSN_P (insn).
17428 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
17430 PR rtl-optimization/57268
17431 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17432 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17434 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
17436 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
17437 (ix86_avoid_lea_for_addr): Likewise.
17438 (exact_dependency_1): Likewise.
17439 (ix86_adjust_cost): Likewise.
17440 (swap_top_of_ready_list): Fix formatting and !reload_completed check
17442 (do_reorder_for_imul): Fix typo, formatting and
17443 !reload_completed check removed.
17444 (ix86_sched_reorder): Fix typo and formatting.
17445 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
17448 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
17450 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
17452 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
17454 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
17455 <STRING_CST>: Likewise.
17456 <VECTOR_CST>: Likewise.
17458 2013-06-01 Janus Weil <janus@gcc.gnu.org>
17459 Mikael Morin <mikael@gcc.gnu.org>
17461 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
17462 * config.in: Regenerated.
17463 * configure: Regenerated.
17465 2013-06-01 Jan Hubicka <jh@suse.cz>
17467 PR middle-end/57366
17468 * cgraphunit.c (compile): When weakref is not supported,
17469 set up transparent aliases before final output pass.
17470 * varasm.c (assemble_alias): Do not try to do it here.
17472 2013-06-01 Jan Hubicka <jh@suse.cz>
17474 PR middle-end/57467
17475 * passes.c (for_per_function): Skip unanalyzed functions.
17477 2013-06-01 Jan Hubicka <jh@suse.cz>
17479 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
17480 (lto_symtab_merge_symbols_1): ... this one.
17481 (lto_symtab_merge_cgraph_nodes): Rename to ...
17482 (lto_symtab_merge_symbols): ... this one; simplify.
17483 * cgraph.c (same_body_aliases_done): Rename to ...
17484 (cpp_implicit_aliases_done): ... this one.
17485 (cgraph_create_function_alias): Update.
17486 (cgraph_same_body_alias): Update.
17487 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
17488 (verify_edge_corresponds_to_fndecl): Simplify.
17489 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
17490 (cgraph_node): Remove same_body_alias.
17491 (varpool_node): Remove alias_of and extra_name_alias.
17492 (same_body_aliases_done): Rename to ..
17493 (cpp_implicit_aliases_done): ... this one.
17494 (symtab_alias_ultimate_target): Add default parameter.
17495 (symtab_resolve_alias): New function.
17496 (fixup_same_cpp_alias_visibility): Declare.
17497 (cgraph_function_node): Add default parameter.
17498 (cgraph_node_asm_name): Likewise.
17499 (cgraph_function_or_thunk_node): Add default parameter; do
17500 not ICE when it is NULL.
17501 (varpool_variable_node): Likewise.
17502 * tree-emutls.c (create_emultls_var): Update.
17503 (ipa_lower_emutls): Update.
17504 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
17505 (cgraph_reset_node): Reset alias info.
17506 (cgraph_finalize_function): Update.
17507 (fixup_same_cpp_alias_visibility): Move to symtab.c.
17508 (analyze_function): Simplify.
17509 (cgraph_process_same_body_aliases): Simplify.
17510 (analyze_functions): Fixup same body aliases.
17511 (handle_alias_pairs): Simplify.
17512 (assemble_thunk): Update.
17513 (assemble_thunks_and_aliases): Update.
17514 (output_weakrefs): Rewrite.
17515 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
17516 (lto_output_varpool_node): Likewise.
17517 (compute_ltrans_boundary): Remve assert.
17518 (get_alias_symbol): New functoin.
17519 (input_node): Rewrite alias handling.
17520 (input_varpool_node): Likewise.
17521 * ipa-pure-const.c (propagate_pure_const): Fix formating.
17522 * ipa.c (process_references): Handle weakrefs correctly.
17523 (symtab_remove_unreachable_nodes): Likewise.
17524 * trans-mem.c (get_cg_data): Update.
17525 (ipa_tm_create_version_alias): Update.
17526 (ipa_tm_execute): Update.
17527 * symtab.c (dump_symtab_base): Dump aliases.
17528 (verify_symtab_base): Verify aliases.
17529 (symtab_node_availability): New function.
17530 (symtab_alias_ultimate_target): Simplify.
17531 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
17532 handle all the fixup cases.
17533 (symtab_resolve_alias): New function.
17534 * passes.c (ipa_write_summaries): Handle weakrefs.
17535 * varpool.c (varpool_analyze_node): Simplify.
17536 (assemble_aliases): Update.
17537 (varpool_create_variable_alias): Simplify.
17538 (varpool_extra_name_alias): Simplify.
17539 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
17540 (lto_symtab_merge_symbols): ... this one.
17542 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
17545 PR rtl-optimization/57268
17546 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17547 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17549 2013-06-01 Tobias Burnus <burnus@net-b.de>
17551 Partially reverted:
17552 2013-05-31 Tobias Burnus <burnus@net-b.de>
17554 PR middle-end/57073
17555 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
17558 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
17560 PR rtl-optimization/57268
17561 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17562 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17564 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
17566 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
17567 unordered comparison operators when -fno-trapping-math is in effect
17569 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
17570 and implement unordered comparison operators properly on the e500.
17572 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
17574 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
17575 for constant scalar integers.
17576 (simplify_relational_operation_1): Likewise.
17578 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
17580 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
17581 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
17584 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
17585 Igor Zamyatin <igor.zamyatin@intel.com>
17587 Silvermont (SLM) architecture performance tuning.
17588 * config/i386/i386.h (enum ix86_tune_indices): Add
17589 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
17590 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
17592 * config/i386/i386.c (initial_ix86_tune_features)
17593 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
17594 (ix86_lea_outperforms): Handle Silvermont tuning.
17595 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
17597 (ix86_use_lea_for_mov): Likewise.
17598 (ix86_avoid_lea_for_addr): Likewise.
17599 (ix86_lea_for_add_ok): Likewise.
17600 (exact_dependency_1): New function.
17601 (exact_store_load_dependency): Likewise.
17602 (ix86_adjust_cost): Handle Silvermont tuning.
17603 (do_reoder_for_imul): Likewise.
17604 (swap_top_of_ready_list): New function.
17605 (ix86_sched_reorder): Changed to handle Silvermont tuning.
17607 * config/i386/i386.md (peepholes that split memory operand in fp
17610 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
17612 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
17613 Remove un-necessary braces.
17615 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
17617 * config/aarch64/aarch64.c (aarch64_classify_symbol):
17618 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
17620 2013-05-31 Tobias Burnus <burnus@net-b.de>
17622 PR middle-end/57073
17623 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
17625 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17628 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
17629 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
17630 * config/arm/neon.md (iordi3_neon): Remove.
17631 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
17632 * config/arm/predicates.md (imm_for_neon_logic_operand):
17633 Move to earlier in the file.
17634 (neon_logic_op2): Likewise.
17635 (arm_iordi_operand_neon): New predicate.
17637 2013-05-31 Richard Biener <rguenther@suse.de>
17639 PR tree-optimization/57478
17640 PR tree-optimization/57453
17641 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
17644 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17646 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
17647 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
17649 2013-05-30 Tobias Burnus <burnus@net-b.de>
17650 Thomas Koenig <tkoenig@gcc.gnu.org>
17652 PR middle-end/57073
17653 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
17654 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
17656 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
17658 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
17660 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
17662 * target.def (register_usage_leveling_p): New hook.
17663 * targhooks.c (default_register_usage_leveling_p): New.
17664 * targhooks.h (default_register_usage_leveling_p): New prototype.
17665 * lra-assigns.c (register_usage_leveling_p): Use the hook.
17666 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
17667 * doc/tm.texi: Update.
17668 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
17670 2013-05-30 Ian Bolton <ian.bolton@arm.com>
17672 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
17673 (*insv_reg<mode>): New define_insn.
17675 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
17677 PR rtl-optimization/57439
17678 * postreload.c (move2add_valid_value_p): Check that we have
17679 a zero subreg_regno_offset when accessing the register in
17680 the requested mode.
17682 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
17683 Igor Zamyatin <igor.zamyatin@intel.com>
17685 Silvermont (SLM) architecture pipeline model, tuning and
17687 * config.gcc: Add slm config options and target.
17689 * config/i386/slm.md: New.
17691 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
17693 * config/i386/i386-c.c (ix86_target_macros_internal): New case
17695 (ix86_target_macros_internal): Likewise.
17697 * config/i386/i386.c (slm_cost): New cost.
17698 (m_SLM): New macro flag.
17699 (initial_ix86_tune_features): Set m_SLM.
17700 (x86_accumulate_outgoing_args): Likewise.
17701 (x86_arch_always_fancy_math_387): Likewise.
17702 (processor_target_table): Add slm cost.
17703 (cpu_names): Add slm cpu name.
17704 (x86_option_override_internal): Set SLM ISA.
17705 (ix86_issue_rate): New case PROCESSOR_SLM.
17706 (ia32_multipass_dfa_lookahead): Likewise.
17707 (fold_builtin_cpu): Add slm.
17709 * config/i386/i386.h (TARGET_SLM): New target macro.
17710 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
17711 (processor_type): Add PROCESSOR_SLM.
17713 * config/i386/i386.md (cpu): Add new value "slm".
17714 (slm.md): Include slm.md.
17716 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
17717 Zhenqiang Chen <zhenqiang.chen@linaro.org>
17719 * config/arm/arm-protos.h: Add and update function protos.
17720 * config/arm/arm.c (use_simple_return_p): New added.
17721 (thumb2_expand_return): Check simple_return flag.
17722 * config/arm/arm.md: Add simple_return and conditional simple_return.
17723 * config/arm/iterators.md: Add iterator for return and simple_return.
17725 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
17727 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
17728 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
17729 (arm_emit_vfp_multi_reg_pop): Likewise.
17730 (thumb2_emit_ldrd_pop): Likewise.
17731 (arm_expand_epilogue): Add misc REG_CFA notes.
17732 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
17734 2013-05-29 Lawrence Crowl <crowl@google.com>
17736 * config/arm/t-arm: Update for below.
17738 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
17739 Change type to hash_table. Update dependent calls and types.
17741 * config/i386/t-cygming: Update for below.
17743 * config/i386/t-interix: Update for below.
17745 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
17746 Change type to hash_table. Update dependent calls and types.
17747 (i386_find_on_wrapper_list::wrappers): Likewise.
17749 * config/ia64/t-ia64: Update for below.
17751 * config/ia64/ia64.c (bundle_state_table):
17752 Change type to hash_table. Update dependent calls and types.
17754 * config/mips/mips.c (mips_reorg_process_insns::htab):
17755 Change type to hash_table. Update dependent calls and types.
17757 * config/sol2.c (solaris_comdat_htab):
17758 Change type to hash_table. Update dependent calls and types.
17760 * config/t-sol2: Update for above.
17762 2013-05-29 Teresa Johnson <tejohnson@google.com>
17764 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
17765 functions are not yet marked as defined.
17767 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
17768 Pat Haugen <pthaugen@us.ibm.com>
17769 Peter Bergner <bergner@vnet.ibm.com>
17771 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
17775 (vrotl<mode>3): Likewise.
17776 (vashl<mode>3): Likewise.
17777 (vlshr<mode>3): Likewise.
17778 (vashr<mode>3): Likewise.
17780 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17781 support for power8 V2DI builtins.
17783 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
17784 power8 V2DI builtins.
17785 (vupkhsw): Likewise.
17786 (vupklsw): Likewise.
17787 (vaddudm): Likewise.
17788 (vminsd): Likewise.
17789 (vmaxsd): Likewise.
17790 (vminud): Likewise.
17791 (vmaxud): Likewise.
17792 (vpkudum): Likewise.
17793 (vpksdss): Likewise.
17794 (vpkudus): Likewise.
17795 (vpksdus): Likewise.
17800 (vsubudm): Likewise.
17801 (vcmpequd): Likewise.
17802 (vcmpgtsd): Likewise.
17803 (vcmpgtud): Likewise.
17804 (vcmpequd_p): Likewise.
17805 (vcmpgtsd_p): Likewise.
17806 (vcmpgtud_p): Likewise.
17807 (vupkhsw): Likewise.
17808 (vupklsw): Likewise.
17809 (vaddudm): Likewise.
17810 (vmaxsd): Likewise.
17811 (vmaxud): Likewise.
17812 (vminsd): Likewise.
17813 (vminud): Likewise.
17814 (vpksdss): Likewise.
17815 (vpksdus): Likewise.
17816 (vpkudum): Likewise.
17817 (vpkudus): Likewise.
17822 (vsubudm): Likewise.
17824 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
17825 support for power8 V2DI instructions.
17827 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
17828 power8 V2DI instructions. Combine pack and unpack insns to use an
17829 iterator for each mode. Check whether a particular mode supports
17830 Altivec instructions instead of just checking TARGET_ALTIVEC.
17831 (UNSPEC_VPKUWUM): Likewise.
17832 (UNSPEC_VPKSHSS): Likewise.
17833 (UNSPEC_VPKSWSS): Likewise.
17834 (UNSPEC_VPKUHUS): Likewise.
17835 (UNSPEC_VPKSHUS): Likewise.
17836 (UNSPEC_VPKUWUS): Likewise.
17837 (UNSPEC_VPKSWUS): Likewise.
17838 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
17839 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
17840 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
17841 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
17842 (UNSPEC_VUPKHSB): Likewise.
17843 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
17844 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
17845 (UNSPEC_VUPKHSH): Likewise.
17846 (UNSPEC_VUPKLSB): Likewise.
17847 (UNSPEC_VUPKLSH): Likewise.
17849 (VI_char): Likewise.
17850 (VI_scalar): Likewise.
17851 (VI_unit): Likewise.
17853 (VP_small): Likewise.
17854 (VP_small_lc): Likewise.
17855 (VU_char): Likewise.
17856 (add<mode>3): Likewise.
17857 (altivec_vaddcuw): Likewise.
17858 (altivec_vaddu<VI_char>s): Likewise.
17859 (altivec_vadds<VI_char>s): Likewise.
17860 (sub<mode>3): Likewise.
17861 (altivec_vsubcuw): Likewise.
17862 (altivec_vsubu<VI_char>s): Likewise.
17863 (altivec_vsubs<VI_char>s): Likewise.
17864 (altivec_vavgs<VI_char>): Likewise.
17865 (altivec_vcmpbfp): Likewise.
17866 (altivec_eq<mode>): Likewise.
17867 (altivec_gt<mode>): Likewise.
17868 (altivec_gtu<mode>): Likewise.
17869 (umax<mode>3): Likewise.
17870 (smax<mode>3): Likewise.
17871 (umin<mode>3): Likewise.
17872 (smin<mode>3): Likewise.
17873 (altivec_vpkuhum): Likewise.
17874 (altivec_vpkuwum): Likewise.
17875 (altivec_vpkshss): Likewise.
17876 (altivec_vpkswss): Likewise.
17877 (altivec_vpkuhus): Likewise.
17878 (altivec_vpkshus): Likewise.
17879 (altivec_vpkuwus): Likewise.
17880 (altivec_vpkswus): Likewise.
17881 (altivec_vpks<VI_char>ss): Likewise.
17882 (altivec_vpks<VI_char>us): Likewise.
17883 (altivec_vpku<VI_char>us): Likewise.
17884 (altivec_vpku<VI_char>um): Likewise.
17885 (altivec_vrl<VI_char>): Likewise.
17886 (altivec_vsl<VI_char>): Likewise.
17887 (altivec_vsr<VI_char>): Likewise.
17888 (altivec_vsra<VI_char>): Likewise.
17889 (altivec_vsldoi_<mode>): Likewise.
17890 (altivec_vupkhsb): Likewise.
17891 (altivec_vupkhs<VU_char>): Likewise.
17892 (altivec_vupkls<VU_char>): Likewise.
17893 (altivec_vupkhsh): Likewise.
17894 (altivec_vupklsb): Likewise.
17895 (altivec_vupklsh): Likewise.
17896 (altivec_vcmpequ<VI_char>_p): Likewise.
17897 (altivec_vcmpgts<VI_char>_p): Likewise.
17898 (altivec_vcmpgtu<VI_char>_p): Likewise.
17899 (abs<mode>2): Likewise.
17900 (vec_unpacks_hi_v16qi): Likewise.
17901 (vec_unpacks_hi_v8hi): Likewise.
17902 (vec_unpacks_lo_v16qi): Likewise.
17903 (vec_unpacks_hi_<VP_small_lc>): Likewise.
17904 (vec_unpacks_lo_v8hi): Likewise.
17905 (vec_unpacks_lo_<VP_small_lc>): Likewise.
17906 (vec_pack_trunc_v8h): Likewise.
17907 (vec_pack_trunc_v4si): Likewise.
17908 (vec_pack_trunc_<mode>): Likewise.
17910 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
17912 (vec_vmaxsd): Likewise.
17913 (vec_vmaxud): Likewise.
17914 (vec_vminsd): Likewise.
17915 (vec_vminud): Likewise.
17916 (vec_vpksdss): Likewise.
17917 (vec_vpksdus): Likewise.
17918 (vec_vpkudum): Likewise.
17919 (vec_vpkudus): Likewise.
17920 (vec_vrld): Likewise.
17921 (vec_vsld): Likewise.
17922 (vec_vsrad): Likewise.
17923 (vec_vsrd): Likewise.
17924 (vec_vsubudm): Likewise.
17925 (vec_vupkhsw): Likewise.
17926 (vec_vupklsw): Likewise.
17928 2013-05-29 Jan Hubicka <jh@suse.cz>
17930 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
17931 flags; reorder rest of fields in more consistent way.
17932 (varpool_node): Remove analyzed, finalized and alias.
17933 (cgraph_ndoe): Likewise.
17934 (symtab_alias_ultimate_target): New function.
17935 (cgraph_function_node): Move offline.
17936 (cgraph_reset_node): Declare.
17937 (cgraph_comdat_can_be_unshared_p): Remove.
17938 (varpool_remove_initializer): Declare.
17939 (varpool_first_defined_variable, varpool_next_defined_variable
17940 cgraph_first_defined_function, cgraph_next_defined_function): Update.
17941 (cgraph_function_with_gimple_body_p): Update.
17942 (varpool_all_refs_explicit_p): Update.
17943 (symtab_alias_target): New function.
17944 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
17945 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
17946 (cgraph_function_or_thunk_node): Simplify using
17947 symtab_alias_ultimate_target.
17948 (varpool_variable_node): Likewise.
17949 * cgraph.c (cgraph_create_function_alias): Update.
17950 (cgraph_add_thunk): Update.
17951 (cgraph_remove_node): Update.
17952 (dump_cgraph_node): Do not dump removed flags.
17953 (cgraph_function_body_availability): Update.
17954 (cgraph_propagate_frequency): Update.
17955 (verify_cgraph_node): Check sanity of local flag.
17956 (cgraph_function_node): Move here from cgraph.h; revamp for
17957 cgraph_function_or_thunk_node.
17958 * lto-symtab.c (lto_varpool_replace_node): Update.
17959 (lto_symtab_resolve_can_prevail_p): Update.
17960 (lto_symtab_merge_cgraph_nodes): Update.
17961 * ipa-cp.c (determine_versionability, initialize_node_lattices,
17962 propagate_constants_accross_call, devirtualization_time_bonus,
17963 ipcp_propagate_stage): Update.
17964 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
17965 * ipa-inline-transform.c (clone_inlined_nodes,
17966 preserve_function_body_p): Update.
17967 * ipa-reference.c (propagate): Update.
17968 (write_node_summary_p): Update.
17969 * toplev.c (wrapup_global_declaration_2): Update.
17970 * cgraphunit.c (cgraph_analyze_function): Rename to ...
17971 (analyze_function) ... this one.
17972 (cgraph_process_new_functions): Update.
17973 (cgraph_reset_node): Export.
17974 (cgraph_finalize_function): Update.
17975 (cgraph_add_new_function): Update.
17976 (process_function_and_variable_attributes): Update.
17977 (varpool_finalize_decl): Update.
17978 (symbol_finalized): Remove.
17979 (symbol_finalized_and_needed): Rename to ...
17980 (symbol_defined_and_needed): ... update.
17981 (cgraph_analyze_functions): Update.
17982 (handle_alias_pairs): Update.
17983 (mark_functions_to_output): Update.
17984 (assemble_thunk): Update.
17985 (output_in_order): Update.
17986 (output_weakrefs): Update.
17987 (finalize_compilation_unit): Update.
17988 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
17989 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
17990 input_node, input_varpool_node): Update.
17991 * dbxout.c (dbxout_expand_expr): Update.
17992 * cgraphclones.c (cgraph_clone_node): Update.
17993 (cgraph_copy_node_for_versioning): Update.
17994 (cgraph_materialize_clone): Update.
17995 (cgraph_materialize_all_clones): Update.
17996 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
17997 propagate_pure_const, propagate_nothrow): Update.
17998 * lto-streamer-out.c (lto_output, write_symbol): Update.
17999 * ipa-utils.c (ipa_reverse_postorder): Update.
18000 * ipa-inline.c (can_inline_edge_p): Update.
18001 (update_caller_keys, ipa_inline): Update.
18002 * dwarf2out.c (reference_to_unused,
18003 premark_types_used_by_global_vars_helper): Update.
18004 * tree-eh.c (tree_could_trap_p): Update.
18005 * ipa-split.c (consider_split, execute_split_functions): Update.
18006 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18007 has_addr_references_p): Update; move ahead in file for better
18009 (process_references): Simplify.
18010 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18011 bodies are removed.
18012 (cgraph_comdat_can_be_unshared_p): Make static.
18013 (cgraph_externally_visible_p): Update.
18014 (varpool_externally_visible_p): Update.
18015 (function_and_variable_visibility): Update.
18016 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18017 ipa_tm_mark_force_output_node): Update.
18018 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18019 estimate_edge_devirt_benefit, inline_generate_summary,
18020 inline_write_summary): Update.
18021 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18022 * ipa-prop.c (ipa_compute_jump_functions): Update.
18023 (ipa_print_node_params, ipa_prop_read_section,
18024 ipa_update_after_lto_read, read_replacements_section): Update.
18025 * varasm.c (mark_decl_referenced): Update.
18026 (assemble_alias, dump_tm_clone_pairs): Update.
18027 * tree-inline.c (copy_bb): Update.
18028 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18030 * symtab.c (dump_symtab_base): Print new flags.
18031 (verify_symtab_base): Verify new flags.
18032 (symtab_alias_ultimate_target): New function.
18033 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18034 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18036 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18038 * i386.c (ix86_get_function_versions_dispatcher,
18039 ix86_generate_version_dispatcher_body): Update.
18040 (fold_builtin_cpu): Use varpool_add_new_variable.
18041 * varpool.c (varpool_remove_initializer): Break out from ...
18042 (varpool_remove_node): ... this one.
18043 (dump_varpool_node, varpool_node_for_asm,
18044 cgraph_variable_initializer_availability, varpool_analyze_node,
18045 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18046 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18049 2013-05-29 Jan Hubicka <jh@suse.cz>
18051 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18053 2013-05-29 Easwaran Raman <eraman@google.com>
18055 PR tree-optimization/57442
18056 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18057 when control exits the main loop.
18059 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18061 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18063 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18064 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18065 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18068 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18070 PR tree-optimization/57441
18071 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18072 Don't limit size of incr_vec to number of candidates.
18074 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18076 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18077 and mips16 directories.
18078 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18079 (MULTILIB_DIRNAMES): Ditto.
18080 (MULTILIB_EXCEPTIONS): Add new exceptions.
18081 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18082 (MULTILIB_DIRNAMES): Ditto.
18083 (MULTILIB_EXCEPTIONS): Add new exceptions.
18085 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18086 Marcus Shawcroft <marcus.shawcroft@arm.com>
18088 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18089 SYMBOL_TINY_ABSOLUTE.
18090 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18091 SYMBOL_TINY_ABSOLUTE.
18092 (aarch64_expand_mov_immediate): Likewise.
18093 (aarch64_classify_symbol): Likewise.
18094 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18095 Permit SYMBOL_TINY_ABSOLUTE.
18096 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18098 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18099 Marcus Shawcroft <marcus.shawcroft@arm.com>
18101 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18102 Refactor if/switch. Replace gcc_assert with if.
18104 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18106 * config/i386/i386.c (initial_ix86_tune_features): Enable
18107 FP Reassociation for AMD bdver1 and bdver2.
18109 2013-05-29 Martin Jambor <mjambor@suse.cz>
18111 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18112 and IMAGPART_EXPR do not occur within other handled_components.
18114 2013-05-29 Richard Biener <rguenther@suse.de>
18116 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18117 access on whether the use is in the BB we currently try to
18119 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18120 vectorize to vect_bb_slp_scalar_cost.
18122 2013-05-29 Richard Biener <rguenther@suse.de>
18124 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18125 computing scalar cost offsetted by stmts that are kept live
18127 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18128 for computation of scalar cost.
18130 2013-05-28 Steve Ellcey <sellcey@mips.com>
18132 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18134 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18136 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18137 array notation built-in reduction functions.
18138 * doc/passes.texi (Passes): Added documentation about changes done
18140 * doc/invoke.texi (C Dialect Options): Added documentation about
18141 the -fcilkplus flag.
18142 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18143 (BUILTINS_DEF): Depend on cilkplus.def.
18144 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18145 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18146 * cilkplus.def: New file.
18148 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18150 PR rtl-optimization/57439
18151 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18153 2013-05-28 Easwaran Raman <eraman@google.com>
18155 PR tree-optimization/57337
18156 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18157 (find_insert_point): Correctly identify the insertion point
18158 when two statements with the same UID is compared.
18160 2013-05-28 Richard Biener <rguenther@suse.de>
18162 PR tree-optimization/56787
18163 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18164 from the list of data references.
18165 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18167 (vect_analyze_loop_operations): Likewise.
18168 (vect_transform_loop): Remove clobbers.
18170 2013-05-28 Martin Jambor <mjambor@suse.cz>
18172 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18173 and REALPART_EXPRs have scalar type.
18175 2013-05-28 Richard Biener <rguenther@suse.de>
18177 PR tree-optimization/57411
18178 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18180 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18181 virtual operand propagation.
18183 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18185 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18186 destination register for bmasksi_vis.
18187 (vector_init_bshuffle): Likewise.
18188 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18190 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18192 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18193 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18194 mode if the instruction isn't available in the original mode.
18195 * config/sparc/sparc.opt (mfix-ut699): New option.
18196 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18197 (divdf3): Turn into expander.
18198 (divdf3_nofix): New insn.
18199 (divdf3_fix): Likewise.
18200 (divsf3): Disable if -mfix-ut699.
18201 (sqrtdf2): Turn into expander.
18202 (sqrtdf2_nofix): New insn.
18203 (sqrtdf2_fix): Likewise.
18204 (sqrtsf2): Disable if -mfix-ut699.
18206 2013-05-27 Richard Biener <rguenther@suse.de>
18208 PR middle-end/57412
18209 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18210 block for the new loop.
18212 2013-05-27 Richard Biener <rguenther@suse.de>
18214 PR tree-optimization/57343
18215 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
18216 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
18217 (number_of_iterations_cond): Do not build the folded tree.
18219 2013-05-27 Richard Biener <rguenther@suse.de>
18222 PR middle-end/57381
18223 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18224 OEP_CONSTANT_ADDRESS_OF retained.
18226 PR tree-optimization/57417
18227 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
18228 for unchanged base.
18229 (set_ssa_val_to): Compare addresses using
18230 get_addr_base_and_unit_offset.
18232 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
18234 PR rtl-optimization/56833
18235 * postreload.c (move2add_record_mode): New function.
18236 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
18237 (move2add_use_add2_insn): Use move2add_record_sym_value.
18238 (move2add_use_add3_insn): Likewise.
18239 (reload_cse_move2add): Use move2add_valid_value_p and
18240 move2add_record_mode. Invalidate call-clobbered and REG_INC
18241 affected regs by setting reg_mode to VOIDmode.
18242 (move2add_note_store): Don't pretend the inside of a SUBREG is
18243 the actual destination. Invalidate single/leading registers by
18244 setting reg_mode to VOIDmode.
18245 Use move2add_record_sym_value, move2add_valid_value_p and
18246 move2add_record_mode.
18248 2013-05-27 Richard Biener <rguenther@suse.de>
18250 PR tree-optimization/57396
18251 * tree-affine.c (double_int_constant_multiple_p): Properly
18252 return false for val == 0 and div != 0.
18254 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18256 * config/mips/mips.h: Use #elif in preprocessor conditions.
18258 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18261 * config/mips/constraints.md (kl): New constraint.
18262 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
18263 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
18264 constraint for operand 0. Split after CSE for MIPS16. Emit a move
18265 from LO for MIPS16.
18266 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
18268 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18271 * config/mips/mips.c (mips_can_inline_p): New function.
18272 (TARGET_CAN_INLINE_P): Define.
18274 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
18276 * sched-int.h (ds_t, dw_t): Make unsigned int.
18277 Fix documentation that describes how all the ds_t bits are used.
18278 Reserve the last bit for delayed-branch scheduling.
18279 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
18280 (BITS_PER_DEP_WEAK): Fix definition and documentation.
18281 (gen_dep_weak_1): Remove prototype.
18282 * sched-deps.c (get_dep_weak_1): Make static.
18283 * target.def (speculate_insn, needs_block_p, gen_spec_check,
18284 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
18285 * doc/tm.texi: Regenerate.
18286 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
18288 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
18291 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
18293 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
18294 Sandra Loosemore <sandra@codesourcery.com>
18296 * config.gcc (powerpc-*): Allow native for with-cpu.
18298 2013-05-24 Jeff Law <law@redhat.com>
18300 PR tree-optimization/57124
18301 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
18302 conversion feeding a condition if the range has an overflow
18303 if -fstrict-overflow. Add warnings for when we do make the
18306 2013-05-24 Dehao Chen <dehao@google.com>
18308 * tree-cfg.c (locus_discrim_map): Fix the typo.
18309 (locus_discrim_hasher): Likewise.
18310 (locus_discrim_hasher::hash): Likewise.
18311 (locus_discrim_hasher::equal): Likewise.
18313 2013-05-24 Martin Jambor <mjambor@suse.cz>
18315 PR tree-optimization/57294
18316 * cgraph.h (ipa_record_stmt_references): Declare.
18317 * cgraphbuild.c (ipa_record_stmt_references): New function.
18318 (build_cgraph_edges): Use ipa_record_stmt_references.
18319 (rebuild_cgraph_edges): Likewise.
18320 (cgraph_rebuild_references): Likewise.
18321 * ipa-prop.c (ipa_modify_call_arguments): Discard references
18322 associated with the old statement and build references from the
18323 newly built statements.
18324 * ipa-ref.c (ipa_remove_stmt_references): New function.
18325 * ipa-ref.h (ipa_remove_stmt_references): Declare.
18327 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
18329 * lra-constraints.c (emit_spill_move): Use smaller mode for
18331 (check_and_process_move): Consider mem-reg moves for secondary
18333 (curr_insn_transform): Don't lose insns emitted before for
18334 secondary memory moves.
18335 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
18336 reg set up in the current insn.
18338 2013-05-24 Dehao Chen <dehao@google.com>
18340 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
18342 (locus_descrim_hasher::equal): Likewise.
18343 (build_gimple_cfg): New discriminator assignment algorithm.
18344 (make_edges): Likewise.
18345 (next_discriminator_for_locus): Likewise.
18346 (same_line_p): Likewise.
18347 (assign_discriminators): Likewise.
18348 (make_cond_expr_edges): Likewise.
18349 (make_gimple_switch_edges): Likewise.
18350 (make_goto_expr_edges): Likewise.
18351 (make_gimple_asm_edges): Likewise.
18353 2013-05-24 Ian Bolton <ian.bolton@arm.com>
18355 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
18356 X format specifier to only display bottom 16 bits.
18357 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
18358 immediate to match for operand 2, since it will be masked.
18360 2013-05-24 Richard Biener <rguenther@suse.de>
18362 PR tree-optimization/57287
18363 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
18364 all SSA names that occur in abnormal PHIs.
18366 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
18369 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
18370 that index is not negative.
18372 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
18374 PR rtl-optimization/55177
18375 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
18376 (simplify_byte_swapping_operation): New.
18377 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
18378 (simplify_relational_operation_1): Deal with BSWAP.
18380 2013-05-23 Richard Henderson <rth@redhat.com>
18383 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
18384 (ix86_reorg): Call it.
18386 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
18389 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
18390 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
18391 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
18393 2013-05-23 Christian Bruel <christian.bruel@st.com>
18396 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
18398 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18399 Marcus Shawcroft <marcus.shawcroft@arm.com>
18401 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
18402 * config/aarch64/constraints.md (Usa): Remove.
18403 * doc/md.texi (AArch64 Usa): Remove.
18405 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18406 Marcus Shawcroft <marcus.shawcroft@arm.com>
18408 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
18409 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
18410 * config/aarch64/predicates.md (aarch64_const_address): Remove.
18411 (aarch64_mov_operand): Use aarch64_mov_operand_p.
18413 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
18415 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
18416 instruction (AdvSIMD).
18417 * config/aarch64/aarch64-builtins.c
18418 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
18419 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
18421 2013-05-23 Martin Jambor <mjambor@suse.cz>
18423 PR middle-end/57347
18424 * tree.h (contains_bitfld_component_ref_p): Declare.
18425 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
18426 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
18428 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
18429 not access a bit-field. Assert all final offsets are byte-aligned.
18431 2013-05-23 Richard Biener <rguenther@suse.de>
18433 PR tree-optimization/57380
18434 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
18435 least one invariant or re-used load.
18436 * passes.c (init_optimization_passes): Move pass_phiprop before
18439 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
18441 * config/aarch64/aarch64-simd.md
18442 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
18444 2013-05-23 Richard Biener <rguenther@suse.de>
18446 PR middle-end/57381
18447 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18448 OEP_CONSTANT_ADDRESS_OF retained.
18450 2013-05-23 Jakub Jelinek <jakub@redhat.com>
18452 PR middle-end/57344
18453 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
18454 don't lower unit. Handle unit not being always BITS_PER_WORD.
18456 2013-05-23 Richard Biener <rguenther@suse.de>
18458 PR rtl-optimization/57341
18459 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
18460 instead of true_dependence.
18462 2013-05-22 David Malcolm <dmalcolm@redhat.com>
18464 * bb-reorder.c (branch_threshold): Make const.
18465 (exec_threshold): Ditto.
18467 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
18468 Pat Haugen <pthaugen@us.ibm.com>
18469 Peter Bergner <bergner@vnet.ibm.com>
18471 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
18472 documentation for the power8 crypto builtins.
18474 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
18476 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
18477 macros for defining power8 builtin functions.
18478 (BU_P8V_AV_2): Likewise.
18479 (BU_P8V_AV_P): Likewise.
18480 (BU_P8V_VSX_1): Likewise.
18481 (BU_P8V_OVERLOAD_1): Likewise.
18482 (BU_P8V_OVERLOAD_2): Likewise.
18483 (BU_CRYPTO_1): Likewise.
18484 (BU_CRYPTO_2): Likewise.
18485 (BU_CRYPTO_3): Likewise.
18486 (BU_CRYPTO_OVERLOAD_1): Likewise.
18487 (BU_CRYPTO_OVERLOAD_2): Likewise.
18488 (XSCVSPDP): Fix typo, point to the correct instruction.
18489 (VCIPHER): Add power8 crypto builtins.
18490 (VCIPHERLAST): Likewise.
18491 (VNCIPHER): Likewise.
18492 (VNCIPHERLAST): Likewise.
18493 (VPMSUMB): Likewise.
18494 (VPMSUMH): Likewise.
18495 (VPMSUMW): Likewise.
18496 (VPERMXOR_V2DI): Likewise.
18497 (VPERMXOR_V4SI: Likewise.
18498 (VPERMXOR_V8HI: Likewise.
18499 (VPERMXOR_V16QI: Likewise.
18500 (VSHASIGMAW): Likewise.
18501 (VSHASIGMAD): Likewise.
18502 (VPMSUM): Likewise.
18503 (VPERMXOR): Likewise.
18504 (VSHASIGMA): Likewise.
18506 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
18507 __CRYPTO__ if the crypto instructions are available.
18508 (altivec_overloaded_builtins): Add support for overloaded power8
18511 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
18512 support for power8 crypto builtins.
18513 (builtin_function_type): Likewise.
18514 (altivec_init_builtins): Add support for builtins that take vector
18515 long long (V2DI) arguments.
18517 * config/rs6000/crypto.md: New file, define power8 crypto
18520 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
18521 Pat Haugen <pthaugen@us.ibm.com>
18522 Peter Bergner <bergner@vnet.ibm.com>
18524 * doc/invoke.texi (Option Summary): Add power8 options.
18525 (RS/6000 and PowerPC Options): Likewise.
18527 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
18528 constraints.md instead of rs6000.h. Reorder w* constraints. Add
18529 wm, wn, wr documentation.
18531 * config/rs6000/constraints.md (wm): New constraint for VSX
18532 registers if direct move instructions are enabled.
18533 (wn): New constraint for no registers.
18534 (wq): New constraint for quad word even GPR registers.
18535 (wr): New constraint if 64-bit instructions are enabled.
18536 (wv): New constraint if power8 vector instructions are enabled.
18537 (wQ): New constraint for quad word memory locations.
18539 * config/rs6000/predicates.md (const_0_to_15_operand): New
18540 constraint for 0..15 for crypto instructions.
18541 (gpc_reg_operand): If VSX allow registers in VSX registers as well
18542 as GPR and floating point registers.
18543 (int_reg_operand): New predicate to match only GPR registers.
18544 (base_reg_operand): New predicate to match base registers.
18545 (quad_int_reg_operand): New predicate to match even GPR registers
18546 for quad memory operations.
18547 (vsx_reg_or_cint_operand): New predicate to allow vector logical
18548 operations in both GPR and VSX registers.
18549 (quad_memory_operand): New predicate for quad memory operations.
18550 (reg_or_indexed_operand): New predicate for direct move support.
18552 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
18553 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
18554 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
18555 (POWERPC_MASKS): Add power8 options.
18556 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
18559 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
18560 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
18562 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
18563 (-mpower8-fusion): New power8 options.
18564 (-mpower8-fusion-sign): Likewise.
18565 (-mpower8-vector): Likewise.
18566 (-mcrypto): Likewise.
18567 (-mdirect-move): Likewise.
18568 (-mquad-memory): Likewise.
18570 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
18571 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
18573 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
18574 (rs6000_debug_vector_unit): Add p8_vector.
18575 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
18576 definitions. Also print fusion state.
18577 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
18578 (rs6000_builtin_mask_calculate): Add power8 builtin support.
18579 (rs6000_option_override_internal): Add support for power8.
18580 (rs6000_common_init_builtins): Add debugging for skipped builtins
18581 if -mdebug=builtin.
18582 (rs6000_adjust_cost): Add power8 support.
18583 (rs6000_issue_rate): Likewise.
18584 (insn_must_be_first_in_group): Likewise.
18585 (insn_must_be_last_in_group): Likewise.
18586 (force_new_group): Likewise.
18587 (rs6000_register_move_cost): Likewise.
18588 (rs6000_opt_masks): Likewise.
18590 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
18591 power8 capable assembler, default to power7 options.
18592 (TARGET_DIRECT_MOVE): Likewise.
18593 (TARGET_CRYPTO): Likewise.
18594 (TARGET_P8_VECTOR): Likewise.
18595 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
18596 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
18597 (VECTOR_MEM_P8_VECTOR_P): Likewise.
18598 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
18599 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
18600 (TARGET_XSCVDPSPN): Likewise.
18601 (TARGET_XSCVSPDPN): Likewsie.
18602 (TARGET_SYNC_HI_QI): Likewise.
18603 (TARGET_SYNC_TI): Likewise.
18604 (MASK_CRYPTO): Likewise.
18605 (MASK_DIRECT_MOVE): Likewise.
18606 (MASK_P8_FUSION): Likewise.
18607 (MASK_P8_VECTOR): Likewise.
18608 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
18609 temporary used by some of the direct move instructions to get two FP
18610 temporary registers does not force creation of a stack frame.
18611 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
18612 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
18613 that any VSX registers are tieable, even if they are also an
18614 Altivec vector mode.
18615 (r6000_reg_class_enum): Add wm, wr, wv constraints.
18616 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
18617 (RS6000_BTM_CRYPTO): Likewise.
18618 (RS6000_BTM_COMMON): Likewise.
18620 * config/rs6000/rs6000.md (cpu attribute): Add power8.
18621 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
18622 (enum rs6000_vector): Add power8 vector support.
18624 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18628 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
18629 (any_sibcall_could_use_r3): this and handle indirect calls.
18630 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
18632 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18634 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
18636 2013-05-22 Richard Biener <rguenther@suse.de>
18638 PR middle-end/57349
18639 * profile.c (branch_prob): Do not split blocks that are
18640 abnormally receiving from ECF_RETURNS_TWICE functions.
18642 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18644 * recog.c (offsettable_address_addr_space_p): Fix calculation of
18645 address mode. Move pointer mode initialization to the same place.
18647 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
18649 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
18650 while it has any effect.
18652 2013-05-21 Easwaran Raman <eraman@google.com>
18654 PR tree-optimization/57322
18655 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
18656 UID of the statement added to the BB to be 1.
18658 2013-05-21 Jakub Jelinek <jakub@redhat.com>
18660 PR tree-optimization/57331
18661 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
18662 of conversion from pointer type to integral type with integer.
18664 2013-05-21 Martin Jambor <mjambor@suse.cz>
18667 * ipa-prop.c (ipa_read_node_info): Process param_used and
18668 controlled_uses in the same order as when writing.
18670 2013-05-21 Magnus Granberg <baldrick@free.fr>
18673 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
18675 2013-05-21 Richard Biener <rguenther@suse.de>
18677 PR tree-optimization/57318
18678 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
18679 estimate stmts with side-effects as likely eliminated.
18681 2013-05-21 Richard Biener <rguenther@suse.de>
18683 PR tree-optimization/57330
18684 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
18685 preserve the call stmts fntype.
18687 2013-05-21 Richard Biener <rguenther@suse.de>
18689 PR tree-optimization/57303
18690 * tree-ssa-sink.c (statement_sink_location): Improve killing
18691 stmt detection and properly handle self-assignments.
18693 2013-05-21 Christian Bruel <christian.bruel@st.com>
18695 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
18696 spanning registers. LEAF_REG_REMAP is supported only for contiguous
18697 registers. Set register size out of the PARALLEL loop.
18699 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
18702 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
18703 (*fmasf4, *fmasf4_media): New insns.
18705 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
18707 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
18708 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
18709 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
18710 (mips_idiv_insns): Update the comments to say that the returned
18711 instruction counts are in units of BASE_INSN_LENGTH.
18712 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
18713 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
18714 using 2 rather than 4 as the length of indirect MIPS16 and
18715 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
18716 length of a NOP. Don't divide MIPS16 lengths by 2.
18717 (mips16_split_long_branches): Assume a branch is long if the
18718 length is greater than 4 rather than 8.
18719 * config/mips/mips.md (length): Give MIPS16 lengths directly,
18720 rather than multiplying them by 2. Multiply instruction counts
18721 by BASE_INSN_LENGTH rather than 4.
18722 (*jump_mips16, tls_get_tp_mips16_<mode>)
18723 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
18725 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
18727 * config/mips/mips.md (extended_mips16): Remove branch case.
18728 (length): Remove duplicated extended_mips16 test.
18730 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
18732 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
18734 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
18736 * recog.h (Recog_data): Rename to...
18737 (recog_data_d): ...this.
18738 (recog_data): Update accordingly.
18739 * recog.c (recog_data): Likewise.
18740 * reload.c (save_recog_data): Likewise.
18741 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
18742 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
18744 2013-05-17 Julian Brown <julian@codesourcery.com>
18746 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
18747 found in a REG_EQUAL note, invalidate it.
18749 2013-05-17 Easwaran Raman <eraman@google.com>
18751 * tree-ssa-reassoc.c (find_insert_point): New function.
18752 (insert_stmt_after): Likewise.
18753 (get_def_stmt): Likewise.
18754 (ensure_ops_are_available): Likewise.
18755 (not_dominated_by): Likewise.
18756 (rewrite_expr_tree): Do not move statements beyond what is
18757 necessary. Remove call to swap_ops_for_binary_stmt...
18758 (reassociate_bb): ... and move it here.
18759 (build_and_add_sum): Assign UIDs for new statements.
18760 (linearize_expr): Likewise.
18761 (do_reassoc): Renumber gimple statement UIDs.
18763 2013-05-17 Jan Hubicka <jh@suse.cz>
18765 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
18767 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
18768 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
18770 (output_weakrefs): Update.
18772 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
18773 Martin Jambor <mjambor@suse.cz>
18775 PR middle-end/57276
18776 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
18777 value that corresponds to the given aggval is found in values vector.
18779 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
18781 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
18782 sse, sse2, sse3, ssse3 and sse4a flags to options.
18784 2013-05-17 David Malcolm <dmalcolm@redhat.com>
18786 * gengtype-state.c: (s_expr_writer): New class, to handle
18787 prettifying of output layout of s-expressions.
18788 (state_writer): New class, to write out gtype.state.
18789 (state_written_type_count): Move this variable into member data of
18791 (s_expr_writer::s_expr_writer): New code: constructor for new class
18792 (state_writer::state_writer(): ditto
18793 (s_expr_writer::write_new_line): New function
18794 (s_expr_writer::write_any_indent): ditto
18795 (s_expr_writer::begin_s_expr): ditto
18796 (s_expr_writer::end_s_expr): ditto
18797 (write_state_fileloc): convert to method of state_writer...
18798 (state_writer:: write_state_fileloc): ...and use methods of
18799 s_expr_writer to write indentation into the gtype.state output file
18800 to visually represent the hierarchical structure of the list
18802 (write_state_fields): ditto, renaming to...
18803 (state_writer::write_state_fields)
18804 (write_state_a_string): ditto, renaming to...
18805 (state_writer::write_state_a_string)
18806 (write_state_string_option): ditto, renaming to...
18807 (state_writer::write_state_string_option)
18808 (write_state_type_option): ditto, renaming to...
18809 (state_writer::write_state_type_option)
18810 (write_state_nested_option): ditto, renaming to...
18811 (state_writer::write_state_nested_option)
18812 (write_state_option): ditto, renaming to...
18813 (state_writer::write_state_option)
18814 (write_state_options): ditto, renaming to...
18815 (state_writer::write_state_options)
18816 (write_state_lang_bitmap): ditto, renaming to...
18817 (state_writer::write_state_lang_bitmap)
18818 (write_state_version): ditto, renaming to...
18819 (state_writer::write_state_version)
18820 (write_state_scalar_type): ditto, renaming to...
18821 (state_writer::write_state_scalar_type)
18822 (write_state_string_type): ditto, renaming to...
18823 (state_writer::write_state_string_type)
18824 (write_state_undefined_type): ditto, renaming to...
18825 (state_writer::write_state_undefined_type)
18826 (write_state_struct_union_type): ditto, renaming to...
18827 (state_writer::write_state_struct_union_type)
18828 (write_state_struct_type): ditto, renaming to...
18829 (state_writer::write_state_struct_type)
18830 (write_state_user_struct_type): ditto, renaming to...
18831 (state_writer::write_state_user_struct_type)
18832 (write_state_lang_struct_type): ditto, renaming to...
18833 (state_writer::write_state_lang_struct_type)
18834 (write_state_param_struct_type): ditto, renaming to...
18835 (state_writer::write_state_param_struct_type)
18836 (write_state_pointer_type): ditto, renaming to...
18837 (state_writer::write_state_pointer_type)
18838 (write_state_array_type): ditto, renaming to...
18839 (state_writer::write_state_array_type)
18840 (write_state_gc_used): ditto, renaming to...
18841 (state_writer::write_state_gc_used)
18842 (write_state_common_type_content): ditto, renaming to...
18843 (state_writer::write_state_common_type_content)
18844 (write_state_type): ditto, renaming to...
18845 (state_writer::write_state_type)
18846 (write_state_pair_list): ditto, renaming to...
18847 (state_writer::write_state_pair_list)
18848 (write_state_pair): ditto, renaming to...
18849 (state_writer::write_state_pair)
18850 (write_state_typedefs): ditto, renaming to...
18851 (state_writer::write_state_typedefs)
18852 (write_state_structures): ditto, renaming to...
18853 (state_writer::write_state_structures)
18854 (write_state_param_structs): ditto, renaming to...
18855 (state_writer::write_state_param_structs)
18856 (write_state_variables): ditto, renaming to...
18857 (state_writer::write_state_variables)
18858 (write_state_srcdir): ditto, renaming to...
18859 (state_writer::write_state_srcdir)
18860 (write_state_files_list): ditto, renaming to...
18861 (state_writer::write_state_files_list)
18862 (write_state_languages): ditto, renaming to...
18863 (state_writer::write_state_languages)
18864 (write_state): create a state_writer instance and use it when
18865 writing out the state file
18867 2013-05-17 Mike Stump <mikestump@comcast.net>
18869 PR rtl-optimization/57304
18870 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
18871 accessing DF_REF_REAL_LOC.
18873 2013-05-17 Jakub Jelinek <jakub@redhat.com>
18875 PR rtl-optimization/57281
18876 PR rtl-optimization/57300
18877 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
18878 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
18879 what the other splitter did if the registers are dead.
18881 2013-05-17 Richard Biener <rguenther@suse.de>
18883 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
18886 2013-05-17 Jakub Jelinek <jakub@redhat.com>
18888 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
18891 2013-05-17 Marek Polacek <polacek@redhat.com>
18893 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
18894 length when doing non-zero store of storing '\0' to '\0'.
18896 2013-05-17 Jakub Jelinek <jakub@redhat.com>
18898 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
18899 vect_external_def oprnd1 with loop_vinfo, try to emit
18900 optional cast, negation and and stmts on the loop preheader
18901 edge instead of into the pattern def seq.
18903 PR tree-optimization/57051
18904 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
18905 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
18907 2013-05-16 Nick Clifton <nickc@redhat.com>
18909 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
18910 (rl78_is_naked_func): New function.
18911 (rl78_expand_prologue): Skip prologue generation for naked functions.
18912 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
18913 * doc/extend.texi (naked): Add RL78 to the list of processors
18914 that supports this attribute.
18916 2013-05-16 Jeff Law <law@redhat.com>
18918 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
18920 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
18922 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
18923 cache parameters using detect_caches_amd also for CYRIX,
18924 NSC and TM2 signatures.
18926 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
18927 Dzianis Kahanovich <mahatma@eu.by>
18931 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
18932 VIA/Centaur processors and determine their cache parameters
18933 using detect_caches_amd.
18935 2013-05-16 Teresa Johnson <tejohnson@google.com>
18937 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
18938 (rtl_verify_edges): New function.
18939 (rtl_verify_bb_insns): Ditto.
18940 (rtl_verify_bb_pointers): Ditto.
18941 (rtl_verify_bb_insn_chain): Ditto.
18942 (rtl_verify_fallthru): Ditto.
18943 (rtl_verify_bb_layout): Ditto.
18944 (rtl_verify_flow_info_1): Outline checks into new functions.
18945 (rtl_verify_flow_info): Ditto.
18947 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
18949 * cfghooks.c (copy_bbs): Add update_dominance argument.
18950 * cfghooks.h (copy_bbs): Update prototype.
18951 * tree-cfg.c (gimple_duplicate_sese_region):
18952 Add update_dominance argument.
18953 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
18954 * tree-ssa-loop-ch.c (copy_loop_headers): Update
18955 gimple_duplicate_sese_region call.
18956 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
18957 Update copy_bbs call.
18958 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
18959 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
18961 2013-05-16 Jakub Jelinek <jakub@redhat.com>
18963 * tree-vectorizer.h (NUM_PATTERNS): Increment.
18964 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
18965 vect_recog_rotate_pattern.
18966 (vect_recog_rotate_pattern): New function.
18968 2013-05-16 Jason Merrill <jason@redhat.com>
18970 * Makefile.in (LLINKER): New variable.
18971 (mostlyclean): Remove link mutex.
18972 * configure.ac: Handle --enable-link-mutex.
18973 * lock-and-run.sh: New script.
18975 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18978 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
18981 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18983 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
18985 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
18987 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
18988 * config/arm/arm.c (next_consecutive_mem): New function.
18989 (gen_movmem_ldrd_strd): Likewise.
18990 * config/arm/arm.md (movmemqi): Update condition and code.
18991 (unaligned_loaddi, unaligned_storedi): New patterns.
18993 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18995 * config.gcc: Obsolete *-*-solaris2.9*.
18996 * doc/install.texi (Specific, *-*-solaris2*): Document it.
18998 2013-05-16 Richard Biener <rguenther@suse.de>
19000 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19001 earlier, after GRAPHITE transforms and IV canonicalization.
19003 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19005 * omp-low.c (extract_omp_for_data): For collapsed loops,
19006 if at least one of the loops is known at compile time to
19007 iterate zero times, set count to 0.
19008 (expand_omp_regimplify_p): New function.
19009 (expand_omp_for_generic): For collapsed loops, if at least
19010 one of the loops isn't known to iterate at least once,
19011 add runtime check with setting count to 0.
19012 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19013 For unsigned types if it isn't known at compile time that
19014 the loop will iterate at least once, add runtime check to bypass
19015 the whole loop if initial condition isn't true.
19017 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19019 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19021 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19023 PR middle-end/57286
19024 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19025 transformations to avoid an infinite loop.
19027 2013-05-16 Marek Polacek <polacek@redhat.com>
19029 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19031 2013-05-15 Leif Ekblad <leif@rdos.net>
19033 * config/i386/i386.c (ix86_decompose_address): Use
19034 DEFAULT_TLS_SEG_REG to access TLS segment register.
19035 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19036 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19037 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19039 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19042 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19043 sibling calls to functions that would normally be lazily bound,
19044 unless $gp is call-clobbered.
19046 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19048 * config/i386/i386.c (ix86_option_override_internal): Update
19049 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19050 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19051 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19052 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19054 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19056 2013-05-15 Andreas Schwab <schwab@suse.de>
19058 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19059 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19062 2013-05-15 Teresa Johnson <tejohnson@google.com>
19064 * loop-unroll.c (report_unroll_peel): Check decision before
19065 emitting unroll/peel message.
19067 2013-05-15 Teresa Johnson <tejohnson@google.com>
19069 * function.h (has_bb_partition): New rtl_data flag.
19070 (bb_reorder_complete): Ditto.
19071 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19072 instead of flag_reorder_blocks_and_partition.
19073 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19074 with some enhancements.
19075 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19076 * bb-reorder.c (connect_traces): Check for has_bb_partition
19077 instead of flag_reorder_blocks_and_partition.
19078 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19079 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19080 verify_hot_cold_block_grouping.
19081 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19083 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19086 * config/arm/predicates.md (call_insn_operand): New predicate.
19087 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19088 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19089 if insn is not a tail call.
19090 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19092 * config/arm/arm.h (enum reg_class): New caller save register class.
19093 (REG_CLASS_NAMES): Likewise.
19094 (REG_CLASS_CONTENTS): Likewise.
19095 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19098 2013-05-15 Richard Biener <rguenther@suse.de>
19100 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19101 of MSG_OPTIMIZED_LOCATIONS.
19102 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19103 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19105 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19106 of MSG_OPTIMIZED_LOCATIONS.
19107 (execute_vect_slp): Likewise.
19108 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19109 (vect_create_cond_for_alias_checks): Likewise.
19110 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19111 (vect_recog_widen_mult_pattern): Likewise.
19112 (vect_recog_widen_sum_pattern): Likewise.
19113 (vect_recog_over_widening_pattern): Likewise.
19114 (vect_recog_widen_shift_pattern): Likewise.
19115 (vect_recog_vector_vector_shift_pattern): Likewise.
19116 (vect_recog_divmod_pattern): Likewise.
19117 (vect_recog_mixed_size_cond_pattern): Likewise.
19118 (vect_recog_bool_pattern): Likewise.
19119 (vect_pattern_recog_1): Likewise.
19121 2013-05-15 Martin Jambor <mjambor@suse.cz>
19123 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19124 non-functions to builtin_unreachable.
19125 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19126 correct when new direct edges were discovered.
19128 2013-05-15 Martin Jambor <mjambor@suse.cz>
19130 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19131 header, print symbol order instead of node uid, print more information
19132 about indirect edge targets.
19133 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19135 (ipa_make_edge_direct_to_target): Likewise.
19136 (remove_described_reference): Likewise.
19137 (propagate_controlled_uses): Likewise.
19138 (ipa_print_node_params): Also print symbol order.
19139 (ipcp_transform_function): Print symbol order instead of node uids.
19140 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19141 (cgraph_get_create_real_symbol_node): Likewise.
19142 * ipa-cp.c (print_lattice): Likewise.
19143 (print_all_lattices): Likewise.
19144 (determine_versionability): Likewise.
19145 (initialize_node_lattices): Likewise.
19146 (estimate_local_effects): Likewise.
19147 (update_profiling_info): Likewise.
19148 (create_specialized_node): Likewise.
19149 (perhaps_add_new_callers): Likewise.
19150 (decide_about_value): Likewise.
19151 (decide_whether_version_node): Likewise.
19152 (identify_dead_nodes): Likewise.
19153 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19154 (dump_inline_summary): Likewise.
19155 (estimate_node_size_and_time): Likewise.
19156 (inline_analyze_function): Likewise.
19157 * ipa-inline.c (report_inline_failed_reason): Likewise.
19158 (want_early_inline_function_p): Likewise.
19159 (edge_badness): Likewise.
19160 (update_edge_key): Likewise.
19161 (inline_small_functions): Likewise. Add dumping of order to two other
19163 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19164 instead of node uids.
19165 (propagate_pure_const): Likewise.
19166 (propagate_pure_const): Likewise.
19167 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19168 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19169 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19171 * tree-pretty-print.c (dump_function_header): Likewise.
19172 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19173 Print symbol order instead of node uids.
19175 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19177 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19178 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19180 2013-05-15 Richard Biener <rguenther@suse.de>
19182 PR tree-optimization/57275
19183 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19184 return value for fail to do runtime alias checks for gather loads.
19186 2013-05-15 Jan Hubicka <jh@suse.cz>
19190 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19191 weakrefs are not external.
19192 (lto_symtab_merge_decls): Fix thinko when dealing with
19193 non-lto_symtab decls.
19194 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19195 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19196 * varpool.c (dump_varpool_node): Dump more flags.
19198 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19200 * config/i386/i386.c (processor_alias_table): Add instruction
19201 FSGSBASE for AMD bdver3 architecture.
19203 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19205 * tree.c (warn_deprecated_use): Print file:line using locus color.
19206 * diagnostic.c (diagnostic_report_current_module): Print file:line
19207 and file:line:column using locus color.
19209 2013-05-14 Mike Stump <mikestump@comcast.net>
19211 * gdbinit.in: Add __null.
19213 2013-05-14 Mike Stump <mikestump@comcast.net>
19215 * recog.h: Rename struct recog_data to Recog_data.
19216 * recog.c: Likewise.
19217 * reload.c (can_reload_into): Likewise.
19218 * config/picochip/picochip.c: Likewise.
19220 2013-05-14 Mike Stump <mikestump@comcast.net>
19222 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
19224 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
19226 * resource.h (struct resources): Remove unch_memory member.
19227 (CLEAR_RESOURCE): Don't clear unch_memory.
19228 * resource.c (mark_referenced_resources): Don't set it.
19229 (mark_set_resources): Likewise.
19230 (mark_target_live_regs): Don't clear it.
19231 (init_resource_info): Likewise.
19232 * reorg.c (resource_conflicts_p): Don't compare it.
19233 (redundant_insn): Don't set it.
19235 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
19237 * emit-rtl.c (next_label): Remove unused function.
19238 (skip_consecutive_labels, link_cc0_insns): Move to ...
19239 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
19240 only place where these functions are used, and make them static.
19242 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19244 * fold-const.c (fold_negate_expr): Handle vectors.
19245 (fold_truth_not_expr): Make it static.
19246 (fold_invert_truthvalue): New static function.
19247 (invert_truthvalue_loc): Handle vectors. Do not call
19248 fold_truth_not_expr directly.
19249 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
19250 <TRUTH_NOT_EXPR>: Do not cast to boolean.
19251 (fold_comparison): Handle vector constants.
19252 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
19253 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
19254 * tree.h (fold_truth_not_expr): Remove declaration.
19256 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
19258 * config/aarch64/aarch64-simd.md
19259 (aarch64_vcond_internal<mode>): Rename to...
19260 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
19261 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
19262 float modes. Clarify all iterator modes.
19263 (vcond<mode><mode>): Use new name for vcond expanders.
19264 (vcond<v_cmp_result><mode>): Likewise.
19265 (vcondu<mode><mode>: Likewise.
19266 * config/aarch64/iterators.md (VDQF_COND): New.
19268 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19271 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
19272 variable for the shift amount. Check that we shift by non-negative
19275 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
19278 * config/arm/arm.h (EPILOGUE_USES): Only return true
19279 for LR_REGNUM after epilogue_completed.
19281 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
19283 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
19284 is error_mark_node.
19286 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19289 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
19290 and Solaris 11+/x86 with gld.
19291 * configure: Regenerate.
19293 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19295 * expmed.c (expand_shift_1): Canonicalize rotates by
19296 constant bitsize / 2 to bitsize - 1.
19297 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
19298 case ROTATERT>: Likewise.
19301 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19303 * config/i386/i386.md (rotateinv): New code attr.
19304 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19305 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19306 roll $31, %eax, etc.
19308 2013-05-14 Richard Biener <rguenther@suse.de>
19310 PR middle-end/57235
19311 * tree-eh.c (sink_clobbers): Give up for successors with
19312 multiple predecessors and no virtual uses.
19314 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
19316 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
19317 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
19319 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19321 PR middle-end/57251
19322 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
19323 the case when both op0 and op1 have VOIDmode.
19325 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19327 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
19328 in multiply-accumulate mode.
19330 2013-05-13 Guozhi Wei <carrot@google.com>
19332 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
19334 2013-05-13 Kai Tietz <ktietz@redhat.com>
19337 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
19338 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
19339 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
19340 * config/i386/i386.c (ix86_option_override_internal): Likewise.
19341 (ix86_expand_prologue): Likewise.
19342 (ix86_expand_split_stack_prologue): Likewise.
19343 (legitimate_pic_address_disp_p): Likewise.
19344 (legitimize_pic_address): Likewise.
19345 (legitimize_tls_address): Likewise.
19346 (legitimize_pe_coff_symbol): Likewise.
19347 (output_pic_addr_const): Likewise.
19348 (construct_plt_address): Likewise.
19349 (ix86_expand_call): Likewise.
19350 (x86_output_mi_thunk): Likewise.
19351 (x86_function_profiler): Likewise.
19353 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
19355 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
19356 similar switch cases.
19357 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
19358 (aarch64_simd_mov_to_<mode>low): Delete.
19359 (aarch64_simd_mov_to_<mode>high): Delete.
19360 (move_lo_quad_<mode>): Add w<-r alternative.
19361 (aarch64_simd_move_hi_quad_<mode>): Likewise.
19362 (aarch64_simd_mov_from_*): Update type attribute.
19363 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
19366 2013-05-13 Jan Hubicka <jh@suse.cz>
19368 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
19369 * config/i386/i386.c (ix86_compute_frame_layout,
19370 ix86_expand_epilogue, emit_i387_cw_initialization,
19371 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
19372 ix86_local_alignment): Fix use of size/speed predicates.
19374 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19376 PR tree-optimization/45216
19377 PR tree-optimization/57157
19378 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
19379 the (-Y) & (B - 1) variant if OP is |.
19380 * expmed.c (expand_shift_1): For rotations by const0_rtx just
19381 return shifted. Use (-op1) & (prec - 1) as other_amount
19382 instead of prec - op1.
19384 2013-05-13 Martin Jambor <mjambor@suse.cz>
19386 PR middle-end/42371
19387 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
19388 (ipa_constant_data): New type.
19389 (ipa_jump_func): Use ipa_constant_data to hold information about
19390 constant jump functions.
19391 (ipa_get_jf_constant): Adjust to jump function type changes.
19392 (ipa_get_jf_constant_rdesc): New function.
19393 (ipa_param_descriptor): New field controlled_uses.
19394 (ipa_get_controlled_uses): New function.
19395 (ipa_set_controlled_uses): Likewise.
19396 * ipa-ref.h (ipa_find_reference): Declare.
19397 * ipa-prop.c (ipa_cst_ref_desc): New type.
19398 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
19400 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
19401 New parameter cs. Adjust all callers.
19402 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
19403 (remove_described_reference): New function.
19404 (jfunc_rdesc_usable): Likewise.
19405 (try_make_edge_direct_simple_call): Decrement controlled use count,
19406 attempt to remove reference if it hits zero.
19407 (combine_controlled_uses_counters): New function.
19408 (propagate_controlled_uses): Likewise.
19409 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
19410 (ipa_edge_duplication_hook): Duplicate reference descriptions.
19411 (ipa_print_node_params): Print described use counter.
19412 (ipa_write_jump_function): Adjust to jump function type changes.
19413 (ipa_read_jump_function): New parameter CS, pass it to
19414 ipa_set_jf_constant. Adjust caller.
19415 (ipa_write_node_info): Stream controlled use count
19416 (ipa_read_node_info): Likewise.
19417 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
19419 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
19420 count. Remove cloning-added reference if it reaches zero.
19421 * ipa-ref.c (ipa_find_reference): New function.
19423 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19425 * config/i386/i386.c (processor_target_table): Modified default
19426 alignment values for AMD BD and BT architectures.
19428 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
19430 * tree-vect-generic.c (uniform_vector_p): Move ...
19431 * tree.c (uniform_vector_p): ... here.
19432 * tree.h (uniform_vector_p): Declare it.
19433 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
19436 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19438 PR tree-optimization/57230
19439 * tree-ssa-strlen.c (handle_char_store): Record length for
19440 array store from STRING_CST.
19442 PR tree-optimization/57230
19443 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
19446 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
19448 * config/epiphany/epiphany.c (epiphany_init): Check size of
19449 NUM_MODES_FOR_MODE_SWITCHING.
19450 (epiphany_expand_prologue):
19451 Remove CONFIG_REGNUM initial value handling code.
19452 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
19453 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
19454 (emit_set_fp_mode, epiphany_mode_after): Likewise.
19455 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
19456 Don't return 1 for FP_MODE_NONE.
19457 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
19458 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
19459 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
19460 * config/epiphany/epiphany.md (save_config): New pattern.
19462 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
19464 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
19466 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
19468 * config/i386/i386.md (memory): Handle sseishft1.
19469 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
19470 (*vec_extractv2di_1): Ditto.
19472 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
19474 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
19477 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
19479 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
19480 Add mthumb/march=armv7-a multilib.
19481 Add mthumb/march=armv7-r multilib.
19482 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
19484 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
19486 * config/v850/t-rtems: Add more multilibs.
19488 2013-05-10 Richard Biener <rguenther@suse.de>
19490 PR tree-optimization/57214
19491 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
19492 not propagate from SSA names that occur in abnormal PHI nodes.
19494 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
19496 * stor-layout.c (element_precision): New function.
19497 * machmode.h (element_precision): Declare it.
19498 * tree.c (build_minus_one_cst): New function.
19499 (element_precision): Likewise.
19500 * tree.h (build_minus_one_cst): Declare new function.
19501 (element_precision): Likewise.
19502 * fold-const.c (operand_equal_p): Use element_precision.
19503 (fold_binary_loc): Handle vector types.
19504 * convert.c (convert_to_integer): Use element_precision.
19505 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
19508 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
19510 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
19511 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
19512 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
19513 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
19514 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
19515 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
19516 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
19517 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
19518 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
19519 (Uuw8): New constraints.
19520 (Usb4): Move into alphabetical order.
19521 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
19522 (sd8_operand, ub8_operand, uw8_operand): New predicates.
19523 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
19524 previously unnamed patterns.
19525 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
19526 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
19527 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
19528 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
19529 of set_attr_alternative/if_then_else. Use extended_mips16 instead
19530 of specific lengths.
19532 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19534 * config/i386/i386.md (rotateinv): New code attr.
19535 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19536 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19537 roll $31, %eax, etc.
19539 PR tree-optimization/45216
19540 PR tree-optimization/57157
19541 * tree-ssa-forwprop.c (simplify_rotate): New function.
19542 (ssa_forward_propagate_and_combine): Call it.
19544 2013-05-10 Richard Biener <rguenther@suse.de>
19546 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
19547 disable peeling when we version for aliasing.
19548 (vector_alignment_reachable_p): Honor explicit user alignment.
19549 (vect_supportable_dr_alignment): Likewise.
19550 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
19551 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
19552 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
19553 then peeling to arrange for the cost-model check to come first.
19555 2013-05-10 Alan Modra <amodra@gmail.com>
19557 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
19558 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
19559 * configure: Regenerate.
19561 2013-05-10 Alan Modra <amodra@gmail.com>
19564 * varasm.c (default_elf_select_section): Move !DECL_P check..
19565 (get_named_section): ..to here before calling get_section_name.
19567 (default_section_type_flags): Add DECL_P check.
19568 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
19569 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
19571 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
19573 * config/epiphany/epiphany.c (epiphany_expand_prologue):
19574 When using gen_stack_adjust_str with a register offset, add a
19575 REG_FRAME_RELATED_EXPR note.
19577 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
19579 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
19580 (*vec_extractv4si_zext_mem): Ditto.
19581 (*vec_extractv2di): Add 0->x and x->x alternatives.
19582 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
19583 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
19585 2013-05-09 Jason Merrill <jason@redhat.com>
19587 N3639 C++1y VLA support
19588 * gimplify.c (gimplify_vla_decl): Don't touch an existing
19591 * tree.c (build_constructor_va): New.
19592 * tree.h: Declare it.
19594 2013-05-09 Martin Jambor <mjambor@suse.cz>
19597 * gimple-fold.c (canonicalize_constructor_val): Call
19598 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
19600 2013-05-09 Jan Hubicka <jh@suse.cz>
19601 Richard Biener <rguenther@suse.de>
19604 * symtab.c (symtab_make_decl_local): Do not add private names.
19606 2013-05-09 Jan Hubicka <jh@suse.cz>
19609 * symtab.c (insert_to_assembler_name_hash): Handle clones.
19610 (unlink_from_assembler_name_hash): Likewise.
19611 (symtab_prevail_in_asm_name_hash, symtab_register_node,
19612 symtab_unregister_node, symtab_initialize_asm_name_hash,
19613 change_decl_assembler_name): Update.
19615 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
19617 * config/aarch64/aarch64.md: New movtf split.
19618 (*movtf_aarch64): Update.
19619 (aarch64_movdi_tilow): Handle TF modes and rename to
19620 aarch64_movdi_<mode>low.
19621 (aarch64_movdi_tihigh): Handle TF modes and rename to
19622 aarch64_movdi_<mode>high
19623 (aarch64_movtihigh_di): Handle TF modes and rename to
19624 aarch64_mov<mode>high_di
19625 (aarch64_movtilow_di): Handle TF modes and rename to
19626 aarch64_mov<mode>low_di
19627 (aarch64_movtilow_tilow): Remove spurious whitespace.
19628 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
19630 (aarch64_print_operand): Update.
19632 2013-05-09 Alan Modra <amodra@gmail.com>
19634 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
19636 * configure: Regenerate.
19638 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
19640 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
19641 splitter preparation statements.
19642 * config/i386/sse.md (*vec_extract* splitters): Ditto.
19643 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
19646 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19648 * gimple-ssa-strength-reduction.c (count_candidates): Change
19649 return value to int.
19650 (analyze_candidates_and_replace): Change type of length to int.
19652 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
19654 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
19655 (*vec_extract<mode>): Use VI12_128 mode iterator.
19656 (*vec_extract<mode>_mem): Ditto.
19657 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
19660 2013-05-08 Diego Novillo <dnovillo@google.com>
19665 2012-08-17 Diego Novillo <dnovillo@google.com>
19668 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
19669 * config.in: Regenerate.
19670 * configure: Regenerate.
19671 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
19673 2013-05-08 Jan Hubicka <jh@suse.cz>
19676 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
19677 * cgraph.h (symtab_node_base): Add unique_name.
19678 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
19679 input_overwrite_node, input_varpool_node): Stream unique_name.
19680 * cgraphclones.c (cgraph_create_virtual_clone,
19681 cgraph_function_versioning): Set unique_name.
19682 * ipa.c (function_and_variable_visibility): Set unique_name.
19684 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19686 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
19687 (alloc_cand_and_find_basis): Restrict conditional candidate
19688 processing to CAND_MULTs.
19690 2013-05-08 Jan Hubicka <jh@suse.cz>
19693 lto-symtab.c (lto_symtab_symbol_p): New function.
19694 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
19695 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
19696 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
19697 Skip static symbols.
19699 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
19701 PR tree-optimization/57200
19702 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
19703 Only call inform if the preceding warning_at returns true.
19705 2013-05-07 Han Shen <shenhan@google.com>
19707 * cfgexpand.c (record_or_union_type_has_array_p): New function.
19708 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
19709 * common.opt (fstack-protector-strong): New option.
19710 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
19711 * doc/invoke.texi (Optimization Options): Document
19712 "-fstack-protector-strong".
19713 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
19715 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
19717 * config/mips/mips.c (mips_machine_reorg2): Return 0.
19719 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
19721 * ira.c (update_equiv_regs): Add insn having equiv memory even if
19722 it is not lhs of the insn.
19723 (setup_reg_equiv): Remove insn having equiv memory which it is not
19725 * lra-constraints.c (process_address): Try to improve generation
19726 code for address base + disp.
19727 (lra_constraints): Make correct the code for checking insn setting
19728 up backward equivalence. Remove insn only if it is in the init
19730 * lra-eliminations.c (update_reg_eliminate): Change return value.
19731 (lra_eliminate): Use the result.
19733 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
19735 * config/i386/sse.md (ssescalarnummask): New mode attribute.
19736 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
19737 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
19738 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
19739 register target operands.
19740 (*vec_extractv8hi_sse2): New pattern.
19741 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
19742 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
19743 (*vec_extract<mode>_mem): New insn and split pattern.
19745 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
19747 * config/arm/arm.c (arm_asan_shadow_offset): New function.
19748 (TARGET_ASAN_SHADOW_OFFSET): Define.
19749 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
19750 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
19752 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19754 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
19755 (incr_vec_index): Return -1 if increment not found.
19756 (create_add_on_incoming_edge): Assert if increment not found.
19757 (record_increment): Limit number of increments recorded.
19758 (all_phi_incrs_profitable): Return false if an increment not found.
19759 (replace_profitable_candidates): Don't process increments that were
19761 (analyze_candidates_and_replace): Limit size of incr_vec.
19763 2013-05-07 Richard Biener <rguenther@suse.de>
19765 * calls.c (special_function_p): setjmp-like functions are leaf.
19766 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
19767 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
19769 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
19771 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
19772 (aarch64_simd_mov<mode>): New expander.
19773 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
19774 (aarch64_simd_mov_to_<mode>high): Likewise.
19775 (aarch64_simd_mov_from_<mode>low): Likewise.
19776 (aarch64_simd_mov_from_<mode>high): Likewise.
19777 (aarch64_dup_lane<mode>): Update.
19778 (aarch64_dup_lanedi): New instruction pattern.
19779 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
19780 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
19782 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19784 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
19785 (replace_mult_candidate): Remove unnecessary argument; remove
19786 unnecessary parameter from call to introduce_cast_before_cand.
19787 (replace_unconditional_candidate): Remove unnecessary parameter
19788 from call to replace_mult_candidate.
19789 (replace_conditional_candidate): Likewise.
19790 (insert_initializers): Use make_temp_ssa_name.
19791 (introduce_cast_before_cand): Remove unnecessary argument; use
19792 make_temp_ssa_name.
19793 (replace_one_candidate): Remove unnecessary argument; remove
19794 unnecessary parameter from calls to introduce_cast_before_cand.
19795 (replace_profitable_candidates): Remove unnecessary parameters
19796 from calls to replace_one_candidate.
19798 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19800 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
19801 phi def as possibly hiding a basis for a CAND_ADD whose operands
19802 have been commuted in the analysis.
19803 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
19805 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19807 * config/aarch64/aarch64.md
19808 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
19809 shift value between 0-4.
19811 2013-05-07 Richard Biener <rguenther@suse.de>
19813 * double-int.h (rshift): New overload.
19814 * double-int.c (rshift): New function.
19815 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
19816 (create_reference_ops_from_ref): Remove.
19817 (vn_reference_insert): Use shared ops for constructing the
19818 reference and copy it.
19820 2013-05-07 Richard Biener <rguenther@suse.de>
19822 PR middle-end/57190
19823 * tree-eh.c (sink_clobbers): Properly propagate
19824 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
19826 2013-05-07 Jakub Jelinek <jakub@redhat.com>
19828 PR tree-optimization/57149
19829 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
19830 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
19831 collect_phi_def_edges, execute_late_warn_uninitialized): Use
19832 uninit_undefined_value_p instead of ssa_undefined_value_p.
19835 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
19836 for modifier == EXPAND_INITIALIZER.
19838 2013-05-07 Anton Blanchard <anton@samba.org>
19840 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
19841 for powerpc64 little endian.
19842 * configure: Regenerate.
19844 2013-05-06 Graham Stott <grahams@btinternet.com>
19846 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
19847 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
19848 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
19849 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
19851 2013-05-06 Graham Stott <grahams@btinternet.com>
19853 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
19854 codes which allow non-lvalues.
19856 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
19858 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
19859 components are all 1s.
19860 (integer_minus_onep): New function.
19861 * tree.h (integer_minus_onep): Declare it.
19862 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
19863 integer_minus_onep instead of integer_all_onesp.
19865 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
19868 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
19869 variations of these patterns.
19871 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
19873 * config/i386/i386.md (isa): Add x64_sse4 member.
19874 (enabled): Handle x64_sse4.
19875 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
19876 instruction for 64bit SSE4_1 targets. Update insn attributes.
19877 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
19878 instruction for SSE4_1 targets. Update insn attributes.
19879 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
19880 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
19881 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
19883 (*vec_extractv4si): Rename from *sse4_1_pextrd.
19884 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
19885 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
19887 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
19890 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
19892 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
19894 * final.c (do_assembler_dialects): Don't handle curly braces and
19895 vertical bar escaped by % as dialect delimiters.
19896 (output_asm_insn): Print curly braces and vertical bar if escaped
19897 by % and ASSEMBLER_DIALECT defined.
19898 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
19899 * doc/tm.texi: Regenerated.
19901 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
19903 * config/mips/mips.c: Include tree-pass.h.
19904 (mips_reorg): Split in pre- and post-dbr_schedule parts.
19905 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
19906 (pass_mips_machine_reorg2): New machine specific pass.
19907 (insert_pass_mips_machine_reorg2): New pass plugin definition.
19908 (mips_option_override): Register the new pass.
19909 * rtl.h (cleanup_barriers): Remove prototype.
19910 (dbr_schedule): Likewise.
19911 * jump.c (cleanup_barriers): Make static.
19912 * reorg.c (dbr_schedule): Likewise.
19914 2013-05-06 Richard Biener <rguenther@suse.de>
19916 PR tree-optimization/57185
19917 * tree-parloops.c (add_field_for_reduction): Handle anonymous
19918 SSA names properly.
19920 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
19923 * config/i386/i386.c (add_parameter_dependencies): Add dependence
19924 between "first_arg" and "insn", not "last" and "insn".
19926 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19928 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
19929 (find_candidates_in_block): Re-enable slsr_process_phi.
19930 (create_phi_basis): Fix double counting of candidate adjustment.
19932 2013-05-06 Richard Biener <rguenther@suse.de>
19934 PR middle-end/57147
19935 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
19936 the edge is also fallthru, preserve it and just clear the
19938 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
19939 also complex, preserve that and just clear the fallthru flag.
19940 * tree-inline.c (update_ssa_across_abnormal_edges): Also
19941 update virtual operands.
19943 2013-05-06 Alan Modra <amodra@gmail.com>
19945 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
19946 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
19947 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
19948 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
19949 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
19950 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
19952 2013-05-06 Alan Modra <amodra@gmail.com>
19954 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
19955 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
19956 (DEFAULT_ASM_ENDIAN): Define.
19957 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
19958 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
19959 Update -K PIC clause from sysv4.h.
19960 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
19961 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
19963 2013-05-06 Alan Modra <amodra@gmail.com>
19965 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
19966 twice for little-endian.
19967 (ashrdi3_no_power, ashrdi3): Support little-endian.
19969 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
19972 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
19973 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
19975 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
19976 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
19977 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
19980 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
19981 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19983 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
19984 * config/pa/pa.opt: Make mbig-switch a no-op.
19985 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
19986 (CASE_VECTOR_MODE): Always return SImode.
19987 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
19988 for the !TARGET_BIG_SWITCH case.
19989 * config/pa/pa-linux.h: Likewise.
19990 * config/pa/pa-openbsd.h: Likewise.
19991 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
19992 * config/pa/pa.md (short_jump): Remove define_insn.
19993 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
19994 (casesi0): Remove define_insn.
19995 (type): Remove btable_branch.
19996 (pa_combine_type): Likewise.
19997 (in_nullified_branch_delay): Likewise.
19998 (in_call_delay): Likewise.
19999 (define_delay): Likewise.
20000 (define_insn_reservation "Z3"): Likewise.
20001 (define_insn_reservation "Z4"): Likewise.
20002 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20003 (pa_adjust_insn_length): Remove adjustment for btable branches.
20004 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20007 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20009 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20010 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20011 Add m->r,x alternatives.
20012 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20013 splitters using SWI48x mode iterator.
20014 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20015 TARGET_64BIT. Add m->x alternative.
20016 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20017 Add o->x alternative. Enable for TARGET_SSE.
20018 (sse_storeq): Remove expander.
20019 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20020 with memory input operand.
20021 (*vec_extractv2di_1 splitter): New.
20022 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20023 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20025 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20027 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20028 (INT_LOWPART): Delete.
20029 (extract_MB): Adjust.
20030 (extract_ME): Adjust.
20031 (print_operand): Adjust.
20033 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20035 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20036 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20037 (reg_or_logical_cint_operand, easy_fp_constant,
20038 logical_const_operand): Delete "CONST_DOUBLE" case.
20039 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20040 "HOST_BITS_PER_WIDE_INT == 64" test.
20041 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20042 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20043 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20044 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20046 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20047 CONST_DOUBLE DImode/VOIDmode case.
20048 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20049 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20050 CONST_DOUBLE VOIDmode case.
20051 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20052 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20053 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20054 Delete CONST_DOUBLE case.
20055 (splitters for mov FMOVE64 const_double): Delete
20056 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20057 "HOST_BITS_PER_WIDE_INT >= 64" test.
20058 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20060 (mov DI const_double): Delete.
20062 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20064 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20065 on op shows all bits zero in mode of a lowpart subreg, return zero.
20067 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20070 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20071 to save TFmode registers and DImode to save TImode registers for
20072 caller save operations.
20073 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20074 mark being partially clobbered since they only use the first
20077 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20078 and TDmode only use the upper 64-bits of each VSX register.
20080 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20082 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20083 (find_candidates_in_block): Disable slsr_process_phi.
20085 2013-05-03 Guozhi Wei <carrot@google.com>
20087 * coverage.c (coverage_obj_init): Move the construction of gcov
20089 (build_init_ctor): ... here.
20091 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20093 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20094 (slsr_cand_d): Redefine def_phi.
20095 (stride_status, phi_adjust_status, count_phis_status): New enums.
20096 (find_phi_def): New.
20097 (find_basis_for_base_expr): New.
20098 (find_basis_for_candidate): Handle hidden bases.
20099 (alloc_cand_and_find_basis): Handle phi candidates.
20100 (slsr_process_phi): New.
20101 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20102 (create_mul_imm_cand): Likewise.
20103 (create_add_ssa_cand): Exclude phi base candidates.
20104 (create_add_imm_cand): Likewise.
20105 (slsr_process_cast): Likewise.
20106 (slsr_process_copy): Likewise.
20107 (find_candidates_in_block): Handle phi candidates.
20108 (dump_candidate): Likewise.
20109 (unconditional_cands): Delete.
20110 (unconditional_cands_with_known_stride_p): Delete.
20111 (phi_dependent_cand_p): New.
20112 (cand_increment): Handle phi-dependent candidates.
20113 (replace_dependent): Delete.
20114 (replace_mult_candidate): New.
20115 (replace_unconditional_candidate): New.
20116 (incr_vec_index): Move to avoid forward reference.
20117 (create_add_on_incoming_edge): New.
20118 (create_phi_basis): New.
20119 (replace_dependents): Delete.
20120 (replace_conditional_candidate): New.
20121 (phi_add_costs): New.
20122 (replace_uncond_cands_and_profitable_phis): New.
20123 (record_increment): Handle phi adjustments.
20124 (record_phi_increments): New.
20125 (record_increments): Handle phi adjustments.
20126 (phi_incr_cost): New.
20127 (lowest_cost_path): Handle phis.
20128 (total_savings): Likewise.
20129 (analyze_increments): Likewise.
20130 (ncd_with_phi): New.
20131 (ncd_of_cand_and_phis): New.
20132 (nearest_common_dominator_for_cands): Handle phi increments.
20133 (all_phi_incrs_profitable): New.
20134 (replace_profitable_candidates): Handle phi-dependent candidates.
20135 (analyze_candidates_and_replace): Likewise.
20137 2013-05-03 Teresa Johnson <tejohnson@google.com>
20140 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20141 do not exceed REG_BR_PROB_BASE.
20143 2013-05-03 Jeff Law <law@redhat.com>
20145 PR tree-optimization/57144
20146 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20147 operand of the condition will bit into the new type when eliminating
20148 a cast feeding a condition.
20150 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20152 PR rtl-optimization/57130
20153 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20154 of COMPARE as in_code to the recursive call if needed.
20156 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20158 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20159 (enabled): Handle new members.
20160 * config/i386/sse.md (*vec_concatv2si): Merge from
20161 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20162 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20164 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20166 PR tree-optimization/57027
20167 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20168 for fnms opportunity, check we got the prerequisite kind
20169 of tree / gimple before using accessor functions.
20171 2013-05-03 Richard Biener <rguenther@suse.de>
20173 * double-int.h (lshift): New overload without precision
20174 and arith argument.
20175 (operator *=, operator +=, operator -=): Move ...
20176 * double-int.c (operator *=, operator +=, operator -=): ... here
20177 and implement more efficiently.
20178 (mul_double_with_sign): Remove.
20179 (lshift_double): Adjust to take unsinged shift argument, push
20180 dispatching code to callers.
20181 (mul_double_wide_with_sign): Add early out for callers that
20182 are not interested in high parts or overflow.
20183 (lshift): New function.
20184 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20185 dispatch code here.
20186 (lrotate, rrotate): Use logical shifts.
20187 * expr.c (get_inner_reference): Use lshift.
20188 * fixed-value.c (do_fixed_divide): Likewise.
20189 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20190 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20191 (indirect_refs_may_alias_p): Likewise.
20192 (stmt_kills_ref_p_1): Likewise.
20194 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20196 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20198 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20200 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20201 scalar form of FABD instruction.
20203 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20205 * lra-constraints.c (process_alt_operands): Add checking alt
20206 number to choose the best alternative.
20208 2013-05-02 Richard Biener <rguenther@suse.de>
20210 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20211 bitmap and its handling.
20212 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20214 2013-05-02 Richard Biener <rguenther@suse.de>
20216 PR middle-end/57140
20217 * tree-inline.c (copy_loops): Properly handle removed loops.
20218 (copy_cfg_body): Mark destination loops for fixup if source
20219 loops needed fixup.
20221 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
20224 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
20225 generating simple_return for naked functions.
20227 2013-05-02 Martin Jambor <mjambor@suse.cz>
20229 PR middle-end/56988
20230 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
20231 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
20233 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
20234 ipa_agg_replacement_value structures.
20235 (known_aggs_to_agg_replacement_list): Likewise.
20236 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
20237 (read_agg_replacement_chain): Likewise.
20238 (ipcp_transform_function): Also check that by_ref flags match.
20240 2013-05-02 Richard Biener <rguenther@suse.de>
20242 * graphds.h (struct graph): Add obstack member.
20243 * graphds.c (new_graph): Initialize obstack and allocate
20245 (add_edge): Allocate edge from the obstack.
20246 (free_graph): Free the obstack instead of all edges and vertices.
20248 2013-05-02 Teresa Johnson <tejohnson@google.com>
20250 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
20252 * cfg.c (update_bb_profile_for_threading): Ditto.
20253 * tree-inline.c (copy_bb): Ditto.
20254 (copy_edges_for_bb): Ditto.
20255 (initialize_cfun): Ditto.
20256 (copy_cfg_body): Ditto.
20257 (expand_call_inline): Ditto.
20258 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
20259 (estimate_node_size_and_time): Ditto.
20260 (inline_merge_summary): Ditto.
20261 * cgraphclones.c (cgraph_clone_edge): Ditto.
20262 (cgraph_clone_node): Ditto.
20263 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20264 (compute_trg_info): Ditto.
20266 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20268 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
20269 S reg when fp attribute set.
20270 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
20272 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20274 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
20276 (*and_one_cmplsi3_compare0_uxtw): Likewise.
20277 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
20278 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
20280 2013-05-02 Richard Biener <rguenther@suse.de>
20282 * tree-scalar-evolution.c (scev_info_hasher): Remove.
20283 (struct instantiate_cache_entry): New type.
20284 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
20285 (struct instantiate_cache_type): New type.
20286 (set_instantiated_value, get_instantiated_value): Remove.
20287 (get_instantiated_value_entry): New function.
20288 (instantiate_scev_name): Use the new cache and adjust.
20289 (instantiate_scev_poly): Adjust.
20290 (instantiate_scev_binary): Likewise.
20291 (instantiate_array_ref): Likewise.
20292 (instantiate_scev_convert): Likewise.
20293 (instantiate_scev_not): Likewise.
20294 (instantiate_scev_3): Likewise.
20295 (instantiate_scev_2): Likewise.
20296 (instantiate_scev_r): Likewise.
20297 (instantiate_scev): Likewise.
20298 (resolve_mixers): Likewise.
20300 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
20303 * lra-constraints.c (best_small_class_operands_num): Remove.
20304 (process_alt_operands): Remove small_class_operands_num. Take
20305 small classes operands into losers and only if the operand is not
20306 matched. Modify debugging output.
20307 (curr_insn_transform): Remove best_small_class_operands_num.
20310 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20312 * config/aarch64/aarch64-builtins.c
20313 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
20314 * config/aarch64/aarch64-simd-builtins.def
20315 (reduc_splus_): Add new modes.
20316 (reduc_uplus_): New.
20317 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
20318 (reduc_uplus_v4sf): Likewise.
20319 (reduc_splus_v4sf): Likewise.
20320 (aarch64_addv<mode>): Likewise.
20321 (reduc_uplus_<mode>): Likewise.
20322 (reduc_splus_<mode>): Likewise.
20323 (aarch64_addvv2di): Likewise.
20324 (reduc_uplus_v2di): Likewise.
20325 (reduc_splus_v2di): Likewise.
20326 (aarch64_addvv2si): Likewise.
20327 (reduc_uplus_v2si): Likewise.
20328 (reduc_splus_v2si): Likewise.
20329 (reduc_<sur>plus_<mode>): New.
20330 (reduc_<sur>plus_v2di): Likewise.
20331 (reduc_<sur>plus_v2si): Likewise.
20332 (reduc_<sur>plus_v4sf): Likewise.
20333 (aarch64_addpv4sf): Likewise.
20334 * config/aarch64/arm_neon.h
20335 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
20336 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
20337 add UNSPEC_SADDV, UNSPEC_UADDV.
20339 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
20341 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20343 * config/aarch64/arm_neon.h
20344 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
20346 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20348 * config/aarch64/aarch64-builtins
20349 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
20351 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20353 * config/aarch64/aarch64-simd-builtins.def
20354 (reduc_smax_): New.
20355 (reduc_smin_): Likewise.
20356 (reduc_umax_): Likewise.
20357 (reduc_umin_): Likewise.
20358 (reduc_smax_nan_): Likewise.
20359 (reduc_smin_nan_): Likewise.
20362 (smax): Update for V2SF, V4SF and V2DF modes.
20365 (smin_nan): Likewise.
20366 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
20367 (<su><maxmin><mode>3): ...This, refactor.
20368 (s<maxmin><mode>3): New.
20369 (<maxmin_uns><mode>3): Likewise.
20370 (reduc_<maxmin_uns>_<mode>): Refactor.
20371 (reduc_<maxmin_uns>_v4sf): Likewise.
20372 (reduc_<maxmin_uns>_v2si): Likewise.
20373 (aarch64_<fmaxmin><mode>: Remove.
20374 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
20376 (vmin<q>_f<32,64>): Likewise.
20377 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20379 (su): Add mappings for smax, smin, umax, umin.
20381 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20382 (FMAXMIN): Rename as...
20383 (FMAXMIN_UNS): ...This.
20385 (fmaxminv): Likewise.
20386 (fmaxmin): Likewise.
20388 (maxmin_uns_op): Likewise.
20390 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20392 * config/aarch64/arm_neon.h
20393 (vac<ge, gt><sd>_f<32, 64>): Rename to...
20394 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
20395 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
20397 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20399 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
20400 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
20402 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20404 * config/aarch64/aarch64-simd.md
20405 (vcond<mode>_internal): Handle special cases for constant masks.
20406 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
20407 (vcondu<mode><mode>): Likewise.
20408 (vcond<v_cmp_result><mode>): New.
20410 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20412 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
20413 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
20414 * config/aarch64/aarch64-simd-builtins.def
20415 (cmeq): Update to BUILTIN_VALLDI.
20420 * config/aarch64/arm_neon.h
20421 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
20422 to builtins or C as appropriate.
20424 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20426 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
20428 (cmhi): Rename to...
20430 * config/aarch64/aarch64-simd.md
20431 (simd_mode): Add SF.
20432 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
20433 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
20434 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
20435 (cstore<mode>_neg): ...This.
20436 * config/aarch64/iterators.md
20438 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
20439 (COMPARISONS): New.
20440 (UCOMPARISONS): Likewise.
20441 (optab): Add missing comparisons.
20447 (VCMP_S): Likewise.
20448 (VCMP_U): Likewise.
20449 (V_cmp_result): Add DF, SF modes.
20450 (v_cmp_result): Likewise.
20452 (vmtype): Likewise.
20453 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
20455 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
20457 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
20458 define_insn to define_insn_and_split.
20459 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
20460 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
20461 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
20462 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
20463 (thumb2_negscc): Likewise.
20465 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
20467 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
20469 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
20471 * config/arm/thumb2.md: Remove trailing whitespaces.
20473 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
20475 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
20476 Use gen_int_mode rather than GEN_INT.
20478 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
20480 * value-prof.c (stream_in_histogram_value): Remove the strayed
20483 2013-04-30 Richard Biener <rguenther@suse.de>
20485 PR middle-end/57122
20486 * cfghooks.c (split_edge): Properly check for the loop latch edge.
20488 2013-04-30 Richard Biener <rguenther@suse.de>
20490 PR middle-end/57107
20491 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
20493 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
20495 PR rtl-optimization/56957
20496 PR rtl-optimization/57105
20497 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
20498 variable. Use just INSN_UID for determining whether an insn
20499 should be only disconnected from the insn stream.
20500 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
20502 2013-04-30 Jakub Jelinek <jakub@redhat.com>
20504 PR tree-optimization/57104
20505 * tsan.c (instrument_expr): Don't instrument accesses to
20506 DECL_HARD_REGISTER VAR_DECLs.
20508 2013-04-30 Richard Biener <rguenther@suse.de>
20510 * function.h (loops_for_fn): New inline function.
20511 (set_loops_for_fn): Likewise.
20512 * cfgloop.h (place_new_loop): Add struct function parameter.
20513 (get_loop): Likewise.
20514 (get_loops): Likewise.
20515 (number_of_loops): Likewise.
20516 (fel_next): Adjust.
20517 (fel_init): Likewise.
20518 * cfg.c (get_loop_copy): Adjust.
20519 * cfgloop.c (flow_loops_dump): Likewise.
20520 (record_loop_exits): Likewise.
20521 (verify_loop_structure): Likewise.
20522 * cfgloopanal.c (mark_irreducible_loops): Likewise.
20523 (estimate_reg_pressure_cost): Likewise.
20524 (mark_loop_exit_edges): Likewise.
20525 * cfgloopmanip.c (place_new_loop): Likewise.
20526 (add_loop): Likewise.
20527 (duplicate_loop): Likewise.
20528 * graph.c (draw_cfg_nodes): Likewise.
20529 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
20530 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
20531 (extract_affine_chrec): Likewise.
20532 (build_scop_iteration_domain): Likewise.
20533 * graphite.c (graphite_initialize): Likewise.
20534 * ira-build.c (create_loop_tree_nodes): Likewise.
20535 (more_one_region_p): Likewise.
20536 (rebuild_regno_allocno_maps): Likewise.
20537 (mark_loops_for_removal): Likewise.
20538 (mark_all_loops_for_removal): Likewise.
20539 (remove_unnecessary_regions): Likewise.
20540 (ira_build): Likewise.
20541 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
20542 * loop-init.c (fix_loop_structure): Likewise.
20543 (gate_rtl_move_loop_invariants): Likewise.
20544 (gate_rtl_unswitch): Likewise.
20545 (gate_rtl_unroll_and_peel_loops): Likewise.
20546 (rtl_doloop): Likewise.
20547 * lto-streamer-in.c (input_cfg): Likewise.
20548 * lto-streamer-out.c (output_cfg): Likewise.
20549 * modulo-sched.c (sms_schedule): Likewise.
20550 * predict.c (tree_estimate_probability): Likewise.
20551 (tree_estimate_probability_driver): Likewise.
20552 (estimate_loops): Likewise.
20553 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
20554 (move_sese_region_to_fn): Likewise.
20555 (debug_loop_num): Likewise.
20556 * tree-chrec.c (chrec_evaluate): Likewise.
20557 (hide_evolution_in_other_loops_than_loop): Likewise.
20558 (chrec_component_in_loop_num): Likewise.
20559 (reset_evolution_in_loop): Likewise.
20560 (evolution_function_is_invariant_rec_p): Likewise.
20561 * tree-if-conv.c (main_tree_if_conversion): Likewise.
20562 * tree-inline.c (copy_loops): Likewise.
20563 (copy_cfg_body): Likewise.
20564 (tree_function_versioning): Likewise.
20565 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
20566 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
20568 (add_to_evolution_1): Likewise.
20569 (scev_const_prop): Likewise.
20570 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
20571 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
20572 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
20573 (tree_ssa_lim_initialize): Likewise.
20574 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
20575 (verify_loop_closed_ssa): Likewise.
20576 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
20577 (tree_ssa_loop_im): Likewise.
20578 (tree_ssa_loop_unswitch): Likewise.
20579 (tree_vectorize): Likewise.
20580 (check_data_deps): Likewise.
20581 (tree_ssa_loop_ivcanon): Likewise.
20582 (tree_ssa_loop_bounds): Likewise.
20583 (tree_complete_unroll): Likewise.
20584 (tree_complete_unroll_inner): Likewise.
20585 (tree_parallelize_loops): Likewise.
20586 (tree_ssa_loop_prefetch): Likewise.
20587 (tree_ssa_loop_ivopts): Likewise.
20588 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
20589 * tree-vectorizer.c (vectorize_loops): Likewise.
20591 2013-04-29 Mike Frysinger <vapier@gentoo.org>
20593 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
20594 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
20595 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
20596 with EABI_LINK_SPEC.
20598 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
20601 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
20604 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
20607 * lra-constraints.c (process_alt_operands): Discourage a bit more
20608 using memory for pseudos. Print cost dump for alternatives.
20609 Modify cost values for conflicts with early clobbers.
20610 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
20612 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
20615 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
20617 2013-04-29 Ian Bolton <ian.bolton@arm.com>
20619 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
20620 from/to S register.
20621 (movdi_aarch64): Support LDR/STR from/to D register.
20623 2013-04-29 Ian Bolton <ian.bolton@arm.com>
20625 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
20626 or higher by default.
20628 2013-04-29 Richard Biener <rguenther@suse.de>
20630 PR middle-end/57075
20631 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
20632 even if not adding abnormal edges for calls that can make
20635 2013-04-29 Richard Biener <rguenther@suse.de>
20637 PR middle-end/57103
20638 * tree-cfg.c (move_stmt_op): Fix condition under which to update
20640 (move_stmt_r): Remove redundant checking.
20642 2013-04-29 Teresa Johnson <tejohnson@google.com>
20645 * basic-block.h (apply_scale): New function.
20646 (apply_probability): Use apply_scale.
20647 * gimple-streamer-in.c (input_bb): Ditto.
20648 * lto-streamer-in.c (input_cfg): Ditto.
20649 * lto-cgraph.c (merge_profile_summaries): Ditto.
20650 * tree-optimize.c (execute_fixup_cfg): Ditto.
20651 * tree-inline.c (copy_bb): Update comment to use apply_scale.
20652 (copy_edges_for_bb): Ditto.
20653 (copy_cfg_body): Ditto.
20655 2013-04-29 Tom de Vries <tom@codesourcery.com>
20657 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
20658 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
20659 (tail_merge_optimize): Handle current_loops == NULL.
20661 2013-04-26 Jeff Law <law@redhat.com>
20663 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
20664 (simplify_cond_using_ranges): Generalize code to simplify
20665 COND_EXPRs where one argument is a constant and the other
20666 is an SSA_NAME created by an integral type conversion.
20668 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20670 * config/arm/arm.md (store_minmaxsi): Use only when
20671 optimize_insn_for_size_p.
20673 2013-04-29 Christian Bruel <christian.bruel@st.com>
20676 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
20678 2013-04-29 Richard Biener <rguenther@suse.de>
20680 PR middle-end/57089
20681 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
20682 loop tree make sure to schedule a fixup for the child as well.
20683 (expand_omp_for_generic): Properly add loops.
20684 (expand_omp_for_static_nochunk): Likewise.
20685 (expand_omp_for_static_chunk): Likewise.
20686 (expand_omp_for): For the degenerate case fixup loops.
20687 (expand_omp_sections): Fix default bb placement in loops.
20688 (expand_omp_atomic_pipeline): Properly add loops.
20690 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20692 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
20694 2013-04-29 Tom de Vries <tom@codesourcery.com>
20696 * tree-ssa-tail-merge.c: Update header comment.
20698 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20700 * config/aarch64/arm_neon.h
20701 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
20702 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
20703 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
20704 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
20705 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
20706 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
20707 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
20708 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
20710 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20712 * config/aarch64/aarch64-simd.md
20713 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
20714 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
20715 fix_trunc, fixuns_trunc.
20716 (ftrunc<VDQF:mode>2): New.
20717 * config/aarch64/iterators.md (optab): Add fix, fixuns.
20718 (fix_trunc_optab): New.
20720 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20722 * config/aarch64/aarch64-builtins.c
20723 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
20724 iceilf, lround, iroundf.
20726 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
20729 * config/i386/i386.h (enum ix86_tune_indices)
20730 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
20731 New, split from X86_TUNE_INTER_UNIT_MOVES.
20732 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
20733 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
20734 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
20735 (TARGET_INTER_UNIT_MOVES): Remove.
20736 * config/i386/i386.c (initial_ix86_tune_features): Update.
20737 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
20738 (ix86_expand_convert_uns_didf_sse): Use
20739 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
20740 (ix86_expand_vector_init_one_nonzero): Ditto.
20741 (ix86_expand_vector_init_interleave): Ditto.
20742 (inline_secondary_memory_needed): Return true for moves from SSE class
20743 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
20744 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
20745 * config/i386/constraints.md (Yi, Ym): Depend on
20746 TARGET_INTER_UNIT_MOVES_TO_VEC.
20747 (Yj, Yn): New constraints.
20748 * config/i386/i386.md (*movdi_internal): Change constraints of
20749 operand 1 from Yi to Yj and from Ym to Yn.
20750 (*movsi_internal): Ditto.
20751 (*movdf_internal): Ditto.
20752 (*movsf_internal): Ditto.
20753 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
20754 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
20755 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
20756 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
20757 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
20758 * config/i386/sse.md (movdi_to_sse): Ditto.
20759 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
20760 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
20761 TARGET_INTER_UNIT_MOVES.
20762 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
20763 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
20764 instead of TARGET_INTER_UNIT_MOVES.
20765 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
20766 operand 1 from Yi to Yj and from Ym to Yn.
20768 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20770 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
20771 (float_truncate_hi_): Likewise.
20772 (float_extend_lo_): Likewise.
20773 (float_truncate_lo_): Likewise.
20774 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
20775 (aarch64_float_extend_lo_v2df): Likewise.
20776 (vec_unpacks_hi_v4sf): Likewise.
20777 (aarch64_float_truncate_lo_v2sf): Likewise.
20778 (aarch64_float_truncate_hi_v4sf): Likewise.
20779 (vec_pack_trunc_v2df): Likewise.
20780 (vec_pack_trunc_df): Likewise.
20782 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20784 * config/aarch64/aarch64-builtins.c
20785 (aarch64_fold_builtin): Fold float conversions.
20786 * config/aarch64/aarch64-simd-builtins.def
20787 (floatv2si, floatv4si, floatv2di): New.
20788 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
20789 * config/aarch64/aarch64-simd.md
20790 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
20791 * config/aarch64/iterators.md (FLOATUORS): New.
20792 (optab): Add float, floatuns.
20793 (su_optab): Likewise.
20795 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20797 * config/aarch64/aarch64-builtins.c
20798 (aarch64_builtin_vectorized_function): Use new names for
20800 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
20801 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
20802 (fcvtzu): Split as...
20803 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
20804 (fcvtas): Split as...
20805 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
20806 (fcvtau): Split as...
20807 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
20808 (fcvtps): Split as...
20809 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
20810 (fcvtpu): Split as...
20811 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
20812 (fcvtms): Split as...
20813 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
20814 (fcvtmu): Split as...
20815 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
20816 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
20817 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
20818 (lfrintnusf, lfrintnudf): Likewise.
20819 * config/aarch64/aarch64-simd.md
20820 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
20822 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
20823 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
20824 (fcvt_pattern): Likewise.
20826 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20828 * config/aarch64/aarch64-simd.md
20829 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
20830 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
20832 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20834 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
20835 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
20836 (vrnd<a,m,n,p>_f32): Implement using builtins.
20837 (vrnd<i,x><q>_f<32, 64>): New.
20839 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
20841 * config/aarch64/aarch64-builtins.c
20842 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
20843 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
20844 (frintz): Rename to...
20846 (frintp): Rename to...
20848 (frintm): Rename to...
20850 (frinti): Rename to...
20851 (nearbyint): ...this.
20852 (frintx): Rename to...
20854 (frinta): Rename to...
20856 * config/aarch64/aarch64-simd.md
20857 (aarch64_frint<frint_suffix><mode>): Delete.
20858 (<frint_pattern><mode>2): Convert to insn.
20859 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
20860 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
20861 (frint_pattern): Likewise.
20862 (frint_suffix): Likewise.
20864 2013-04-29 Richard Biener <rguenther@suse.de>
20866 PR tree-optimization/57081
20867 * loop-init.c: Include tree-flow.h.
20868 (loop_optimizer_finalize): Free number of iteration estimates.
20869 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
20871 2013-04-29 Jakub Jelinek <jakub@redhat.com>
20873 PR tree-optimization/57083
20874 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
20875 non-singleton shift count range, zero extend low_bound for uns case.
20877 * config/i386/predicates.md (general_vector_operand): New predicate.
20878 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
20879 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
20880 if they aren't nonimmediate operands. If their original values
20881 satisfy const_vector_equal_evenodd_p, don't shift them.
20882 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
20883 predicates. For the SSE4.1 case force operands[{1,2}] into registers
20884 if not nonimmediate_operand.
20885 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
20886 instead of register_operand.
20887 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
20889 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
20891 * stor-layout.c (finalize_size_functions): Allocate a structure and
20892 reset cfun before dumping the functions.
20894 2013-04-27 Jakub Jelinek <jakub@redhat.com>
20896 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
20899 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
20900 use xop_pmacsdqh if uns_p.
20901 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
20902 the immediate rotate count.
20904 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
20906 * rtl.h (struct rtx_def): Add comment for field jump.
20907 (LRA_SUBREG_P): New macro.
20908 * recog.c (register_operand): Check LRA_SUBREG_P.
20909 * lra.c (lra): Add note at the end of RTL code. Align non-empty
20911 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
20912 (lra_final_code_change): Skip subreg change for operators.
20913 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
20914 if there are no operand changes.
20915 * lra-constraints.c (curr_insn_set): New.
20916 (match_reload): Set LRA_SUBREG_P.
20917 (emit_spill_move): Ditto.
20918 (check_and_process_move): Use curr_insn_set. Process only single
20919 set insns. Don't initialize sec_mem_p and change_p.
20920 (simplify_operand_subreg): Use LRA_SUBREG_P.
20921 (reg_in_class_p): New function.
20922 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
20923 of #ifdef. Add code to remove cycling.
20924 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
20925 non-null disp. Reload inner instead of disp when base and index
20926 are null. Try to put lo_sum into register.
20927 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
20928 (check_and_process_move): Move code for move cost check to
20929 simple_move_p. Remove equiv_substitution.
20930 (simple_move_p): New function.
20931 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
20932 curr_insn_set. Call check_and_process_move only for single set
20933 insns. Use the new function. Move call of check_and_process_move
20934 after operand equiv substitution and address process.
20936 2013-04-26 Jakub Jelinek <jakub@redhat.com>
20939 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
20940 with nonlocal goto receivers or returns twice calls, ignore
20941 unininitialized values from abnormal edges to nl goto receiver
20942 or returns twice call.
20944 2013-04-26 Jakub Jelinek <jakub@redhat.com>
20946 PR tree-optimization/57051
20947 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
20948 and VEC_RSHIFT_EXPR if shift count is a multiple of element
20951 2013-04-26 Richard Biener <rguenther@suse.de>
20953 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
20954 (expand_omp_taskreg): Likewise. Mark loops for fixup.
20955 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
20956 (fixup_loop_arrays_after_move): New function.
20957 (move_sese_region_to_fn): Properly outline the loop tree parts
20958 of the SESE region.
20960 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
20962 * config/i386/i386.md (type, unit): Fix long lines.
20964 2013-04-26 Richard Biener <rguenther@suse.de>
20966 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
20967 (lto-streamer-out.o): Likewise.
20968 * cfgloop.c (init_loops_structure): Export, add struct function
20969 argument and adjust.
20970 (flow_loops_find): Adjust.
20971 * cfgloop.h (enum loop_estimation): Add EST_LAST.
20972 (init_loops_structure): Declare.
20973 * lto-streamer-in.c: Include cfgloop.h.
20974 (input_cfg): Input the loop tree.
20975 * lto-streamer-out.c: Include cfgloop.h.
20976 (output_cfg): Output the loop tree.
20977 (output_struct_function_base): Do not drop PROP_loops.
20979 2013-03-26 Richard Biener <rguenther@suse.de>
20981 * tree-cfg.c (execute_build_cfg): Build the loop tree.
20982 (pass_build_cfg): Provide PROP_loops.
20983 (move_sese_region_to_fn): Remove loops that are outlined into fn
20985 * tree-inline.c: Include cfgloop.h.
20986 (initialize_cfun): Do not drop PROP_loops.
20987 (copy_loops): New function.
20988 (copy_cfg_body): Copy loop structure.
20989 (tree_function_versioning): Initialize destination loop tree.
20990 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
20991 (pass_parallelize_loops): Do IL verification.
20992 * loop-init.c (loop_optimizer_init): Fixup loops if required.
20993 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
20994 the CFG make sure we fixup loops as well.
20995 * tree-ssa-tail-merge.c: Include cfgloop.h.
20996 (replace_block_by): When merging loop latches mark loops for fixup.
20997 * lto-streamer-out.c (output_struct_function_base): Drop
20998 PROP_loops for now.
20999 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21000 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21001 * ipa-split.c: Include cfgloop.h.
21002 (split_function): Add the new return block to the loop tree root.
21003 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21004 whether we have removed the forwarder block.
21005 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21006 * cfgloop.h (place_new_loop): Declare.
21007 * cfgloopmanip.c (place_new_loop): Export.
21008 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21009 (tree-switch-conversion.o): Likewise.
21010 (tree-complex.o): Likewise.
21011 (tree-inline.o): Likewise.
21012 (tree-ssa-tailmerge.o): Likewise.
21013 (ipa-split.o): Likewise.
21014 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21015 (tree-ssa-copy.o): Likewise.
21016 * tree-switch-conversion.c: Include cfgloop.h
21017 (process_switch): If we emit a bit-test cascade, schedule loops
21019 * tree-complex.c: Include cfgloop.h.
21020 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21021 * asan.c: Include cfgloop.h.
21022 (create_cond_insert_point): Properly add new basic-blocks to
21023 loops, schedule loop fixup.
21024 * cfgloop.c (verify_loop_structure): Check that looks are not
21026 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21028 (expand_omp_for_generic): Likewise.
21029 (expand_omp_sections): Likewise.
21030 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21031 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21032 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21033 is initialized, not when loops are present.
21034 * tree-parloops.c (parallelize_loops): Remove checking here.
21035 * passes.c (init_optimization_passes): Schedule a copy-propagation
21036 pass before complete unrolling of inner loops.
21038 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21040 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21041 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21042 (colorize_init): Add argument to _WIN32 version.
21043 * toplev.c: Include diagnostic-color.h.
21044 (process_options): Default to -fdiagnostics-color=auto if
21045 GCC_COLORS env var is in the environment.
21046 * common.opt (fdiagnostics-color=): Add Var and Init.
21047 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21048 env var is in the environment, the default is auto rather than never.
21050 * diagnostic.h (file_name_as_prefix): Add context argument.
21051 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21052 the string as locus.
21053 * langhooks.c (lhd_print_error_function): Adjust caller.
21055 2013-04-25 Lawrence Crowl <crowl@google.com>
21057 * var-tracking.c (shared_hash_def::htab):
21058 Change type to hash_table. Update dependent calls and types.
21060 2013-04-25 Lawrence Crowl <crowl@google.com>
21062 * Makefile.in: Update as needed below.
21064 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21065 Move declaration to after the type's method definitons.
21067 * attribs.c (htab_t scoped_attributes::attribute_hash):
21068 Change type to hash_table. Update dependent calls and types.
21070 * bitmap.c (htab_t bitmap_desc_hash):
21071 Change type to hash_table. Update dependent calls and types.
21073 * cselib.c (htab_t cselib_hash_table):
21074 Change type to hash_table. Update dependent calls and types.
21076 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21077 (hash_string_slot_node): Move implementation into lto-streamer.h
21078 struct string_slot_hasher.
21079 (eq_string_slot_node): Likewise.
21081 * data-streamer-out.c: Update output_block::string_hash_table
21082 dependent calls and types.
21084 * dwarf2cfi.c (htab_t trace_index):
21085 Change type to hash_table. Update dependent calls and types.
21087 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21088 Change type to hash_table. Update dependent calls and types.
21089 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21090 (htab_t optimize_external_refs::map): Likewise.
21091 (htab_t output_comp_unit::extern_map): Likewise.
21092 (htab_t output_comdat_type_unit::extern_map): Likewise.
21093 (htab_t output_macinfo::macinfo_htab): Likewise.
21094 (htab_t optimize_location_lists::htab): Likewise.
21095 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21097 * except.c (htab_t ehspec_hash_type):
21098 Change type to hash_table. Update dependent calls and types.
21099 (assign_filter_values::ttypes): Likewise.
21100 (assign_filter_values::ehspec): Likewise.
21101 (sjlj_assign_call_site_values::ar_hash): Likewise.
21102 (convert_to_eh_region_ranges::ar_hash): Likewise.
21104 * gcse.c (htab_t pre_ldst_table):
21105 Change type to hash_table. Update dependent calls and types.
21107 * ggc-common.c (htab_t saving_htab):
21108 Change type to hash_table. Update dependent calls and types.
21109 (htab_t loc_hash): Likewise.
21110 (htab_t ptr_hash): Likewise.
21111 (call_count): Rename ggc_call_count.
21112 (call_alloc): Rename ggc_call_alloc.
21113 (loc_descriptor): Rename make_loc_descriptor.
21114 (add_statistics): Rename ggc_add_statistics.
21116 * ggc-common.c (saving_htab):
21117 Change type to hash_table. Update dependent calls and types.
21119 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21120 (push_gimplify_context): Likewise.
21121 (pop_gimplify_context): Likewise.
21122 (struct gimple_temp_hash_elt): Added.
21123 (struct gimplify_hasher): Likewise.
21124 (struct gimplify_ctx.temp_htab):
21125 Change type to hash_table. Update dependent calls and types.
21127 * gimple-fold.c: Include gimplify-ctx.h.
21129 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21130 Change type to hash_table. Update dependent calls and types.
21131 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21132 avoid potential global name collision.
21134 * gimplify.c: Include gimplify-ctx.h.
21135 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21136 (htab_t gimplify_ctx::temp_htab):
21137 Update dependent calls and types for new type hash_table.
21138 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21139 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21141 * gimplify-ctx.h: New.
21142 (struct gimple_temp_hash_elt): Move from gimplify.c.
21143 (class gimplify_hasher): New.
21144 (struct gimplify_ctx): Move from gimple.h.
21145 (htab_t gimplify_ctx::temp_htab):
21146 Change type to hash_table. Update dependent calls and types.
21148 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21149 (htab_t ivs_params::newivs_index):
21150 Change type to hash_table. Update dependent calls and types.
21151 (htab_t ivs_params::params_index): Likewise.
21152 (htab_t print_generated_program::params_index): Likewise.
21153 (htab_t gloog::newivs_index): Likewise.
21154 (htab_t gloog::params_index): Likewise.
21156 * graphite.c: Include graphite-htab.h.
21157 4htab_t graphite_transform_loops::bb_pbb_mapping):
21158 Change type to hash_table. Update dependent calls and types.
21160 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21161 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21162 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21164 * graphite-dependences.c: Include graphite-htab.h.
21165 (loop_is_parallel_p): Change hash table type of parameter.
21167 * graphite-htab.h: New.
21168 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21169 (extern find_pbb_via_hash): Move from graphite-poly.h.
21170 (extern loop_is_parallel_p): Move from graphite-poly.h.
21171 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21173 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21174 (extern loop_is_parallel_p): Move to graphite-htab.h.
21175 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21177 * haifa-sched.c (htab_t delay_htab):
21178 Change type to hash_table. Update dependent calls and types.
21179 (htab_t delay_htab_i2): Likewise.
21181 * ira-color.c (htab_t allocno_hard_regs_htab):
21182 Change type to hash_table. Update dependent calls and types.
21184 * ira-costs.c (htab_t cost_classes_htab):
21185 Change type to hash_table. Update dependent calls and types.
21187 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21188 Change type to hash_table. Update dependent calls and types.
21190 * loop-iv.c (htab_t bivs):
21191 Change type to hash_table. Update dependent calls and types.
21193 * loop-unroll.c (htab_t opt_info::insns_to_split):
21194 Change type to hash_table. Update dependent calls and types.
21195 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21197 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21198 (struct string_slot_hasher): New.
21199 (htab_t output_block::string_hash_table):
21200 Change type to hash_table. Update dependent calls and types.
21202 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21203 (htab_t file_name_hash_table):
21204 Change type to hash_table. Update dependent calls and types.
21206 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21209 * lto-streamer.c (htab_t tree_htab):
21210 Change type to hash_table. Update dependent calls and types.
21212 * omp-low.c: Include gimplify-ctx.h.
21214 * passes.c (htab_t name_to_pass_map):
21215 Change type to hash_table. Update dependent calls and types.
21216 (pass_traverse): Rename to passes_pass_traverse.
21218 * plugin.c (htab_t event_tab):
21219 Change type to hash_table. Update dependent calls and types.
21221 * postreload-gcse.c (htab_t expr_table):
21222 Change type to hash_table. Update dependent calls and types.
21223 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
21225 * sese.c (debug_rename_map_1): Make extern.
21226 (htab_t copy_bb_and_scalar_dependences::rename_map):
21227 Change type to hash_table. Update dependent calls and types.
21229 * sese.h (extern debug_rename_map): Move to .c file.
21231 * store-motion.c (htab_t store_motion_mems_table):
21232 Change type to hash_table. Update dependent calls and types.
21234 * trans-mem.c (htab_t tm_new_mem_hash):
21235 Change type to hash_table. Update dependent calls and types.
21237 * tree-browser.c (htab_t TB_up_ht):
21238 Change type to hash_table. Update dependent calls and types.
21240 * tree-cfg.c (htab_t discriminator_per_locus):
21241 Change type to hash_table. Update dependent calls and types.
21243 * tree-complex.c: Include tree-hasher.h
21244 (htab_t complex_variable_components):
21245 Change type to hash_table. Update dependent calls and types.
21247 * tree-eh.c (htab_t finally_tree):
21248 Change type to hash_table. Update dependent calls and types.
21250 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
21251 struct int_tree_hasher.
21252 (extern int_tree_map_eq): Likewise.
21253 (uid_decl_map_hash): Removed.
21254 (extern decl_tree_map_eq): Likewise.
21256 * tree-hasher.h: New.
21257 (struct int_tree_hasher): New.
21258 (typedef int_tree_htab_type): New.
21260 * tree-inline.c: Include gimplify-ctx.h.
21262 * tree-mudflap.c: Include gimplify-ctx.h.
21264 * tree-parloops.c: Include tree-hasher.h.
21265 (htab_t eliminate_local_variables_stmt::decl_address):
21266 Change type to hash_table. Update dependent calls and types.
21267 (htab_t separate_decls_in_region::decl_copies): Likewise.
21269 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
21270 Change type to hash_table. Update dependent calls and types.
21272 * tree-sra.c (candidates):
21273 Change type to hash_table. Update dependent calls and types.
21275 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
21277 (int_tree_map_hash): Likewise.
21279 * tree-ssa-dom.c (htab_t avail_exprs):
21280 Change type to hash_table. Update dependent calls and types.
21282 * tree-ssa-live.c (var_map_base_init::tree_to_index):
21283 Change type to hash_table. Update dependent calls and types.
21285 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
21286 Change type to hash_table. Update dependent calls and types.
21288 * tree-ssa-phiopt.c (seen_ssa_names):
21289 Change type to hash_table. Update dependent calls and types.
21291 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
21292 Change type to hash_table. Update dependent calls and types.
21294 * tree-ssa-uncprop.c (equiv):
21295 Change type to hash_table. Update dependent calls and types.
21297 2013-04-25 Jakub Jelinek <jakub@redhat.com>
21299 PR rtl-optimization/57003
21300 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
21301 call note_stores with kill_clobbered_value callback again after
21302 killing regs_invalidated_by_call.
21304 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21306 * config/aarch64/aarch64-simd.md
21307 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
21308 (aarch64_simd_bsl<mode>): Likewise.
21309 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
21311 2013-04-25 Marek Polacek <polacek@redhat.com>
21313 PR tree-optimization/57066
21314 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
21316 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21318 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
21320 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21322 * config/aarch64/aarch64-builtins.c
21323 (aarch64_fold_builtin): New.
21324 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
21325 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
21326 * config/aarch64/aarch64-simd-builtins.def (abs): New.
21327 * config/aarch64/arm_neon.h
21328 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
21330 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21331 Tejas Belagod <tejas.belagod@arm.com>
21333 * config/aarch64/aarch64-builtins.c
21334 (aarch64_gimple_fold_builtin): New.
21335 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
21336 * config/aarch64/aarch64-simd-builtins.def (addv): New.
21337 * config/aarch64/aarch64-simd.md (addpv4sf): New.
21338 (addvv4sf): Update.
21339 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
21341 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21343 * config/aarch64/aarch64.md
21344 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
21346 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21348 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
21349 (*ngcsi_uxtw): New pattern.
21351 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21352 Julian Brown <julian@codesourcery.com>
21354 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
21355 (TB_DREG): Add T_V4HF.
21356 (v4hf_UP): New macro.
21357 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21358 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21359 Handle initialisation of V4HF. Adjust initialisation of reinterpret
21361 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21362 (arm_vector_mode_supported_p): Handle V4HF.
21363 (arm_mangle_map): Handle V4HFmode.
21364 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
21365 * config/arm/arm_neon_builtins.def: Add entries for
21366 vcvtv4hfv4sf, vcvtv4sfv4hf.
21367 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
21368 (neon_vcvtv4hfv4sf): Likewise.
21369 * config/arm/neon-gen.ml: Handle half-precision floating point
21371 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
21372 * config/arm/arm_neon.h: Regenerate.
21373 * config/arm/neon.ml (type elts): Add F16.
21374 (type vectype): Add T_float16x4, T_floatHF.
21375 (type vecmode): Add V4HF.
21376 (type features): Add Requires_FP_bit feature.
21377 (elt_width): Handle F16.
21378 (elt_class): Likewise.
21379 (elt_of_class_width): Likewise.
21380 (mode_of_elt): Refactor.
21381 (type_for_elt): Handle F16, fix error messages.
21382 (vectype_size): Handle T_float16x4.
21383 (vcvt_sh): New function.
21384 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
21385 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
21386 (string_of_mode): Handle V4HF.
21387 * doc/arm-neon-intrinsics.texi: Regenerate.
21389 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21391 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
21392 format specifier in 'X' case.
21394 2013-04-25 Alan Modra <amodra@gmail.com>
21397 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
21398 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
21399 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
21400 Repeat for many other rotate/shift and mask patterns using subregs.
21401 Name lshiftrt insns.
21402 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
21403 on WORDS_BIG_ENDIAN.
21405 2013-04-25 Alan Modra <amodra@gmail.com>
21407 * config.gcc: Support little-endian powerpc-linux targets.
21408 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
21409 (LINK_OS_LINUX_SPEC): Define.
21410 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
21411 Preserve MASK_LITTLE_ENDIAN.
21412 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
21413 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
21414 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
21415 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
21416 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
21417 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
21418 Correct fp word order for little-endian. Don't shift toc entries
21419 smaller than a word for little-endian.
21420 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
21421 (bswapdi2 splits): Correct low-part subreg for little-endian.
21422 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
21423 low/high where such is correct only for be.
21424 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
21425 little-endian for -mcall-aixdesc.
21427 2013-04-25 Alan Modra <amodra@gmail.com>
21429 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
21430 replace_equiv_address_nv.
21432 2013-04-25 Alan Modra <amodra@gmail.com>
21434 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
21436 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21439 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21440 * rtl.h (struct rtx_def): ...
21442 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21444 PR rtl-optimizations/57046
21445 * lra-constraints (split_reg): Set up lra_risky_transformations_p
21446 for multi-reg splits.
21448 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
21450 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
21452 2013-04-24 Sterling Augustine <saugustine@google.com>
21454 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
21455 (comp_dir_string, debug_str_dwo_section): New.
21456 (DEBUG_STR_DWO_SECTION): Rename to ...
21457 (DEBUG_DWO_STR_SECTION): ... this.
21458 (DEBUG_NORM_STR_SECTION): Delete.
21459 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
21460 (DEBUG_STR_DWO_SECTION_FLAGS): New.
21461 (find_AT_string): Move most logic to ...
21462 (find_AT_string_in_table): ... here. New.
21463 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
21464 add_skeleton_AT_string. Delete logic.
21465 (output_skeleton_debug_sections): Remove call to
21466 add_top_level_skeleton_die_attrs.
21467 (add_comp_dir_attribute): Move logic to comp_dir_string.
21468 (dwarf2out_init): Initialize debug_str_dwo_section.
21469 (output_indirect_string): Call find_string_form.
21470 (output_indirect_strings): Rewrite.
21471 (prune_unused_types): Empty skeleton_debug_str_hash.
21472 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
21473 (dwarf2out_finish): Call output_indirect_strings.
21475 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
21477 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
21479 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21481 * rtl.h (struct rtx_def): Add comment for field jump.
21482 (LRA_SUBREG_P): New macro.
21483 * recog.c (register_operand): Check LRA_SUBREG_P.
21484 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21486 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21487 (lra_final_code_change): Skip subreg change for operators.
21488 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21489 if there are no operand changes.
21490 * lra-constraints.c (curr_insn_set): New.
21491 (match_reload): Set LRA_SUBREG_P.
21492 (emit_spill_move): Ditto.
21493 (check_and_process_move): Use curr_insn_set. Process only single
21494 set insns. Don't initialize sec_mem_p and change_p.
21495 (simplify_operand_subreg): Use LRA_SUBREG_P.
21496 (reg_in_class_p): New function.
21497 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21498 of #ifdef. Add code to remove cycling.
21499 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21500 non-null disp. Reload inner instead of disp when base and index
21501 are null. Try to put lo_sum into register.
21502 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21503 (check_and_process_move): Move code for move cost check to
21504 simple_move_p. Remove equiv_substitution.
21505 (simple_move_p): New function.
21506 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21507 curr_insn_set. Call check_and_process_move only for single set
21508 insns. Use the new function. Move call of check_and_process_move
21509 after operand equiv substitution and address process.
21511 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
21513 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
21514 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
21515 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
21517 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
21519 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
21521 2013-04-24 Marek Polacek <polacek@redhat.com>
21523 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
21524 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
21525 (select_loops_exit_conditions): Likewise.
21526 (number_of_iterations_for_all_loops): Likewise.
21527 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
21528 (scev_analysis): Likewise.
21530 2013-04-02 Catherine Moore <clm@codesourcery.com>
21531 Chao-ying Fu <fu@mips.com>
21533 * config/mips/micromips.md (jraddiusp): New pattern.
21534 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
21535 instruction if possible.
21537 2013-04-24 Alan Modra <amodra@gmail.com>
21539 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
21541 2013-04-24 Julian Brown <julian@codesourcery.com>
21542 Chung-Lin Tang <cltang@codesourcery.com>
21544 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
21545 dependency behavior in enumeration type DIE generation. Add TODO note
21546 to comments about future DW_FORM_sdata/udata re-work of related code.
21548 2013-04-23 Lawrence Crowl <crowl@google.com>
21550 * Makefile.in: Update as needed below.
21552 * hash-table.h (class hash_table):
21553 Correct many methods with parameter types compare_type to the correct
21554 value_type. (Correct code was unlikely to notice the change.)
21555 (hash_table::elements_with_deleted) New.
21556 (class hashtable::iterator): New.
21557 (hashtable::begin()): New.
21558 (hashtable::end()): New.
21559 (FOR_EACH_HASH_TABLE_ELEMENT): New.
21561 * statistics.c (statistics_hashes):
21562 Change type to hash_table. Update dependent calls and types.
21564 * tree-into-ssa.c (var_infos):
21565 Change type to hash_table. Update dependent calls and types.
21567 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
21568 Change type to hash_table. Update dependent calls and types.
21570 * tree-ssa-loop-im.c (struct mem_ref.refs):
21571 Change type to hash_table. Update dependent calls and types.
21573 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
21574 Change type to hash_table. Update dependent calls and types.
21576 * tree-ssa-sccvn.c (vn_tables_s::nary):
21577 Change type to hash_table. Update dependent calls and types.
21578 (vn_tables_s::phis): Likewise.
21579 (vn_tables_s::references): Likewise.
21581 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
21582 (vn_reference_eq): Update parameter and return types.
21584 * tree-ssa-structalias.c (pointer_equiv_class_table):
21585 Change type to hash_table. Update dependent calls and types.
21586 (location_equiv_class_table): Likewise.
21588 * tree-vect-data-refs.c: Consequential changes for making
21589 peeling a hash_table.
21591 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
21592 (destroy_loop_vec_info): Dependent hash_table update.
21594 * tree-vectorizer.h (peeling_htab):
21595 Change type to hash_table. Update dependent calls and types.
21597 2013-04-23 Shiva Chen <shiva0217@gmail.com>
21599 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
21600 to check the register content is equal or not.
21601 * lra-constraints.c (match_reload): Use lra_assign_reg_val
21602 to assign register content record.
21603 * lra-eliminations.c (update_reg_eliminate): Use
21604 lra_update_reg_val_offset to update register content offset.
21605 * lra-int.h (struct lra_reg): Add offset member.
21606 (lra_reg_val_equal_p): New static inline function.
21607 (lra_update_reg_val_offset): New static inline function.
21608 (lra_assign_reg_val): New static inline function.
21609 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
21610 to assign register content record.
21611 (initialize_lra_reg_info_element): Initial offset to zero.
21613 2013-04-23 Catherine Moore <clm@codesourcery.com>
21615 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
21616 operands. Record compression.
21618 2013-04-23 Xinliang David Li <davidxl@google.com>
21620 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
21622 2013-04-23 Richard Biener <rguenther@suse.de>
21624 PR middle-end/57036
21625 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
21626 parameter, only add abnormal goto edges from the copied body
21627 if the call could perform abnormal gotos.
21628 (copy_cfg_body): Adjust.
21630 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
21632 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
21634 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
21636 * coretypes.h (gimple_stmt_iterator): Add struct to make
21639 2013-04-23 Richard Biener <rguenther@suse.de>
21641 PR tree-optimization/57026
21642 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
21643 from SSA names occuring in abnormal PHI nodes.
21645 2013-04-22 Andi Kleen <ak@linux.intel.com>
21647 * lto/lto.c (print_lto_report_1): Fix LTO report names.
21649 2013-04-22 Andi Kleen <ak@linux.intel.com>
21651 * lto/lto.c (print_lto_report_1): Declare early.
21652 (read_cgraph_and_symbols): Call print_lto_report_1 early.
21654 2013-04-22 Andi Kleen <ak@linux.intel.com>
21656 * common.opt (-flto-report-wpa): Add.
21657 * doc/invoke.texi (-flto-report-wpa): Add.
21658 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
21661 2013-04-22 Xinliang David Li <davidxl@google.com>
21663 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
21664 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
21665 * Makefile.in: New dependency
21667 David Daney <ddaney.cavm@gmail.com>
21669 * configure.ac (gcc_cv_as_micromips_support): Use the
21670 --fatal-warnings option.
21671 * configure: Regenerate.
21673 2013-04-22 Marek Polacek <polacek@redhat.com>
21676 * tsan.c (instrument_expr): Don't instrument expression
21677 in case its size is zero.
21679 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
21683 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
21685 * config/alpha/alpha.c (TARGET_LRA_P): New define.
21687 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
21689 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
21690 (gimple_stmt_iterator): New typedef.
21691 * gimple.h (gimple_stmt_iterator): Rename to...
21692 (gimple_stmt_iterator_d): ... This.
21693 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
21694 trees be valid for GIMPLE and GENERIC.
21695 (TARGET_GIMPLE_FOLD_BUILTIN): New.
21696 * gimple-fold.c (gimple_fold_call): Call target hook
21697 gimple_fold_builtin.
21698 * hooks.c (hook_bool_gsiptr_false): New.
21699 * hooks.h (hook_bool_gsiptr_false): New.
21700 * target.def (fold_stmt): New.
21701 * doc/tm.texi: Regenerate.
21703 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
21706 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
21707 a set sp if no stack realignment.
21709 2013-04-22 Nick Clifton <nickc@redhat.com>
21711 * config.gcc (tilegx-linux): Extend extra_objs rather than
21713 (tilepro-linux): Likewise.
21715 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
21717 * config/aarch64/aarch64-builtins.c
21719 (CF0, CF1, CF2, CF3, CF4, CF10): New.
21720 (VAR<1-12>): Add MAP parameter.
21721 (BUILTIN_*): Likewise.
21722 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
21723 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
21724 (aarch64_ushl_n<mode>): Likewise.
21725 (aarch64_sshr_n<mode>): Likewise.
21726 (aarch64_ushr_n<mode>): Likewise.
21727 (aarch64_<maxmin><mode>): Likewise.
21728 (aarch64_sqrt<mode>): Likewise.
21729 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
21730 (vshr<q>_n_*): Likewise.
21732 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
21734 * config/aarch64/aarch64-builtins.c
21735 (aarch64_simd_builtin_type_mode): Handle SF types.
21737 (BUILTIN_GPF): Define.
21738 (aarch64_init_simd_builtins): Handle SF types.
21739 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
21740 (frecps): Likewise.
21741 (frecpx): Likewise.
21742 * config/aarch64/aarch64-simd.md
21743 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
21744 (aarch64_frecpe<mode>): New.
21745 (aarch64_frecps<mode>): Likewise.
21746 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
21747 (v8type): Add frecp<esx>.
21748 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
21749 (aarch64_frecps<mode>): Likewise.
21750 * config/aarch64/iterators.md (FRECP): New.
21751 (frecp_suffix): Likewise.
21752 * config/aarch64/arm_neon.h
21753 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
21755 2013-04-22 Christian Bruel <christian.bruel@st.com>
21758 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
21759 (REG_CLASS_NAMES): Idem.
21760 (REG_CLASS_CONTENTS): Idem.
21761 (REGCLASS_HAS_FP_REG): Idem.
21762 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
21763 (sh_conditional_register_usage): Idem.
21765 2013-04-21 Jeff Law <law@redhat.com>
21767 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
21768 (ssa_forward_propagate_and_combine): Use it.
21770 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
21772 * lra.c: Update the flow chart diagram.
21774 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
21776 PR rtl-optimization/56847
21777 * lra-constraints.c (process_alt_operands): Discourage alternative
21778 with non-matche doffsettable memory constraint fro memory with
21781 2013-04-19 Richard Biener <rguenther@suse.de>
21783 PR tree-optimization/56982
21784 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
21786 * gimplify.c (gimplify_call_expr): Notice special calls.
21787 (gimplify_modify_expr): Likewise.
21788 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
21789 abnormal control flow receivers.
21790 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
21791 in the same way as cfun->has_nonlocal_labels.
21792 (gimple_purge_dead_abnormal_call_edges): Likewise.
21793 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
21794 receivers start a basic-block.
21796 2013-04-19 Richard Biener <rguenther@suse.de>
21798 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
21800 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
21801 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
21802 (SLP_TREE_LOAD_PERMUTATION): Add.
21803 (vect_transform_slp_perm_load): Adjust prototype.
21804 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
21805 (vect_free_slp_instance): Likewise.
21806 (vect_create_new_slp_node): Likewise.
21807 (vect_supported_slp_permutation_p): Remove.
21808 (vect_slp_rearrange_stmts): Adjust.
21809 (vect_supported_load_permutation_p): Likewise. Inline
21810 vect_supported_slp_permutation_p here.
21811 (vect_analyze_slp_instance): Compute load permutations per
21812 slp node instead of per instance.
21813 (vect_get_slp_defs): Adjust.
21814 (vect_transform_slp_perm_load): Likewise.
21815 (vect_schedule_slp_instance): Remove redundant code.
21816 (vect_schedule_slp): Remove hack for PR56270, add it ...
21817 * tree-vect-stmts.c (vectorizable_load): ... here, do not
21818 CSE loads for SLP. Adjust.
21820 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
21822 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
21823 spelling in two comments.
21825 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
21828 * config/arm/arm.c (load_multiple_sequence): Require SP
21829 as base register for loads if SP is in the register list.
21831 2013-04-19 Martin Jambor <mjambor@suse.cz>
21833 PR tree-optimization/56718
21834 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
21835 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
21836 and made public. Adjusted all callers.
21837 (ipa_intraprocedural_devirtualization): New function.
21838 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
21839 (ipa_intraprocedural_devirtualization): Likewise.
21840 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
21842 2013-04-19 Richard Biener <rguenther@suse.de>
21844 PR tree-optimization/57000
21845 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
21847 2013-04-19 Terry Guo <terry.guo@arm.com>
21849 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
21851 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
21852 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
21853 (cortex_m4_fmacs): Use new reservations.
21854 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
21856 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
21858 PR rtl-optimization/56999
21859 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
21861 (lra_coalesce): Remove split_origin_bitmap and related code.
21862 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
21863 ranges if necessary.
21865 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
21867 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
21869 (ix86_expand_call): Remove clobbered_registers array and use
21870 x86_64_ms_sysv_extra_clobbered_registers instead.
21871 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
21873 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
21875 * config/i386/i386.md (*call_rex64_ms_sysv): Use
21876 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
21877 (*call_value_rex64_ms_sysv): Ditto.
21879 2013-04-18 Cary Coutant <ccoutant@google.com>
21881 * dwarf2out.c (output_pubnames): Check die_perennial_p of
21882 parent instead of die_mark.
21884 2013-04-18 Diego Novillo <dnovillo@google.com>
21886 * gimple.c (create_gimple_tmp): New.
21887 (get_expr_type): New.
21888 (build_assign): New.
21889 (build_type_cast): New.
21890 * gimple.h (enum ssa_mode): Define.
21891 (gimple_seq_set_location): New.
21892 * asan.c (build_check_stmt): Change some gimple_build_* calls
21893 to use build_assign and build_type_cast.
21895 2013-04-18 Richard Biener <rguenther@suse.de>
21897 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
21898 handle negative step. Remove redundant checks.
21899 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
21900 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
21901 for negative step and grouped loads fail to vectorize.
21903 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
21905 * emit-rtl.c (reset_insn_used_flags): New function.
21906 (reset_all_used_flags): Use it.
21907 (verify_insn_sharing): New function.
21908 (verify_rtl_sharing): Fix verification for SEQUENCEs.
21910 2013-04-18 Jakub Jelinek <jakub@redhat.com>
21912 PR tree-optimization/56984
21913 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
21914 and (x >> M) >= N don't register any assertion if N << M is the
21917 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
21919 * lower-subreg.c (resolve_simple_move): If called self-recursive,
21920 do not delete_insn insns that have not yet been emitted, only
21921 unlink them with remove_insn.
21922 * df-scan.c (df_insn_delete): Revert r197492.
21924 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
21926 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
21927 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
21929 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
21931 * config/arm/arm.md (movsicc_insn): Convert define_insn into
21932 define_insn_and_split.
21933 (and_scc,ior_scc,negscc): Likewise.
21934 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
21936 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
21938 * config/arm/arm.c (use_return_insn): Return 0 for targets that
21939 can benefit from using a sequence of LDRD instructions in epilogue
21940 instead of a single LDM instruction.
21942 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
21945 * doc/extend.texi: Fix typo.
21947 2013-04-17 Richard Biener <rguenther@suse.de>
21949 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
21950 (vect_build_slp_tree): ... here.
21951 (vect_build_slp_tree_1): Compute which stmts of the SLP group
21952 match. Remove special-casing of mismatched complex loads.
21953 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
21954 re-try the match with swapped commutative operands.
21955 (vect_supported_load_permutation_p): Remove special-casing of
21956 mismatched complex loads.
21957 (vect_analyze_slp_instance): Adjust.
21959 2013-04-17 Richard Biener <rguenther@suse.de>
21961 PR rtl-optimization/56921
21962 * cfgloop.h (struct loop): Add simple_loop_desc member.
21963 (struct niter_desc): Mark with GTY(()).
21964 (simple_loop_desc): Do not use aux field but simple_loop_desc.
21965 * loop-iv.c (get_simple_loop_desc): Likewise.
21966 (free_simple_loop_desc): Likewise.
21969 2013-04-16 Richard Biener <rguenther@suse.de>
21971 PR rtl-optimization/56921
21972 * loop-init.c (pass_rtl_move_loop_invariants): Add
21973 TODO_do_not_ggc_collect to todo_flags_finish.
21974 (pass_rtl_unswitch): Same.
21975 (pass_rtl_unroll_and_peel_loops): Same.
21976 (pass_rtl_doloop): Same.
21978 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
21980 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
21981 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
21982 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
21984 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
21985 * tree-streamer.c (record_common_node): Adjust reference in comment.
21987 2013-04-17 Terry Guo <terry.guo@arm.com>
21989 * config/arm/cortex-m4.md: Add a new bypass.
21991 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21993 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
21995 (*subs_<optab><mode>_multp2): New pattern.
21996 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
21997 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
21999 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22001 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22002 (*subs_mul_imm_<mode>): New pattern.
22004 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22007 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22008 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22009 (vsx_movti_32bit): Change j->wa to O->wa.
22011 2013-04-16 Richard Biener <rguenther@suse.de>
22013 PR rtl-optimization/56921
22014 * loop-init.c (pass_rtl_move_loop_invariants): Add
22015 TODO_do_not_ggc_collect to todo_flags_finish.
22016 (pass_rtl_unswitch): Same.
22017 (pass_rtl_unroll_and_peel_loops): Same.
22018 (pass_rtl_doloop): Same.
22020 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22022 * config/arm/arm.c (emit_multi_reg_push): New declaration
22023 for an existing function.
22024 (arm_emit_strd_push): New function.
22025 (arm_expand_prologue): Used here.
22026 (arm_emit_ldrd_pop): New function.
22027 (arm_expand_epilogue): Used here.
22028 (arm_get_frame_offsets): Update condition.
22029 (arm_emit_multi_reg_pop): Add a special case for load of a single
22030 register with writeback.
22032 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22034 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22037 2013-04-16 Richard Biener <rguenther@suse.de>
22039 PR tree-optimization/56756
22040 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22041 (first_mem_ref_loc): New.
22042 (execute_sm): Place the load temporarily before a previous
22043 access instead of in the latch edge to ensure its SSA dependencies
22044 are defined at points dominating the load.
22046 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22048 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22049 correct fix by moving header and footer insn to the footer of
22050 the merged basic block. Clear BB_END of the merged-away block.
22052 PR middle-end/43631
22053 * emit-rtl.c (make_note_raw): New function.
22054 (link_insn_into_chain): New static inline function.
22055 (add_insn): Use it.
22056 (add_insn_before, add_insn_after): Factor insn chain linking code...
22057 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22058 using link_insn_into_chain.
22059 (note_outside_basic_block_p): New helper function for emit_note_after
22060 and emit_note_before.
22061 (emit_note_after): Use nobb variant of add_insn_after if the note
22062 should not be contained in a basic block.
22063 (emit_note_before): Use nobb variant of add_insn_before if the note
22064 should not be contained in a basic block.
22065 (emit_note_copy): Use make_note_raw.
22066 (emit_note): Likewise.
22067 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22068 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22069 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22070 the moved barrier the tail of the basic block it follows.
22071 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22073 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22075 PR tree-optimization/56962
22076 * gimple-ssa-strength-reduction.c (record_increment): Only set
22077 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22078 either rhs1 or rhs2 is equal to c->base_expr.
22080 2013-04-15 Richard Biener <rguenther@suse.de>
22082 PR tree-optimization/56933
22083 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22085 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22086 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22087 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22088 dependence check ...
22089 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22091 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22092 GROUP_READ_WRITE_DEPENDENCE.
22094 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22096 * emit-rtl.c (reset_all_used_flags): New function.
22097 (verify_rtl_sharing): Call reset_all_used_flags before and after
22098 performing the checks.
22100 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22102 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22103 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22104 * config/arm/constraints.md (De): New constraint.
22105 * config/arm/neon.md (anddi3_neon): Delete.
22106 (neon_vand<mode>): Expand to standard anddi3 pattern.
22107 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22108 Move earlier in the file.
22109 (neon_inv_logic_op2): Likewise.
22110 (arm_anddi_operand_neon): New predicate.
22112 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22114 * configure.ac (gcc_cv_ld_as_needed): Set
22115 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22116 Use -z ignore, -z record on *-*-solaris2*.
22117 (HAVE_LD_AS_NEEDED): Update comment.
22118 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22119 * configure: Regenerate.
22120 * config.in: Regenerate.
22121 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22122 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22123 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22124 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22125 equivalents. Fix markup.
22126 * doc/tm.texi: Regenerate.
22128 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22130 * config/i386/i386.opt: New option mstack-protector-guard=.
22131 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22132 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22133 TARGET_SSP_TLS_GUARD.
22134 * config/i386/i386.c (ix86_option_override_internal): Set
22135 ix86_stack_protector_guard.
22136 * config/i386/i386.md (stack_protect_set): Enable for
22137 TARGET_SSP_TLS_GUARD only.
22138 (stack_protect_set_<mode>): Ditto.
22139 (stack_protect_test): Ditto.
22140 (stack_protect_test_<mode>): Ditto.
22141 * doc/invoke.texi (i386 Option): Document.
22143 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22146 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22147 (S_MODES): Set H_MODE bit.
22148 (SF_MODES): Set only S_MODE and SF_MODE bits.
22149 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22150 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22151 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22152 <MODE_FLOAT>: Likewise.
22154 2013-04-15 Joey Ye <joey.ye@arm.com>
22156 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22158 2013-04-15 Joey Ye <joey.ye@arm.com>
22160 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22162 (thumb_far_jump_used_p): Count instruction size and set
22165 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22167 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22168 * resource.c (mark_target_live_regs): Reformat conditional block.
22170 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22172 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22173 notes, they are emitted only just before final.
22174 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22176 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22178 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22179 * cfgrtl.c (delete_insn): Call it here instead.
22180 * lra-spills.c (lra_final_code_change): Use delete_insn.
22181 * haifa-sched.c (sched_remove_insn): Likewise.
22182 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22183 returning to the nop pool.
22184 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22185 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22187 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22189 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22190 * doc/tm.texi: Regenerated.
22192 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22194 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22197 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22199 * df-core.c (df_find_def): Compare register numbers.
22200 (df_find_use): Likewise.
22202 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22205 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22206 lra_in_progress for return.
22208 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22210 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22211 define_insn into define_insn_and_split and emit movsicc patterns.
22213 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22215 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
22217 2013-04-12 Richard Biener <rguenther@suse.de>
22219 * tree-pass.h (TODO_do_not_ggc_collect): New.
22220 * passes.c (execute_one_ipa_transform_pass): Honor
22221 TODO_do_not_ggc_collect.
22222 (execute_one_pass): Likewise.
22225 2013-04-10 Richard Biener <rguenther@suse.de>
22227 * passes.c (init_optimization_passes): Remove reload pass.
22228 * ira.c (do_reload): Merge into ...
22230 (rest_of_handle_reload): Remove.
22231 (pass_reload): Likewise.
22232 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22233 of reload for vzeroupper pass placement.
22235 2013-04-12 Jakub Jelinek <jakub@redhat.com>
22237 PR tree-optimization/56918
22238 PR tree-optimization/56920
22239 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
22240 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
22241 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
22242 use wide_mul_with_sign method.
22244 2013-04-12 Richard Biener <rguenther@suse.de>
22246 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
22247 not be considered a gimple constant.
22249 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
22251 * fold-const.c (const_binop): Handle vector shifts by a scalar.
22252 (fold_binary_loc): Call const_binop also for mixed vector-scalar
22255 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
22256 Jakub Jelinek <jakub@redhat.com>
22258 * opts.c: Include diagnostic-color.h.
22259 (common_handle_option): Handle OPT_fdiagnostics_color_.
22260 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
22261 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
22262 (diagnostic-color.o): New.
22263 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
22264 (diagnostic_color_rule): New enum.
22265 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
22266 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
22267 the location string.
22268 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
22269 either NULL, or color kind.
22270 * diagnostic-color.c: New file.
22271 * diagnostic-color.h: New file.
22272 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
22274 * doc/invoke.texi (-fdiagnostics-color): Document.
22275 * pretty-print.h (pp_show_color): Define.
22276 (struct pretty_print_info): Add show_color field.
22277 * diagnostic.c: Include diagnostic-color.h.
22278 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
22279 macros. Colorize error:, warning: etc. strings and also the location
22281 (diagnostic_show_locus): Colorize the caret line.
22282 * pretty-print.c: Include diagnostic-color.h.
22283 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
22284 inside of %< %> quotes or quoted through q format modifier.
22286 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22288 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
22290 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22292 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
22293 code in CC_NZ mode.
22294 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
22297 2013-04-11 Marek Polacek <polacek@redhat.com>
22299 PR tree-optimization/48184
22300 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
22302 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
22304 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
22305 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
22306 (skip_simple_arithmetic): Tidy up.
22307 * tree.h (skip_simple_constant_arithmetic): Declare.
22309 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22311 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
22313 2013-04-11 Richard Biener <rguenther@suse.de>
22315 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22316 generate vector constants.
22318 2013-04-11 Richard Biener <rguenther@suse.de>
22320 PR tree-optimization/56878
22321 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
22322 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
22324 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
22325 Prefer to align the DR with the most invariant base address.
22327 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22329 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
22332 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
22334 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
22335 floating-point vector comparisons against 0.
22337 2013-04-11 Jakub Jelinek <jakub@redhat.com>
22339 PR tree-optimization/56899
22340 * fold-const.c (extract_muldiv_1): Apply distributive law
22341 only if TYPE_OVERFLOW_WRAPS (ctype).
22343 2013-04-11 Bin Cheng <bin.cheng@arm.com>
22346 * ira-costs.c (scan_one_insn): Check whether the source rtx of
22347 loading has side effect.
22349 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
22351 * config/sparc/sparc.c: Include tree-pass.h.
22352 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
22353 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
22354 head of file. Change return type. Split off gate function.
22355 (sparc_gate_work_around_errata): New function.
22356 (pass_work_around_errata): New pass definition.
22357 (insert_pass_work_around_errata) New pass insert definition to
22358 insert pass_work_around_errata just after delayed-branch scheduling.
22359 (sparc_option_override): Insert the pass.
22360 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
22362 2013-04-10 David S. Miller <davem@davemloft.net>
22364 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
22365 or -mcpu=hypersparc.
22367 * target.def (cstore_mode): New hook.
22368 * target.h: Include insn-codes.h
22369 * targhooks.c: Likewise.
22370 (default_cstore_mode): New function.
22371 * targhooks.h: Declare it.
22372 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
22373 * doc/tm.texi: Rebuild.
22374 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
22375 target hook, rather than inspecting the insn_data.
22376 * config/sparc/sparc.c (sparc_cstore_mode): New function.
22377 (TARGET_CSTORE_MODE): Redefine.
22378 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
22380 * config/sparc/predicates.md (cstore_result_operand): New special
22382 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
22383 Use it for operand 0.
22384 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
22385 (*snesi_special): Likewise.
22386 (*snesi_zero): Likewise.
22387 (*seqsi_zero): Likewise.
22388 (*sltu_insn): Likewise.
22389 (*sgeu_insn): Likewise.
22390 (*seqdi_special): Make operand 0 and comparison operation be of
22392 (*snedi_special): Likewise.
22393 (*snedi_special_vis3): Likewise.
22394 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
22395 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
22396 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
22397 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
22398 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
22399 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
22400 (*sltu_extend_sp64): Likewise.
22401 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
22402 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
22403 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
22404 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
22405 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
22407 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
22409 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
22410 (aarch64_start_file): Use the new function.
22412 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22413 Jason Merrill <jason@redhat.com>
22415 * common.opt: Add -gdwarf.
22416 * opts.c (common_handle_option): Handle it.
22417 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
22419 2013-04-10 Richard Biener <rguenther@suse.de>
22421 * passes.c (execute_todo): Do not call ggc_collect conditional here.
22422 (execute_one_ipa_transform_pass): But unconditionally here.
22423 (execute_one_pass): And here.
22424 (init_optimization_passes): Remove reload pass.
22425 * tree-pass.h (TODO_ggc_collect): Remove.
22426 (pass_reload): Likewise.
22427 * ira.c (do_reload): Merge into ...
22429 (rest_of_handle_reload): Remove.
22430 (pass_reload): Likewise.
22431 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22432 of reload for vzeroupper pass placement.
22433 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
22434 and todo_flags_finish of all passes.
22436 2013-04-10 Richard Biener <rguenther@suse.de>
22438 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
22439 first_const_oprnd field, rename first_def_type to first_op_type.
22440 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
22441 (vect_get_and_check_slp_defs): Always use the type of the
22442 operand. Allow mixed vect_external_def, vect_constant_def types.
22443 (vect_get_constant_vectors): Handle mixed vect_external_def,
22444 vect_constant_def types.
22446 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
22448 PR tree-optimization/55524
22449 * tree-ssa-math-opts.c
22450 (convert_mult_to_fma): Don't use an fms construct
22451 when we don't have an fms operation, but fnma, and it looks
22452 likely that we'll be able to use the latter.
22454 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
22456 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
22458 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
22459 inline fail caused by overwritable functions.
22461 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
22463 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
22464 unnecessary bits in the constant power of two case.
22466 2013-04-10 Richard Biener <rguenther@suse.de>
22468 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
22469 broken code swapping operands.
22470 (vect_build_slp_tree): Do not compute load permutations here.
22471 (vect_analyze_slp_instance): Compute load permutations here,
22472 after building the SLP tree.
22474 2013-04-09 Christian Bruel <christian.bruel@st.com>
22476 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
22477 of next/prev_real_insn.
22479 2013-04-09 Jan Hubicka <jh@suse.cz>
22481 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
22482 Drop aliased parameter.
22483 (function_and_variable_visibility): Do not handle alias pairs.
22484 * cgraph.c (varpool_externally_visible_p): Update prototype.
22485 * varpool.c (varpool_add_new_variable): Update.
22487 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22489 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
22491 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
22493 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
22495 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
22497 2013-04-09 Marek Polacek <polacek@redhat.com>
22499 PR tree-optimization/48762
22500 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
22502 2013-04-09 Richard Biener <rguenther@suse.de>
22504 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
22506 (vect_build_slp_tree): Likewise.
22507 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
22508 calculating the cost of a SLP instance.
22509 (vect_analyze_slp_instance): Use it from here, after building
22512 2013-04-09 Jakub Jelinek <jakub@redhat.com>
22514 PR middle-end/56883
22515 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
22516 expand_omp_for_static_chunk): Use simple_p = true in
22517 force_gimple_operand_gsi calls when assigning to addressable decls.
22519 2013-04-09 Jeff Law <law@redhat.com>
22521 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
22522 when the boolean was created by converting a wider object which
22523 had a boolean range.
22525 2013-04-09 Richard Biener <rguenther@suse.de>
22527 * tree-vectorizer.h (slp_void_p): Remove.
22528 (slp_tree): Typedef before _slp_tree declaration.
22529 (struct _slp_tree): Use a vector of slp_tree as children.
22530 (vect_get_place_in_interleaving_chain): Remove.
22531 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
22533 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
22535 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
22536 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
22537 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
22538 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
22539 Use slp_node instead of slp_void_p and adjust.
22541 2013-04-09 Richard Biener <rguenther@suse.de>
22543 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
22544 work that is not necessary.
22546 2013-04-09 Jakub Jelinek <jakub@redhat.com>
22548 PR tree-optimization/56854
22549 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
22550 forward into clobber stmts if it would change MEM_REF lhs into
22553 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
22555 * tree.c (type_hash_lookup, type_hash_add): Make static.
22556 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
22558 2013-04-09 Richard Biener <rguenther@suse.de>
22560 * tree.h (unsave_expr_now): Remove.
22561 * tree-inline.c (mark_local_for_remap_r): Remove.
22562 (unsave_expr_1): Likewise.
22563 (unsave_r): Likewise.
22564 (unsave_expr_now): Likewise.
22565 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
22566 (propagate_tree_value): Likewise.
22568 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
22570 * doc/rtl.texi (sequence): Rewrite documentation to match the
22571 current use of SEQUENCE rtl objects.
22572 * rtl.def (SEQUENCE): Likewise.
22574 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
22575 Update documentation.
22576 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
22577 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
22579 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
22581 2013-04-08 Teresa Johnson <tejohnson@google.com>
22583 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
22584 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
22586 (estimate_edge_size_and_time): Add comment to suggest using rounding
22588 (estimate_node_size_and_time): Ditto.
22589 (remap_edge_change_prob): Use helper rounding divide methods.
22590 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
22591 (gimple_mod_pow2_value_transform): Ditto.
22592 (gimple_mod_subtract_transform): Ditto.
22593 (gimple_ic_transform): Ditto.
22594 (gimple_stringops_transform): Ditto.
22595 * stmt.c (conditional_probability): Ditto.
22596 (emit_case_dispatch_table): Ditto.
22597 * lto-cgraph.c (merge_profile_summaries): Ditto.
22598 * tree-optimize.c (execute_fixup_cfg): Ditto.
22599 * cfgcleanup.c (try_forward_edges): Ditto.
22600 * cfgloopmanip.c (scale_loop_profile): Ditto.
22602 (duplicate_loop_to_header_edge): Ditto.
22603 (lv_adjust_loop_entry_edge): Ditto.
22604 * tree-vect-loop.c (vect_transform_loop): Ditto.
22605 * profile.c (compute_branch_probabilities): Ditto.
22606 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
22607 * lto-streamer-in.c (input_cfg): Ditto.
22608 * gimple-streamer-in.c (input_bb): Ditto.
22609 * ipa-cp.c (update_profiling_info): Ditto.
22610 (update_specialized_profile): Ditto.
22611 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
22612 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
22614 * sched-rgn.c (compute_dom_prob_ps): Ditto.
22615 (compute_trg_info): Ditto.
22616 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
22617 (purge_dead_edges): Ditto.
22618 * loop-unswitch.c (unswitch_loop): Ditto.
22619 * cgraphclones.c (cgraph_clone_edge): Ditto.
22620 (cgraph_clone_node): Ditto.
22621 * tree-inline.c (copy_bb): Ditto.
22622 (copy_edges_for_bb): Ditto.
22623 (initialize_cfun): Ditto.
22624 (copy_cfg_body): Ditto.
22625 (expand_call_inline): Ditto.
22627 2013-04-08 Kai Tietz <ktietz@redhat.com>
22629 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
22630 TARGET_CYGWIN64 by TARGET_64BIT.
22632 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
22634 * config/epiphany/epiphany.md (GPR_1): New constant.
22635 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
22636 * config/epiphany/epiphany.c (gen_compare_reg):
22637 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
22638 is already in place.
22639 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
22640 Don't require being called during rtl expansion; If y operlaps r0,
22642 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
22643 (epiphany_expand_epilogue): Likewise.
22645 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
22646 Don't use CC_FPmode for ORDERED / UNORDERED.
22647 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
22649 * config/epiphany/constraints.md (CnL): New constraint.
22650 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
22651 * config/epiphany/predicates.md (add_operand): Allow 1024.
22653 * config/epiphany/epiphany.md (logical_op): New code iterator.
22654 (op_mnc): New code attribute.
22655 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
22656 (mov_f+1, mov_f+2): New peephole2 patterns.
22658 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
22659 (cstoresi4): Also allow re-use of zero result when doing a NE
22660 comparison to a non-zero operand.
22661 Use (clobber (scratch)) for first insn if the gpr output is not needed.
22663 * config/epiphany/epiphany.md (<insn_opname>v2si3):
22664 Use gen_addsi3_i / gen_subsi3_i.
22666 2013-04-08 Jakub Jelinek <jakub@redhat.com>
22670 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
22671 contain anything but clobbers, at most one __builtin_stack_restore,
22672 optionally debug stmts and final resx, and if it has at least one
22673 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
22674 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
22675 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
22676 which isn't defaut definition, remove them.
22677 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
22678 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
22679 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
22680 with MEM_REF LHS with SSA_NAME address.
22682 2013-04-08 Jeff Law <law@redhat.com>
22684 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
22686 2013-04-08 Richard Biener <rguenther@suse.de>
22688 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
22690 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
22691 determined vector type.
22692 (vect_analyze_data_refs): Likewise.
22693 (vect_get_new_vect_var): Adjust.
22694 (vect_create_destination_var): Preserve SSA name versions.
22695 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
22696 not dump anything here.
22698 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
22700 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
22701 Add member lr_slot_known.
22702 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
22704 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
22705 Remove code that sets lr_slot_offset according to what a previous
22706 version of epiphany_emit_save_restore used to do.
22707 (epiphany_emit_save_restore): When doing an lr save or restore,
22708 set/verify lr_slot_known and lr_slot_offset.
22710 2013-04-08 Xinyu Qi <xyqi@marvell.com>
22713 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
22716 2013-04-08 Richard Biener <rguenther@suse.de>
22718 * alias.c (find_base_term): Fix thinko in previous change.
22720 2013-04-08 Jakub Jelinek <jakub@redhat.com>
22722 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
22723 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
22724 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
22725 if possible to compute val.
22726 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
22727 For QImode integers don't require anything about precision. Use
22728 const_with_all_bytes_same to find out if the constant doesn't have
22729 repeated bytes in it.
22731 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22733 * config/s390/s390.c (s390_expand_insv): Only accept insertions
22736 2013-04-08 Marek Polacek <polacek@redhat.com>
22738 PR rtl-optimization/48182
22739 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
22742 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22745 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
22746 nuses, make sure we have a label.
22748 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22751 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
22752 (rs6000_emit_swdiv_low_precision): Remove.
22753 (rs6000_emit_swdiv): Rewrite to handle between one and four
22754 iterations of Newton-Raphson generally; modify required number of
22755 iterations for some cases.
22756 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
22758 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
22760 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
22761 set-but-unused variable.
22763 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
22764 basic blocks of released function bodies garbage-collectable.
22766 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
22767 (struct rtl_opt_pass): Add TODO_df_finish.
22769 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
22771 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22773 * config/arm/constraints.md (q): New constraint.
22774 * config/arm/ldrdstrd.md: New file.
22775 * config/arm/arm.md (ldrdstrd.md) New include.
22776 (arm_movdi): Use "q" instead of "r" constraint
22777 for double-word memory access.
22778 (movdf_soft_insn): Likewise.
22779 * config/arm/vfp.md (movdi_vfp): Likewise.
22780 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
22781 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
22782 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
22783 (mem_ok_for_ldrd_strd): Likewise.
22784 (output_move_double): Update assertion.
22786 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22788 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
22790 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22792 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
22793 define_insn_and_split.
22794 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
22796 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22798 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
22799 define_insn_and_split.
22800 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
22801 (shiftsi3_compare): New pattern.
22802 (rrx): New pattern.
22803 * config/arm/unspecs.md (UNSPEC_RRX): New.
22805 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22807 * config/arm/arm.md (negdi_extendsidi): New pattern.
22808 (negdi_zero_extendsidi): Likewise.
22810 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22812 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
22813 define_insn_and_split.
22814 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
22815 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
22817 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22819 * config/arm/arm.md (arm_subdi3): Convert define_insn into
22820 define_insn_and_split.
22821 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
22822 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
22824 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22826 * config/arm/arm.md (subsi3_carryin): New pattern.
22827 (subsi3_carryin_const): Likewise.
22828 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
22829 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
22831 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22833 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
22835 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
22837 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
22838 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
22840 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22842 * config/arm/arm.c (arm_expand_builtin): Change fcode
22843 type to unsigned int.
22845 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22847 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
22849 2013-04-04 Ian Lance Taylor <iant@google.com>
22851 * doc/standards.texi (Standards): The Go frontend supports the Go 1
22854 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
22856 PR middle-end/56729
22857 * df-scan.c (df_insn_delete): Disable failing assert.
22859 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22861 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
22862 New function prototype.
22863 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
22864 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
22865 (arm_builtin_vectorized_function): New function.
22867 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22869 * config/arm/arm_neon_builtins.def: New file.
22870 * config/arm/arm.c (neon_builtin_data): Move contents to
22871 arm_neon_builtins.def.
22872 (enum arm_builtins): Include neon builtin definitions.
22873 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
22874 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
22876 2013-04-04 Marek Polacek <polacek@redhat.com>
22878 PR tree-optimization/48186
22879 * predict.c (maybe_hot_frequency_p): Return false if
22880 HOT_BB_FREQUENCY_FRACTION is 0.
22881 (cgraph_maybe_hot_edge_p): Likewise.
22883 2013-04-04 Richard Biener <rguenther@suse.de>
22885 PR tree-optimization/56826
22886 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
22889 2013-04-04 Richard Biener <rguenther@suse.de>
22891 PR tree-optimization/56213
22892 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
22893 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
22895 2013-04-04 Richard Biener <rguenther@suse.de>
22897 PR tree-optimization/56837
22898 * tree-loop-distribution.c (classify_partition): For non-zero
22899 values require that the value has the same precision as its
22900 mode to be useful as memset value.
22902 2013-04-03 Nick Clifton <nickc@redhat.com>
22904 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
22905 (fmssf4): Use fmsf.s on E3V5 architectures.
22906 (fnmasf4): Use fnmaf.s on E3V5 architectures.
22907 (fnmssf4): Use fnmsf.s on E3V5 architectures.
22909 2013-04-03 Jeff Law <law@redhat.com>
22911 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
22912 (lra-eliminations.o): Likewise.
22914 2013-04-03 Teresa Johnson <tejohnson@google.com>
22916 * gcov-io.c (compute_working_sets): Moved most of body of old
22917 compute_working_sets here from profile.c.
22918 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
22919 (gcov_working_set_t): Moved typedef here from basic-block.h
22920 (compute_working_set): Declare.
22921 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
22922 (get_working_sets): Renamed from compute_working_set,
22923 replace most of body with call to new compute_working_sets.
22924 (get_exec_counts): Replace call to compute_working_sets
22925 to get_working_sets.
22926 * profile.h (get_working_sets): Renamed from compute_working_set.
22927 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
22928 to get_working_sets.
22929 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
22930 * gcov-dump.c (dump_working_sets): New function.
22932 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
22934 * hwint.c (sext_hwi, zext_hwi): New functions.
22935 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
22936 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
22937 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
22938 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
22939 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
22940 (sext_hwi, zext_hwi): New functions.
22942 2013-04-03 Jeff Law <law@redhat.com>
22944 PR tree-optimization/56799
22945 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
22946 back test for widening conversion erroneously dropped in prior change.
22948 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22951 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
22952 instead of next_real_insn.
22954 2013-04-03 Marek Polacek <polacek@redhat.com>
22957 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
22959 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22962 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
22964 (thumb1_output_casesi): Likewise.
22965 (thumb2_output_casesi): Likewise.
22967 2013-04-03 Richard Biener <rguenther@suse.de>
22969 PR tree-optimization/56817
22970 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
22972 (tree_unroll_loops_completely_1): ... new function to manually
22973 walk the loop tree, properly defering outer loops of unrolled
22974 loops to later iterations.
22976 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22978 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
22979 (vectorizable_load): Likewise.
22980 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
22981 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
22983 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22985 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
22988 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
22990 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
22992 2013-04-03 Bin Cheng <bin.cheng@arm.com>
22994 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
22996 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22998 PR tree-optimization/56790
22999 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23002 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23004 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23006 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23007 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23010 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23013 * tree.h (force_folding_builtin_constant_p): New decl.
23014 * builtins.c (force_folding_builtin_constant_p): New variable.
23015 (fold_builtin_constant_p): Fold immediately also if
23016 force_folding_builtin_constant_p.
23018 2013-04-03 Richard Biener <rguenther@suse.de>
23020 PR tree-optimization/56812
23021 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23022 DRs of the same interleaving chain are independent.
23024 2013-04-02 Jason Merrill <jason@redhat.com>
23026 * gdbinit.in (pbb): Use debug fn.
23028 2013-04-02 Lawrence Crowl <crowl@google.com>
23030 * sese.h (struct ivtype_map_elt_s): Remove unused.
23031 (extern debug_ivtype_map): Remove unused.
23032 (extern eq_ivtype_map_elts): Remove unused.
23033 * sese.c (debug_ivtype_map): Removed unused.
23034 (debug_ivtype_map_1): Removed unused.
23035 (debug_ivtype_elt): Remove unused.
23036 (eq_ivtype_map_elts): Remove unused.
23039 2013-04-02 Kai Tietz <ktietz@redhat.com>
23042 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23043 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23044 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23046 (legitimize_pe_coff_symbol): Likewise.
23047 (is_imported_p): New helper-function.
23048 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23049 for Windows x64 targets.
23050 (ix86_expand_prologue): Optimize for pe-coff targets.
23051 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23052 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23053 medium/large code-model.
23054 (legitimize_pic_address): Likewise.
23055 (legitimize_tls_address): Likewise.
23056 (ix86_expand_call): Likewise.
23057 (x86_output_mi_thunk): Likewise.
23058 (get_dllimport_decl): Add new beimport argument.
23059 (construct_plt_address): Don't assert for x64 pe-coff targets.
23060 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23062 (SYMBOL_FLAG_STUBVAR): New macro.
23063 (SYMBOL_REF_STUBVAR_P): Likewise.
23064 * config/i386/winnt.c (stub_list): New structure.
23065 (stub_head): New local variable.
23066 (i386_pe_record_stub): New function.
23067 (i386_pe_file_end): Emit refptr-stubs.
23069 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23071 PR rtl-optimization/56745
23072 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23073 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23076 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23077 and both of them are MEM_REFs, just compare first argument for
23078 equality and attempt to deal even with differing offsets.
23081 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23082 of gimple_clobber_p to be MEM_REF.
23083 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23084 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23085 after gimplification.
23086 * asan.c (get_mem_ref_of_assignment): Don't instrument
23087 gimple_clobber_p stmts.
23088 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23089 gimple_clobber_p stmt if they have MEM_REF lhs and
23090 are dead because of another gimple_clobber_p stmt.
23091 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23092 gimple_clobber_p stmts like debug stmts.
23093 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23094 that refer to unused VAR_DECLs or uninitialized values.
23095 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23096 gimple_clobber_p stmts if they refer to removed parameters.
23097 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23100 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23102 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23103 using SWI48 mode attribute.
23105 2013-04-02 Wei Mi <wmi@google.com>
23107 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23108 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23109 *<rotate_insn><mode>3_mask in i386.md.
23111 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23113 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23115 2013-04-02 Richard Biener <rguenther@suse.de>
23117 PR tree-optimization/56778
23118 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23119 Runtime alias tests are not supported for gather loads.
23120 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23121 stmts referenced from SSA operands before updating SSA form.
23123 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23124 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23126 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23127 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23128 * config/arm/cortex-a53.md: New file.
23129 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23130 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23131 * config/arm/arm.c (arm_issue_rate): Likewise.
23132 * config/arm/arm-tune.md: Regenerate
23133 * config/arm/arm-tables.opt: Regenerate.
23134 * config/arm/arm-cores.def: Add cortex-a53.
23136 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23138 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23141 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23143 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23144 scalar load/store operations using B/H registers.
23145 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23147 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23149 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23151 * config/aarch64/aarch64.c
23152 (aarch64_simd_scalar_immediate_valid_for_move): New.
23153 * config/aarch64/aarch64-protos.h
23154 (aarch64_simd_scalar_immediate_valid_for_move): New.
23155 * config/aarch64/constraints.md (Dh, Dq): New.
23156 * config/aarch64/iterators.md (hq): New.
23158 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23160 * reorg.c (get_branch_condition): Deal with conditional returns.
23161 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23163 2013-04-01 Wei Mi <wmi@google.com>
23165 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23166 Truncate operand 2 using %b asm operand modifier.
23167 (*<shift_insn><mode>3_mask): Ditto.
23168 (*<rotate_insn><mode>3_mask): Ditto.
23170 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23172 PR middle-end/56798
23173 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23175 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23177 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23179 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23181 2013-03-30 Lawrence Crowl <crowl@google.com>
23183 * dse.c (clear_alias_sets): Remove never set.
23184 (disqualified_clear_alias_sets): Remove never set.
23185 (clear_alias_mode_pool): Remove never set.
23186 (dse_step0): Remove condition that is never true.
23187 (canon_address): Remove condition that is never true.
23188 (dse_step7): Remove condition that is never true.
23189 (rest_of_handle_dse): Remove condition that is never true.
23190 (rest_of_handle_dse::did_global): Remove never read from above.
23191 (dse_step2_spill): Remove never called from above.
23192 (dse_step5_spill): Remove never called from above.
23194 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23196 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23197 JUMP_TABLE_DATA changes.
23198 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23199 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23200 (Insns) <jump_table_data>: New entry.
23201 * doc/tm.texi: Regenerate.
23203 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23205 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23206 for table jump at the end of a basic block using tablejump_p.
23207 * targhooks.c (default_invalid_within_doloop): Likewise.
23208 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23209 target hook implementation that is identical to the default hook.
23210 (rs6000_invalid_within_doloop): Remove.
23212 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23213 unused variable from tablejump_p call.
23215 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
23216 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
23217 (INSN_DELETED_P): Likewise.
23218 (emit_jump_table_data): New prototype.
23219 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
23220 after 4th as unused.
23221 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
23222 * sched-vis.c (print_insn): Likewise.
23223 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
23224 insn for compatibility with back ends that use next_active_insn to
23225 identify jump table data.
23226 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
23227 (remove_insn): Likewise.
23228 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
23230 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
23231 (emit_jump_table_data): New function.
23233 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
23234 basic block, a JUMP_TABLE_DATA never is.
23235 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
23236 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
23237 off from code handling real insns.
23238 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
23239 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
23240 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
23241 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
23242 is not a NONDEBUG_INSN_P.
23243 * ira-costs.c (scan_one_insn): Likewise.
23244 * jump.c (mark_all_labels): Likewise.
23245 (mark_jump_label_1): Likewise.
23246 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23247 * lra.c (get_insn_freq): Expect all insns reaching here to be in
23249 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
23250 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
23251 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
23252 JUMP_TABLE_DATA_P insns.
23253 (calculate_elim_costs_all_insns): Likewise.
23254 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
23255 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
23256 (delete_output_reload): Code style fixups.
23257 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
23258 insn flags on this non-insn.
23259 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
23260 as scheduling barriers, for pre-change compatibility.
23261 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
23262 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
23264 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
23265 redundant JUMP_TABLE_DATA_P test.
23266 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
23267 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23268 (frv_for_each_packet): Likewise.
23269 * config/i386/i386.c (min_insn_size): Likewise.
23270 (ix86_avoid_jump_mispredicts): Likewise.
23271 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23272 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
23273 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23274 (mips16_insn_length): Robustify.
23275 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
23276 (mips16_split_long_branches): Likewise.
23277 * config/pa/pa.c (pa_combine_instructions): Likewise.
23278 * config/rs6000/rs6000.c (get_next_active_insn): Treat
23279 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
23280 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
23281 as contributing to pool range lengths.
23282 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
23283 Remove redundant JUMP_TABLE_DATA_P test.
23284 (sh_loop_align): Likewise.
23285 (split_branches): Likewise.
23286 (sh_insn_length_adjustment): Likewise.
23287 * config/spu/spu.c (get_branch_target): Likewise.
23289 2013-03-29 Jan Hubicka <jh@suse.cz>
23291 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
23292 gcov streaming; stream hot bb threshold to ltrans.
23293 * predict.c (get_hot_bb_threshold): Break out from ....
23294 (maybe_hot_count_p): ... here.
23295 (set_hot_bb_threshold): New function.
23296 * lto-section-in.c (lto_section_name): Add profile.
23297 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
23298 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
23299 and data-streamer.h
23300 (histogram_entry): New structure.
23301 (histogram, histogram_pool): New global vars.
23302 (histogram_hash): New structure.
23303 (histogram_hash::hash): New method.
23304 (histogram_hash::equal): Likewise.
23305 (account_time_size): New function.
23306 (cmp_counts): New function.
23307 (dump_histogram): New function.
23308 (ipa_profile_generate_summary): New function.
23309 (ipa_profile_write_summary): New function.
23310 (ipa_profile_read_summary): New function.
23311 (ipa_profile): Decide on threshold.
23312 (pass_ipa_profile): Add ipa_profile_write_summary and
23313 ipa_profile_read_summary.
23314 * Makefile.in (ipa.o): Update dependencies.
23315 * lto-streamer.h (LTO_section_ipa_profile): New section.
23317 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
23319 * tree.h (VAR_P): New.
23321 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
23324 * doc/invoke.texi ([-fwhole-program]): Fix typo.
23326 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
23328 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
23329 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23330 (control_flow_insn_p): Likewise.
23331 * cfgrtl.c (duplicate_insn_chain): Likewise.
23332 * final.c (get_attr_length_1): Likewise.
23333 (shorten_branches): Likewise.
23334 (final_scan_insn): Likewise.
23335 * function.c (instantiate_virtual_regs): Likewise.
23336 * gcse.c (insert_insn_end_basic_block): Likewise.
23337 * ira-costs.c (scan_one_insn): Likewise.
23338 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23339 * lra.c (check_rtl): Likewise.
23340 * reload1.c (elimination_costs_in_insn): Likewise.
23341 * reorg.c (follow_jumps): Likewise.
23343 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
23344 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23345 (thumb_far_jump_used_p): Likewise.
23346 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
23347 (workaround_speculation): Likewise.
23348 (add_sched_insns_for_speculation): Likewise.
23349 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
23350 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23351 (frv_for_each_packet): Likewise.
23352 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
23353 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
23354 (final_emit_insn_group_barriers): Likewise.
23355 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23356 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23357 (mips16_insn_length): Likewise.
23358 * config/pa/pa.c (pa_reorg): Likewise.
23359 (pa_combine_instructions): Likewise.
23360 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
23361 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
23362 (sh_reorg): Likewise.
23363 (split_branches): Likewise.
23364 * config/spu/spu.c (get_branch_target): Likewise.
23366 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
23369 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
23371 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
23372 Fix declaration name.
23374 2013-03-28 Lawrence Crowl <crowl@google.com>
23376 * graphds.h (struct graph.indicies): Remove unused.
23377 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
23378 (SCOP_ORIGINAL_PDDRS): Remove unused.
23379 * sese.h (extern insert_loop_close_phis): Removed unused.
23380 (extern insert_guard_phis): Removed unused.
23381 (extern ivtype_map_elt_info): Removed unused.
23382 (new_ivtype_map_elt): Removed unused.
23383 * sese.c (ivtype_map_elt_info): Removed unused.
23385 2013-03-28 Lawrence Crowl <crowl@google.com>
23387 * Makefile.in: Add several missing include dependences.
23389 (test-dump.o): New. This object is not added to any executable,
23390 but is present for ad-hoc testing.
23392 (debug (const bitmap_head_def &)): New.
23393 (debug (const bitmap_head_def *)): New.
23395 (extern debug (const bitmap_head_def &)): New.
23396 (extern debug (const bitmap_head_def *)): New.
23398 (debug (edge_def &)): New.
23399 (debug (edge_def *)): New.
23401 (debug (basic_block_def &)): New.
23402 (debug (basic_block_def *)): New.
23404 (dump_node (const_tree, int, FILE *)): Correct source file.
23406 (debug (die_struct &)): New.
23407 (debug (die_struct *)): New.
23409 (extern debug (die_struct &)): New.
23410 (extern debug (die_struct *)): New.
23411 * gimple-pretty-print.c
23412 (debug (gimple_statement_d &)): New.
23413 (debug (gimple_statement_d *)): New.
23414 * gimple-pretty-print.h
23415 (extern debug (gimple_statement_d &)): New.
23416 (extern debug (gimple_statement_d *)): New.
23418 (debug (ira_allocno_copy &)): New.
23419 (debug (ira_allocno_copy *)): New.
23420 (debug (ira_allocno &)): New.
23421 (debug (ira_allocno *)): New.
23423 (extern debug (ira_allocno_copy &)): New.
23424 (extern debug (ira_allocno_copy *)): New.
23425 (extern debug (ira_allocno &)): New.
23426 (extern debug (ira_allocno *)): New.
23428 (debug (live_range &)): New.
23429 (debug (live_range *)): New.
23431 (debug (lra_live_range &)): New.
23432 (debug (lra_live_range *)): New.
23434 (debug (lra_live_range &)): New.
23435 (debug (lra_live_range *)): New.
23437 (debug (omega_pb_d &)): New.
23438 (debug (omega_pb_d *)): New.
23440 (extern debug (omega_pb_d &)): New.
23441 (extern debug (omega_pb_d *)): New.
23443 (debug (const rtx_def &)): New.
23444 (debug (const rtx_def *)): New.
23446 (debug_tree (tree): Move within file.
23447 (debug_raw (const tree_node &)): New.
23448 (debug_raw (const tree_node *)): New.
23449 (dump_tree_via_hooks (const tree_node *, int)): New.
23450 (debug (const tree_node &)): New.
23451 (debug (const tree_node *)): New.
23452 (debug_verbose (const tree_node &)): New.
23453 (debug_verbose (const tree_node *)): New.
23454 (debug_head (const tree_node &)): New.
23455 (debug_head (const tree_node *)): New.
23456 (debug_body (const tree_node &)): New.
23457 (debug_body (const tree_node *)): New.
23458 (debug_vec_tree (tree): Move and reimplement in terms of dump.
23459 (debug (vec<tree, va_gc> &)): New.
23460 (debug (vec<tree, va_gc> *)): New.
23462 (extern debug (const rtx_def &)): New.
23463 (extern debug (const rtx_def *)): New.
23465 (debug_raw (simple_bitmap_def &)): New.
23466 (debug_raw (simple_bitmap_def *)): New.
23467 (debug (simple_bitmap_def &)): New.
23468 (debug (simple_bitmap_def *)): New.
23470 (extern debug (simple_bitmap_def &)): New.
23471 (extern debug (simple_bitmap_def *)): New.
23472 (extern debug_raw (simple_bitmap_def &)): New.
23473 (extern debug_raw (simple_bitmap_def *)): New.
23475 (debug (vinsn_def &)): New.
23476 (debug (vinsn_def *)): New.
23477 (debug_verbose (vinsn_def &)): New.
23478 (debug_verbose (vinsn_def *)): New.
23479 (debug (expr_def &)): New.
23480 (debug (expr_def *)): New.
23481 (debug_verbose (expr_def &)): New.
23482 (debug_verbose (expr_def *)): New.
23483 (debug (vec<rtx> &)): New.
23484 (debug (vec<rtx> *)): New.
23486 (extern debug (vinsn_def &)): New.
23487 (extern debug (vinsn_def *)): New.
23488 (extern debug_verbose (vinsn_def &)): New.
23489 (extern debug_verbose (vinsn_def *)): New.
23490 (extern debug (expr_def &)): New.
23491 (extern debug (expr_def *)): New.
23492 (extern debug_verbose (expr_def &)): New.
23493 (extern debug_verbose (expr_def *)): New.
23494 (extern debug (vec<rtx> &)): New.
23495 (extern debug (vec<rtx> *)): New.
23497 (_list_iter_cond_expr): Make inline instead of static.
23499 (debug (sreal &)): New.
23500 (debug (sreal *)): New.
23502 (extern debug (sreal &)): New.
23503 (extern debug (sreal *)): New.
23505 (extern debug_raw (const tree_node &)): New.
23506 (extern debug_raw (const tree_node *)): New.
23507 (extern debug (const tree_node &)): New.
23508 (extern debug (const tree_node *)): New.
23509 (extern debug_verbose (const tree_node &)): New.
23510 (extern debug_verbose (const tree_node *)): New.
23511 (extern debug_head (const tree_node &)): New.
23512 (extern debug_head (const tree_node *)): New.
23513 (extern debug_body (const tree_node &)): New.
23514 (extern debug_body (const tree_node *)): New.
23515 (extern debug (vec<tree, va_gc> &)): New.
23516 (extern debug (vec<tree, va_gc> *)): New.
23518 (debug (struct loop &)): New.
23519 (debug (struct loop *)): New.
23520 (debug_verbose (struct loop &)): New.
23521 (debug_verbose (struct loop *)): New.
23522 * tree-dump.c: Add header dependence.
23524 (extern debug (struct loop &)): New.
23525 (extern debug (struct loop *)): New.
23526 (extern debug_verbose (struct loop &)): New.
23527 (extern debug_verbose (struct loop *)): New.
23529 (debug (data_reference &)): New.
23530 (debug (data_reference *)): New.
23531 (debug (vec<data_reference_p> &)): New.
23532 (debug (vec<data_reference_p> *)): New.
23533 (debug (vec<ddr_p> &)): New.
23534 (debug (vec<ddr_p> *)): New.
23536 (extern debug (data_reference &)): New.
23537 (extern debug (data_reference *)): New.
23538 (extern debug (vec<data_reference_p> &)): New.
23539 (extern debug (vec<data_reference_p> *)): New.
23540 (extern debug (vec<ddr_p> &)): New.
23541 (extern debug (vec<ddr_p> *)): New.
23543 (debug (pt_solution &)): New.
23544 (debug (pt_solution *)): New.
23546 (extern debug (pt_solution &)): New.
23547 (extern debug (pt_solution *)): New.
23549 (debug (_var_map &)): New.
23550 (debug (_var_map *)): New.
23551 (debug (tree_live_info_d &)): New.
23552 (debug (tree_live_info_d *)): New.
23554 (extern debug (_var_map &)): New.
23555 (extern debug (_var_map *)): New.
23556 (extern debug (tree_live_info_d &)): New.
23557 (extern debug (tree_live_info_d *)): New.
23559 2013-03-28 Jan Hubicka <jh@suse.cz>
23561 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
23563 2013-03-28 Ian Bolton <ian.bolton@arm.com>
23565 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
23566 record only when desired or required.
23568 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
23570 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
23571 *vec_extractv2di_1_rex64. Use x64 isa attribute.
23573 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23575 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
23576 (*andsi3_compare0_uxtw): New pattern.
23577 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
23578 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
23580 2013-03-28 Jan Hubicka <jh@suse.cz>
23582 * data-streamer-in.c (streamer_read_gcov_count): New function.
23583 * gimple-streamer-out.c: Include value-prof.h.
23584 (output_gimple_stmt): Output histogram.
23585 (output_bb): Use streamer_write_gcov_count.
23586 * value-prof.c: Include data-streamer.h
23587 (dump_histogram_value): Add HIST_TYPE_MAX.
23588 (stream_out_histogram_value): New function.
23589 (stream_in_histogram_value): New function.
23590 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
23591 (stream_out_histogram_value, stream_in_histogram_value): Declare.
23592 * data-streamer-out.c (streamer_write_gcov_count): New function.
23593 (streamer_write_gcov_count_stream): New function.
23594 * lto-cgraph.c (lto_output_edge): Update counter streaming.
23595 (lto_output_node): Likewise.
23596 (input_node, input_edge): Likewise.
23597 * lto-streamer-out.c (output_cfg): Update streaming.
23598 * lto-streamer-in.c (input_cfg): Likewise.
23599 * data-streamer.h (streamer_write_gcov_count,
23600 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
23601 * gimple-streamer-in.c: Include value-prof.h
23602 (input_gimple_stmt): Input histograms.
23603 (input_bb): Update profile streaming.
23605 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
23607 * genmodes.c (emit_max_int): New function.
23608 (emit_insn_modes_h): Added call to emit_max_function.
23609 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
23611 * machmode.def: Fixed comment.
23613 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
23615 * combine.c (try_combine): Removed useless assert.
23616 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
23618 2013-03-28 Marek Polacek <polacek@redhat.com>
23619 Richard Biener <rguenther@suse.de>
23621 PR tree-optimization/56695
23622 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
23623 build signed result of a vector comparison.
23624 * tree-cfg.c (verify_gimple_comparison): Check that a result
23625 of a vector comparison has signed type.
23627 2013-03-28 Richard Biener <rguenther@suse.de>
23629 PR tree-optimization/37021
23630 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
23631 do not restrict gaps between groups.
23632 * tree-vect-stmts.c (vectorizable_load): Properly account for
23633 a gap between groups.
23635 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
23637 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
23638 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
23641 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
23643 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
23644 * doc/extend.texi (Named Address Spaces): Ditto.
23645 (Variable Attributes): Ditto.
23647 2013-03-27 Kai Tietz <ktietz@redhat.com>
23649 * config.build: Add support for cygwin x64 target.
23650 * config.gcc: Likewise.
23651 * config.host: Likewise.
23652 * configure.ac: Likewise
23653 * configure: Regenerated.
23655 2013-03-27 Kai Tietz <ktietz@redhat.com>
23657 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
23658 * config/i386/t-cygwin-w64: New file.
23659 * config/i386/cygwin-w64.h: New file.
23660 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
23661 and add support for x64-cygwin target.
23662 (CPP_SPEC): Likewise.
23663 (CXX_WRAP_SPEC_LIST): Undefine before define.
23664 (LIBGCJ_SONAME): Use 15 as version.
23666 2013-03-27 Richard Biener <rguenther@suse.de>
23668 PR tree-optimization/56716
23669 * tree-ssa-structalias.c (perform_var_substitution): Adjust
23670 dumping for ref nodes.
23672 2013-03-27 Martin Jambor <mjambor@suse.cz>
23674 PR tree-optimization/55334
23675 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
23676 restricted pointers to arrays.
23678 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
23680 * Makefile.in (.SUFFIXES): Add .cc.
23681 (.c.o): Apply same recipe for implicit rule .cc.o.
23683 2013-03-27 Richard Biener <rguenther@suse.de>
23685 PR tree-optimization/37021
23686 * tree-vect-data-refs.c (vect_check_strided_load): Allow
23687 REALPART/IMAGPART_EXPRs around the supported refs.
23688 * tree-ssa-structalias.c (find_func_aliases): Assume that
23689 floating-point values are not used to transfer pointers.
23691 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
23693 * target.def (TARGET_HAS_IFUNC_P): New target hook.
23694 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
23695 * doc/tm.texi: Regenerate.
23696 * targhooks.h (default_has_ifunc_p): New.
23697 * targhooks.c (default_has_ifunc_p): Ditto.
23698 * config/linux-protos.h: New file.
23699 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
23700 hook for linux which disables support of indirect functions in android.
23701 * config/linux-android.c: New file.
23702 * config/t-linux-android.c: Ditto.
23703 * config.gcc: Added new object file linux-android.o.
23704 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
23705 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
23706 * varasm.c (do_assemble_alias): Likewise.
23707 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
23708 doesn't support indirect functions.
23709 * configure: Regenerate.
23711 2013-03-27 Bin Cheng <bin.cheng@arm.com>
23714 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
23715 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
23718 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23720 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
23722 2013-03-27 Terry Guo <terry.guo@arm.com>
23724 * config/arm/arm-cores.def: Added core cortex-r7.
23725 * config/arm/arm-tune.md: Regenerated.
23726 * config/arm/arm-tables.opt: Regenerated.
23727 * doc/invoke.texi: Added entry for core cortex-r7.
23729 2013-03-27 Walter Lee <walt@tilera.com>
23731 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
23732 double-decrement of next_scratch_regno.
23734 2013-03-27 Walter Lee <walt@tilera.com>
23736 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
23738 (insn_v1mulus): Ditto.
23739 (insn_v2muls): Ditto.
23741 2013-03-27 Walter Lee <walt@tilera.com>
23743 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
23744 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
23746 2013-03-27 Walter Lee <walt@tilera.com>
23748 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
23749 (*sibcall_value): Ditto.
23751 2013-03-27 Walter Lee <walt@tilera.com>
23753 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
23754 (insn_mnz_v8qi): ... this ...
23755 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
23757 (insn_v<n>mnz): Replaced by ...
23758 (insn_v1mnz): ... this ...
23759 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
23761 (insn_mz_<mode>): Replaced by ...
23762 (insn_mz_v8qi): ... this ...
23763 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
23765 (insn_v<n>mz): Replaced by ...
23766 (insn_v1mz): ... this ...
23767 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
23770 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
23772 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
23774 2013-03-26 Roland McGrath <mcgrathr@google.com>
23776 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
23777 than fprintf with a non-constant, non-format string.
23779 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
23781 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
23782 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
23783 operand 0 predicate.
23784 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
23785 attribute. Use general_x64nomem_operand as operand 1 predicate.
23786 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
23787 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
23788 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
23789 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
23790 (mov<mode>_insv_1): Remove expander. Merge insn with
23791 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
23792 Use general_x64nomem_operand as operand 1 predicate.
23793 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
23794 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
23795 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
23796 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
23797 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
23798 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
23799 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
23800 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
23801 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
23802 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
23803 (general_x64nomem_operand): Ditto.
23805 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
23807 * config/rtems.opt: Add -pthread option.
23809 2013-03-26 Richard Biener <rguenther@suse.de>
23811 * alias.c (find_base_term): Avoid redundant and not used recursion.
23812 (base_alias_check): Get the initial base term from the caller.
23813 (true_dependence_1): Compute and pass base terms to base_alias_check.
23814 (write_dependence_p): Likewise.
23815 (may_alias_p): Likewise.
23817 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
23819 * config/aarch64/aarch64.c (aarch64_classify_address): Support
23820 PC-relative load in SI modes and above only.
23822 2013-03-26 Xinyu Qi <xyqi@marvell.com>
23824 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
23825 * config/arm/iwmmxt.md (WCGR0): Update.
23826 (WCGR1, WCGR2, WCGR3): Likewise.
23828 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
23830 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
23831 Use x64 and nox64 isa attributes.
23833 2013-03-26 Richard Biener <rguenther@suse.de>
23835 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
23836 alignment computations and rely on get_object_alignment_1
23837 for the !TYPE_P case.
23838 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
23840 2013-03-26 Walter Lee <walt@tilera.com>
23842 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
23843 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
23845 2013-03-25 Jeff Law <law@redhat.com>
23847 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
23848 check for INTEGRAL_TYPE_P that was missing due to checking in
23849 wrong version of prior patch.
23851 2013-03-25 Walter Lee <walt@tilera.com>
23853 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
23854 TILEGX_INSN_SHUFFLEBYTES1.
23855 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
23857 (tilegx_builtins): Ditto.
23858 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
23860 2013-03-25 Walter Lee <walt@tilera.com>
23862 * config/tilegx/tilegx.md (floatsisf2): New pattern.
23863 (floatunssisf2): New pattern.
23864 (floatsidf2): New pattern.
23865 (floatunssidf2): New pattern.
23867 2013-03-25 Walter Lee <walt@tilera.com>
23869 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
23870 tests for constraint J, K, N, P.
23872 2013-03-25 Walter Lee <walt@tilera.com>
23874 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
23875 Use indirect/pcrel encoding.
23876 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
23879 2013-03-25 Steve Ellcey <sellcey@mips.com>
23881 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
23882 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
23883 * config/mips/mips.c (mips_option_override): Set IMADD default.
23884 * config/mips/mips.h (PTF_AVOID_IMADD): New.
23885 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
23886 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
23887 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
23888 * doc/invoke.texi (-mimadd/-mno-imadd): New.
23890 2013-03-25 Jeff Law <law@redhat.com>
23892 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
23893 slightly to avoid creating and folding useless trees. Simplify
23894 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
23896 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
23898 * config/i386/i386.md (*zero_extendsidi2): Merge with
23899 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
23900 * config/i386/predicates.md (x86_64_zext_operand): Rename from
23901 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
23902 targets. Clarify comment.
23904 2013-03-25 Martin Jambor <mjambor@suse.cz>
23906 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
23907 pass-through jump functions differently.
23908 (ipa_read_jump_function): Likewise. Also use setter functions to set
23911 2013-03-25 Martin Jambor <mjambor@suse.cz>
23913 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
23914 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
23916 (ipa_get_indirect_edge_target): New function.
23917 (devirtualization_time_bonus): New parameter known_aggs, pass it to
23918 ipa_get_indirect_edge_target. Update all callers.
23919 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
23920 ipa_get_indirect_edge_target_1 instead of calling
23921 ipa_get_indirect_edge_target.
23922 (create_specialized_node): Pass aggvlas to
23923 ipcp_discover_new_direct_edges.
23925 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23927 * config/arm/arm.md (f_sels, f_seld): New types.
23928 (*cmov<mode>): New pattern.
23929 * config/arm/predicates.md (arm_vsel_comparison_operator): New
23932 2013-03-25 Kai Tietz <ktietz@redhat.com>
23934 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
23935 POSIX-printf for mingw-hosted builds.
23937 2013-03-25 Richard Biener <rguenther@suse.de>
23939 PR middle-end/56694
23940 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
23941 must-not-throw stmt location.
23943 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23945 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
23946 Emit load-acquire versions when acq is true.
23947 (arm_emit_store_exclusive): Add rel parameter.
23948 Emit store-release versions when rel is true.
23949 (arm_split_compare_and_swap): Use acquire-release instructions
23951 of barriers when appropriate.
23952 (arm_split_atomic_op): Likewise.
23953 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
23954 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
23955 (VUNSPEC_SLX): Likewise.
23956 (VUNSPEC_LDA): Likewise.
23957 (VUNSPEC_STL): Likewise.
23958 * config/arm/sync.md (atomic_load<mode>): New pattern.
23959 (atomic_store<mode>): Likewise.
23960 (arm_load_acquire_exclusive<mode>): Likewise.
23961 (arm_load_acquire_exclusivesi): Likewise.
23962 (arm_load_acquire_exclusivedi): Likewise.
23963 (arm_store_release_exclusive<mode>): Likewise.
23965 2013-03-25 Catherine Moore <clm@codesourcery.com>
23967 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
23968 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
23969 * config/mip/predicates.md (lwsp_swsp_operand,
23970 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
23971 sb16_operand, db4_operand, db7_operand, ib3_operand,
23972 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
23973 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
23974 andi16_operand): New predicates.
23975 * config/mips/mips.md (compression): New attribute.
23976 (enabled): New attribute.
23977 (length): Consider compression in computing length.
23978 (shift_compression): New code attribute.
23979 (*add<mode>3): New operands. Record compression.
23980 (sub<mode>3): Likewise.
23981 (one_cmpl<mode>2): Likewise.
23982 (*and<mode>3): Likewise.
23983 (*ior<mode>3): Likewise.
23984 (unnamed pattern for xor): Likewise.
23985 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
23986 (*<optab><mode>3): Likewise.
23987 (*mov<mode>_internal: Likewise.
23988 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
23989 (mips_unsigned_immediate_p): New.
23990 (umips_lwsp_swsp_address_p): New.
23991 (m16_based_address_p): New.
23992 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
23993 (mips_unsigned_immediate_p): New prototype.
23994 (lwsp_swsp_address_p): New prototype.
23995 (m16_based_address_p): New prototype.
23996 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
23997 (mips_signed_immediate_p): New function.
23998 (m16_based_address_p): New function.
23999 (lwsp_swsp_address_p): New function.
24000 (mips_print_operand_punctuation): Recognize short delay slot insns
24001 for microMIPS.add<mode>3"
24003 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24006 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24007 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24009 2013-03-25 Richard Biener <rguenther@suse.de>
24011 PR tree-optimization/56689
24012 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24015 2013-03-25 Richard Biener <rguenther@suse.de>
24017 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24019 (memory_references): Likewise.
24020 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24021 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24022 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24023 (gather_mem_refs_in_loops): Fold into ...
24024 (analyze_memory_references): ... this. Move initialization
24025 to tree_ssa_lim_initialize.
24026 (fill_always_executed_in): Rename to ...
24027 (fill_always_executed_in_1): ... this.
24028 (fill_always_executed_in): Move contains_call computation to
24029 this new function from ...
24030 (tree_ssa_lim_initialize): ... here.
24031 (tree_ssa_lim): Call fill_always_executed_in.
24033 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24035 * postreload.c (reload_combine): Fix code detecting returns.
24037 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24039 * function.c (emit_use_return_register_into_block): On cc0 targets,
24040 do not emit the sequence between cc0 setter and user.
24042 2013-03-25 Kai Tietz <ktietz@redhat.com>
24044 * config/i386/predicates.md (local_symbolic_operand): Interpret
24045 dll-imported symbols as none-local.
24047 2013-03-25 Richard Biener <rguenther@suse.de>
24049 * tree-ssa-loop-im.c (struct depend): Remove.
24050 (struct lim_aux_data): Make depends a vec of gimples.
24051 (free_lim_aux_data): Adjust.
24052 (add_dependency): Likewise.
24053 (set_level): Likewise.
24055 2013-03-25 Richard Biener <rguenther@suse.de>
24057 PR middle-end/56434
24058 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24059 the pointer returned by calls with ECF_MALLOC set.
24061 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24063 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24065 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24067 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24068 using MMXMODE mode iterator.
24069 (*move<mode>_internal): Merge with *movv2sf_internal and
24070 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24072 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24074 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24075 (record_last_mem_set_info): Likewise.
24077 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24078 of XNEWVEC followed by memset.
24079 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24081 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24083 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24084 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24085 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24086 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24087 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24088 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24089 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24090 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24091 BARRIER_P instead of GET_CODE.
24093 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24095 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24096 inaccuracy in the probing code.
24098 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24099 (ctrapdi4): Likewise.
24101 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24103 * calls.c (expand_call): Add missing guard to code handling return
24104 of non-BLKmode structures in MSB.
24105 * function.c (expand_function_end): Likewise.
24107 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24109 * combine.c (try_combine): Adjust comment. Do not add the set of
24110 insn #0 if the destination indirectly is set or dies in insn #2.
24111 Tidy up code to distribute a new note.
24113 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24115 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24116 also for alternatives 16 and 17.
24118 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24120 * config/i386/sse.md (*mov<mode>_internal): Merge with
24121 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24122 Emit insn template depending on type attribute. Use
24123 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24124 movd instead of movq mnemonic for interunit moves. Rewrite mode
24125 attribute calculation. Remove unit attribute calculation.
24126 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24127 Set prefix_data16 attribute for DImode ssemov types.
24128 Use Ym instead of y for SSE-MMX conversion alternatives.
24129 Reorder operand constraints.
24131 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24133 * df.h (df_insn_delete): Adjust prototype.
24134 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24135 and let it decide whether mark the basic block dirty.
24136 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24137 * df-scan.c (df_insn_info_delete): New helper function, split
24138 off from df_insn_delete.
24139 (df_scan_free_bb_info): Use it.
24140 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24142 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24143 that the insn is actually an insn and it has a non-NULL basic block.
24144 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24146 2013-03-22 Richard Biener <rguenther@suse.de>
24148 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24150 (mem_ref_alloc): Do not allocate them.
24151 (refs_independent_p): Do not query or maintain a cache.
24153 2013-03-22 Richard Biener <rguenther@suse.de>
24155 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24156 (gather_mem_refs_in_loops): Do not compute it.
24157 (analyze_memory_references): Do not allocate it.
24158 (tree_ssa_lim_finalize): Do not free it.
24159 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24161 2013-03-22 Richard Biener <rguenther@suse.de>
24163 * is-a.h (as_a): Use gcc_checking_assert.
24165 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24167 * config/aarch64/aarch64.c (aarch64_print_operand): New
24168 format specifier for printing a constant in hex.
24169 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24170 format specifier for printing second operand.
24172 2013-03-22 Richard Biener <rguenther@suse.de>
24174 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24176 (gather_mem_refs_in_loops): Perform store accumulation here.
24177 (create_vop_ref_mapping_loop): Remove.
24178 (create_vop_ref_mapping): Likewise.
24179 (analyze_memory_references): Initialize refs_stored_in_loop.
24180 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24181 (record_indep_loop): Remove.
24182 (record_dep_loop): New function.
24183 (ref_indep_loop_p_1): Adjust to only walk over references
24184 in the loop, not its subloops.
24185 (ref_indep_loop_p): Rename to ...
24186 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24187 maintaining a more fine-grained cache.
24188 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24189 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24191 2013-03-22 Richard Biener <rguenther@suse.de>
24193 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24194 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24195 aggregate mem_ref_loc.
24196 (free_mem_ref_locs): Inline into ...
24197 (memref_free): ... this and adjust.
24198 (mem_ref_alloc): Adjust.
24199 (mem_ref_locs_alloc): Remove.
24200 (record_mem_ref_loc): Adjust.
24201 (get_all_locs_in_loop): Rewrite into ...
24202 (for_all_locs_in_loop): ... this iterator.
24203 (rewrite_mem_ref_loc): New functor.
24204 (rewrite_mem_refs): Use for_all_locs_in_loop.
24205 (sm_set_flag_if_changed): New functor.
24206 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24207 (ref_always_accessed): New functor.
24208 (ref_always_accessed_p): Use for_all_locs_in_loop.
24210 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24212 * tree-pass.h (PROP_gimple_lvec): New.
24213 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24214 (init_optimization_passes): Move pass_lower_vector.
24215 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
24217 (pass_lower_vector): Provide PROP_gimple_lvec.
24218 (pass_lower_vector_ssa): Likewise.
24219 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
24221 2013-03-21 Mark Wielaard <mjw@redhat.com>
24223 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
24225 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24227 * config/i386/i386.md (*movdi_internal): Disparage slightly
24228 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
24229 conversion alternatives.
24231 2013-03-21 Jakub Jelinek <jakub@redhat.com>
24233 PR middle-end/48087
24234 * diagnostic.def (DK_WERROR): New kind.
24235 * diagnostic.h (werrorcount): Define.
24236 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
24237 promoted to DK_ERROR, increment DK_WERROR counter instead of
24239 * toplev.c (toplev_main): Call print_ignored_options even if
24240 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
24241 even if just werrorcount is non-zero.
24244 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
24246 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
24247 (string_cst_pool_decl): New function.
24248 (optimize_one_addr_into_implicit_ptr): New function.
24249 (resolve_addr_in_expr): Optimize DWARF location expression
24250 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
24251 which doesn't live in memory, but has DW_AT_location or
24252 DW_AT_const_value, or refers to a string literal, into
24253 DW_OP_GNU_implicit_pointer.
24254 (optimize_location_into_implicit_ptr): New function.
24255 (resolve_addr): If removing DW_AT_location of a variable because
24256 it was DW_OP_addr of address of the variable, but the variable doesn't
24257 live in memory, try to emit const value attribute for the initializer.
24259 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24261 * tree.h (VECTOR_TYPE_P): New macro.
24262 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
24263 TYPE_MODE): Use it.
24264 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
24265 VEC_COND_EXPR cannot be lvalues.
24266 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
24268 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24270 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
24271 Restrict the transformation to equal modes.
24273 2013-03-21 Richard Biener <rguenther@suse.de>
24275 PR tree-optimization/39326
24276 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
24277 (MEM_ANALYZABLE): Adjust.
24278 (record_mem_ref_loc): Move bitmap ops ...
24279 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
24280 unanalyzable refs, do not record locations for it.
24281 (analyze_memory_references): Allocate ref zero as shared
24283 (refs_independent_p): Do not test for unanalyzed mems here.
24284 (ref_indep_loop_p_1): Special-case disambiguation against
24285 the unanalyzed ref.
24286 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
24288 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
24290 * config/arm/arm-protos.h (tune_params): Add
24291 prefer_neon_for_64bits field.
24292 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
24293 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
24294 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
24295 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
24296 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
24297 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
24298 (arm_option_override): Handle -mneon-for-64bits new option.
24299 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
24300 (prefer_neon_for_64bits): Declare new variable.
24301 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
24302 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
24303 (arch_enabled): Handle new arch types. Remove support for onlya8
24305 (one_cmpldi2): Use new arch names.
24306 (zero_extend<mode>di2, extend<mode>di2): Ditto.
24307 * config/arm/arm.opt (mneon-for-64bits): Add option.
24308 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
24309 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
24310 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
24312 * doc/invoke.texi (-mneon-for-64bits): Document.
24314 2013-03-21 Richard Biener <rguenther@suse.de>
24316 PR tree-optimization/39326
24317 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
24318 (sort_bbs_in_loop_postorder_cmp): New function.
24319 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
24321 2013-03-21 Richard Biener <rguenther@suse.de>
24323 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
24324 (vect_insert_into_interleaving_chain): Likewise.
24325 (vect_drs_dependent_in_basic_block): Inline ...
24326 (vect_slp_analyze_data_ref_dependence): ... here. New function,
24328 (vect_analyze_data_ref_dependence): ... here. Simplify.
24329 (vect_check_interleaving): Simplify.
24330 (vect_analyze_data_ref_dependences): Likewise. Split out ...
24331 (vect_slp_analyze_data_ref_dependences): ... this new function.
24332 (dr_group_sort_cmp): New function.
24333 (vect_analyze_data_ref_accesses): Compute data-reference groups
24334 here instead of in vect_analyze_data_ref_dependence. Use
24335 a more efficient algorithm.
24336 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
24337 vect_slp_analyze_data_ref_dependences. Call
24338 vect_analyze_data_ref_accesses earlier.
24339 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
24340 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
24341 (vect_slp_analyze_data_ref_dependences): New prototype.
24343 2013-03-21 Richard Biener <rguenther@suse.de>
24345 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
24346 ref is stored in the loop.
24347 (find_refs_for_sm): Walk only over all stores.
24348 (store_motion_loop): Allocate from lim_bitmap_obstack.
24349 (store_motion): Likewise.
24351 2013-03-21 Richard Biener <rguenther@suse.de>
24353 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
24354 Update virtual SSA form.
24356 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24358 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
24359 * configure: Regenerate.
24360 * config.in: Regenerate.
24361 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
24362 if !HAVE_LD_EH_FRAME_CIEV3.
24364 2013-03-21 Richard Biener <rguenther@suse.de>
24366 * tree-cfg.c (verify_expr_no_block): New function.
24367 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
24368 nor DECL_VALUE_EXPR have locations with associated blocks.
24369 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
24370 (clear_unused_block_pointer): Remove code dealing with
24371 blocks in DECL_DEBUG_EXPR locations.
24373 2013-03-21 Richard Biener <rguenther@suse.de>
24375 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
24376 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
24377 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
24378 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
24379 instead of DECL_DEBUG_EXPR_IS_FROM.
24380 * gimplify.c (gimplify_modify_expr): Likewise.
24381 * tree-cfg.c (verify_expr_location_1): Likewise.
24382 * tree-complex.c (create_one_component_var): Likewise.
24383 * tree-sra.c (create_access_replacement): Likewise.
24384 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
24385 (clear_unused_block_pointer): Likewise.
24386 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
24387 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
24388 * var-tracking.c (var_debug_decl): Likewise.
24389 (track_expr_p): Likewise.
24390 * tree-inline.c (add_local_variables): Likewise. Set
24391 DECL_HAS_DEBUG_EXPR_P after copying it.
24392 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
24393 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
24395 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24398 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
24399 * configure: Regenerate.
24400 * config.in: Regenerate.
24401 * config/i386/i386.md (*movdf_internal): Use
24402 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24403 movd instead of movq mnemonic for interunit moves.
24404 (*movdi_internal): Ditto.
24406 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24408 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
24409 (abd<mode>_3): New pattern.
24410 (aba<mode>_3): New pattern.
24411 (fabd<mode>_3): New pattern.
24413 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24415 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
24416 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
24417 occurrence of REGISTER_PREFIX as its empty string.
24419 2013-03-20 Jeff Law <law@redhat.com>
24421 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
24422 addititional equivalences for equality comparisons between an SSA_NAME
24423 and a constant where the SSA_NAME was set from a widening conversion.
24425 2013-03-20 Walter Lee <walt@tilera.com>
24427 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
24429 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24431 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
24432 depending on type attribute.
24433 (*movti_internal): Ditto.
24434 (*movtf_internal): Ditto.
24435 (*movxf_internal): Ditto.
24436 (*movdf_internal): Ditto.
24437 (*movsf_internal): Ditto.
24439 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24441 * config/i386/i386.md (*movti_internal): Set prefix attribute to
24442 maybe_vex for sselog1 and ssemov types.
24443 (*movdi_internal): Reorder operand constraints.
24444 (*movsi_internal): Ditto. Set prefix attribute to
24445 maybe_vex for sselog1 and ssemov types.
24446 (*movtf_internal): Set prefix attribute to maybe_vex
24447 for sselog1 and ssemov types.
24448 (*movdf_internal): Ditto. Set prefix_data16 attribute for
24449 DImode ssemov types. Reorder operand constraints.
24450 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
24451 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
24452 attribute for SImode ssemov types. Reorder operand constraints.
24454 2013-03-20 Martin Jambor <mjambor@suse.cz>
24456 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
24457 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
24459 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
24461 * config/rs6000/predicates.md (indexed_address, update_address_mem
24462 update_indexed_address_mem): New predicates.
24463 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
24464 attribute for load/store instructions.
24465 * config/rs6000/dfp.md (movsd_store): Likewise.
24466 (movsd_load): Likewise.
24467 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
24468 (unnamed HI->DI extend define_insn): Likewise.
24469 (unnamed SI->DI extend define_insn): Likewise.
24470 (unnamed QI->SI extend define_insn): Likewise.
24471 (unnamed QI->HI extend define_insn): Likewise.
24472 (unnamed HI->SI extend define_insn): Likewise.
24473 (unnamed HI->SI extend define_insn): Likewise.
24474 (extendsfdf2_fpr): Likewise.
24475 (movsi_internal1): Likewise.
24476 (movsi_internal1_single): Likewise.
24477 (movhi_internal): Likewise.
24478 (movqi_internal): Likewise.
24479 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
24480 attribute for load/store instructions.
24481 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
24483 (mov<mode>_softfloat): Likewise.
24484 (mov<mode>_hardfloat32): Likewise.
24485 (mov<mode>_hardfloat64): Likewise.
24486 (mov<mode>_softfloat64): Likewise.
24487 (movdi_internal32): Likewise.
24488 (movdi_internal64): Likewise.
24489 (probe_stack_<mode>): Likewise.
24491 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
24493 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
24494 floating point, and decimal floating point to reload iterator.
24496 * config/rs6000/constraints.md (wl constraint): New constraints to
24497 return FLOAT_REGS if certain options are used to reduce the number
24498 of separate patterns that exist in the file.
24499 (wx constraint): Likewise.
24500 (wz constraint): Likewise.
24502 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
24503 -mdebug=reg, print wg, wl, wx, and wz constraints.
24504 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
24505 Initialize the reload functions for 64-bit binary/decimal floating
24507 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
24508 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
24509 create the buffer on the stack to overcome not having a 32-bit
24511 (rs6000_emit_move): Likewise.
24512 (rs6000_secondary_memory_needed_rtx): Likewise.
24513 (rs6000_alloc_sdmode_stack_slot): Likewise.
24514 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
24515 via xxlxor, just like DFmode 0.0.
24517 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
24518 define as 1 if we are running on a power7 or newer.
24519 (enum r6000_reg_class_enum): Add new constraints.
24521 * config/rs6000/dfp.md (movsd): Delete, combine with binary
24522 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
24523 with other moves by using conditional constraits (wg). Use LFIWZX
24524 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
24525 (movsd splitter): Likewise.
24526 (movsd_hardfloat): Likewise.
24527 (movsd_softfloat): Likewise.
24529 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
24530 binary and decimal floating point moves.
24531 (fmove_ok): New attributes to combine binary and decimal floating
24532 point moves, and to combine power6x (mfpgpr) moves along normal
24534 (real_value_to_target): Likewise.
24535 (f32_lr): Likewise.
24536 (f32_lm): Likewise.
24537 (f32_li): Likewise.
24538 (f32_sr): Likewise.
24539 (f32_sm): Likewise.
24540 (f32_si): Likewise.
24541 (movsf): Combine binary and decimal floating point moves. Combine
24542 power6x (mfpgpr) moves with other moves by using conditional
24543 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
24544 (mov<mode> for SFmode/SDmode); Likewise.
24545 (SFmode/SDmode splitters): Likewise.
24546 (movsf_hardfloat): Likewise.
24547 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
24548 (movsf_softfloat): Likewise.
24549 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
24551 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
24552 wx and wz constraints.
24554 * config/rs6000/constraints.md (wg constraint): New constraint to
24555 return FLOAT_REGS if -mmfpgpr (power6x) was used.
24557 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
24560 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
24561 -mdebug=reg, print wg, wl, wx, and wz constraints.
24562 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
24563 Initialize the reload functions for 64-bit binary/decimal floating
24565 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
24566 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
24567 create the buffer on the stack to overcome not having a 32-bit
24569 (rs6000_emit_move): Likewise.
24570 (rs6000_secondary_memory_needed_rtx): Likewise.
24571 (rs6000_alloc_sdmode_stack_slot): Likewise.
24572 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
24573 via xxlxor, just like DFmode 0.0.
24575 * config/rs6000/dfp.md (movdd): Delete, combine with binary
24576 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
24577 with other moves by using conditional constraits (wg). Use LFIWZX
24578 and STFIWX for loading SDmode on power7.
24579 (movdd splitters): Likewise.
24580 (movdd_hardfloat32): Likewise.
24581 (movdd_softfloat32): Likewise.
24582 (movdd_hardfloat64_mfpgpr): Likewise.
24583 (movdd_hardfloat64): Likewise.
24584 (movdd_softfloat64): Likewise.
24586 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
24587 64-bit binary and decimal floating point moves.
24588 (FMOVE64X): Likewise.
24589 (movdf): Combine 64-bit binary and decimal floating point moves.
24590 Combine power6x (mfpgpr) moves with other moves by using
24591 conditional constraits (wg).
24592 (mov<mode> for DFmode/DDmode): Likewise.
24593 (DFmode/DDmode splitters): Likewise.
24594 (movdf_hardfloat32): Likewise.
24595 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
24596 (movdf_softfloat32): Likewise.
24597 (movdf_hardfloat64_mfpgpr): Likewise.
24598 (movdf_hardfloat64): Likewise.
24599 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
24600 (movdf_softfloat64): Likewise.
24601 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
24602 (reload_<mode>_load): Move to later in the file so they aren't in
24603 the middle of the floating point move insns.
24604 (reload_<mode>_store): Likewise.
24606 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
24609 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
24610 constraint if -mdebug=reg.
24611 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
24612 Enable using dd reload support if needed.
24614 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
24615 binary and decimal floating point moves in rs6000.md.
24616 (movtd_internal): Likewise.
24618 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
24619 decimal floating point moves.
24621 (movtf_internal): Likewise.
24622 (mov<mode>_internal, TDmode/TFmode): Likewise.
24623 (movtf_softfloat): Likewise.
24624 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
24626 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
24627 movdi_internal64, using wg constraint for move direct operations.
24628 (movdi_internal64): Likewise.
24630 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
24631 MODES_TIEABLE_P for selected modes. Print the numerical value of
24632 the various virtual registers. Use GPR/FPR first/last values,
24633 instead of hard coding the register numbers. Print which modes
24634 have reload functions registered.
24635 (rs6000_option_override_internal): If -mdebug=reg, trace the options
24636 settings before/after setting cpu, target and subtarget settings.
24637 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
24638 and for secondary reload failures in rs6000_secondary_reload_inner.
24639 (rs6000_secondary_reload_fail): Likewise.
24640 (rs6000_secondary_reload_inner): Likewise.
24642 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
24643 macros for first/last GPR and FPR registers.
24644 (LAST_GPR_REGNO): Likewise.
24645 (FIRST_FPR_REGNO): Likewise.
24646 (LAST_FPR_REGNO): Likewise.
24648 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
24649 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
24650 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
24651 (vcond<mode><mode>): Likewise.
24652 (vcondu<mode><mode>): Likewise.
24653 (vector_gtu<mode>): Likewise.
24654 (vector_gte<mode>): Likewise.
24655 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
24656 to prevent the compiler from converting DImode operations to TImode.
24657 (ior<mode>3): Likewise.
24658 (and<mode>3): Likewise.
24659 (one_cmpl<mode>2): Likewise.
24660 (nor<mode>3): Likewise.
24661 (andc<mode>3): Likewise.
24663 * config/rs6000/constraints.md (wt constraint): New constraint
24664 that returns VSX_REGS if TImode is allowed in VSX registers.
24666 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
24667 constant under VSX.
24669 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
24670 similar to TImode, but it is restricted to being in the GPRs.
24672 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
24673 TImode to occupy a single VSX register.
24675 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
24676 -mvsx-timode for power7/power8.
24677 (power7 cpu): Likewise.
24678 (power8 cpu): Likewise.
24680 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
24681 sure that TFmode/TDmode take up two registers if they are ever
24682 allowed in the upper VSX registers.
24683 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
24685 (rs6000_init_hard_regno_mode_ok): Likewise.
24686 (rs6000_debug_reg_global): Add debugging for PTImode and wt
24687 constraint. Print if LRA is turned on.
24688 (rs6000_option_override_internal): Give an error if -mvsx-timode
24689 and VSX is not enabled.
24690 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
24691 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
24692 to reg+offset addressing. Use PTImode when checking offset
24693 addresses for validity.
24694 (reg_offset_addressing_ok_p): Likewise.
24695 (rs6000_legitimate_offset_address_p): Likewise.
24696 (rs6000_legitimize_address): Likewise.
24697 (rs6000_legitimize_reload_address): Likewise.
24698 (rs6000_legitimate_address_p): Likewise.
24699 (rs6000_eliminate_indexed_memrefs): Likewise.
24700 (rs6000_emit_move): Likewise.
24701 (rs6000_secondary_reload): Likewise.
24702 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
24703 reloads to fpr registers to continue to use reg+offset addressing,
24704 but 64-bit reloads to altivec registers need reg+reg addressing.
24705 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
24706 it. Treat LO_SUM like a PLUS operation.
24707 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
24708 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
24709 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
24710 registers to share a register with a smaller sized type, since VSX
24711 puts scalars in the upper 64-bits.
24712 (print_operand): Add support for PTImode.
24713 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
24714 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
24715 registers, but don't have arithmetic support.
24716 (rs6000_memory_move_cost): Add test for VSX.
24717 (rs6000_opt_masks): Add -mvsx-timode.
24719 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
24722 (VSr): Use wt constraint for TImode.
24723 (VSv): Drop TImode support.
24724 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
24725 (vsx_movti_64bit): Likewise.
24726 (vsx_movti_32bit): Likewise.
24727 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
24728 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
24729 one '?' on the appropriate output constraint. Do not allow TImode
24730 logical operations on 32-bit systems.
24731 (vsx_ior<mode>3): Likewise.
24732 (vsx_xor<mode>3): Likewise.
24733 (vsx_one_cmpl<mode>2): Likewise.
24734 (vsx_nor<mode>3): Likewise.
24735 (vsx_andc<mode>3): Likewise.
24736 (vsx_concat_<mode>): Likewise.
24737 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
24739 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
24740 OPTION_MASK_VSX_TIMODE.
24741 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
24742 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
24744 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
24745 (TI2 iterator): New iterator for TImode, PTImode.
24746 (wd mode attribute): Add values for vector types.
24747 (movti_string): Replace TI move operations with operations for TImode
24748 and PTImode. Add support for TImode being allowed in VSX registers.
24749 (mov<mode>_string, TImode/PTImode): Likewise.
24750 (movti_ppc64): Likewise.
24751 (mov<mode>_ppc64, TImode/PTImode): Likewise.
24752 (TI mode splitters): Likewise.
24754 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
24757 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
24759 PR tree-optimization/56355
24760 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
24761 Also handle integers with undefined overflow.
24763 2013-03-20 Catherine Moore <clm@codesourcery.com>
24764 Maciej W. Rozycki <macro@codesourcery.com>
24765 Tom de Vries <tom@codesourcery.com>
24766 Nathan Sidwell <nathan@codesourcery.com>
24767 Iain Sandoe <iain@codesourcery.com>
24768 Nathan Froyd <froydnj@codesourcery.com>
24769 Chao-ying Fu <fu@mips.com>
24771 * doc/extend.texi: (micromips, nomicromips, nocompression):
24772 Document new function attributes.
24773 * doc/invoke.texi (minterlink-compressed, mmicromips,
24774 m14k, m14ke, m14kec): Document new options.
24775 (minterlink-mips16): Update documentation.
24776 * doc/md.texi (ZC, ZD): Document new constraints.
24777 * configure.ac (gcc_cv_as_micromips): Check if linker
24778 supports the .set micromips directive.
24779 * configure: Regenerate.
24780 * config.in: Regenerate.
24781 * config/mips/mips-tables.opt: Regenerate.
24782 * config/mips/micromips.md: New file.
24783 * constraints.md (ZC, ZD): New constraints.
24784 * config/mips/predicates.md (movep_src_register): New predicate.
24785 (movep_src_operand): New predicate.
24786 (non_volatile_mem_operand): New predicate.
24787 * config/mips/mips.md (multimem): New type.
24788 (length): Differentiate between 17-bit and 18-bit branch offsets.
24789 (MOVEP1, MOVEP2): New mode iterator.
24790 (mov_<load>l): Use ZC constraint.
24791 (mov_<load>r): Likewise.
24792 (mov_<store>l): Likewise.
24793 (mov_<store>r): Likewise.
24794 (*branch_equality<mode>_inverted): Add microMIPS support.
24795 (*branch_equality<mode>): Likewise.
24796 (*jump_absolute): Likewise.
24797 (indirect_jump_<mode>): Likewise.
24798 (tablejump_<mode>): Likewise.
24799 (<optab>_internal): Likewise.
24800 (sibcall_internal): Likewise.
24801 (sibcall_value_internal): Likewise.
24802 (prefetch): Use constraint ZD.
24803 * config/mips/mips.opt (minterlink-compressed): New option.
24804 (minterlink-mips16): Now an alias for minterlink-compressed.
24805 (mmicromips): New option.
24806 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24807 (compare_and_swap_12): Likewise.
24808 (sync_add<mode>): Likewise.
24809 (sync_<optab>_12): Likewise.
24810 (sync_old_<optab>_12): Likewise.
24811 (sync_new_<optab>_12): Likewise.
24812 (sync_nand_12): Likewise.
24813 (sync_old_nand_12): Likewise.
24814 (sync_new_nand_12): Likewise.
24815 (sync_sub<mode>): Likewise.
24816 (sync_old_add<mode>): Likewise.
24817 (sync_old_sub<mode>): Likewise.
24818 (sync_new_add<mode>): Likewise.
24819 (sync_new_sub<mode>): Likewise.
24820 (sync_<optab><mode>): Likewise.
24821 (sync_old_<optab><mode>): Likewise.
24822 (sync_new_<optab><mode>): Likewise.
24823 (sync_nand<mode>): Likewise.
24824 (sync_old_nand<mode>): Likewise.
24825 (sync_new_nand<mode>): Likewise.
24826 (sync_lock_test_and_set<mode>): Likewise.
24827 (test_and_set_12): Likewise.
24828 (atomic_compare_and_swap<mode>): Likewise.
24829 (atomic_exchange<mode>_llsc): Likewise.
24830 (atomic_fetch_add<mode>_llsc): Likewise.
24831 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24832 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24833 (umips_save_restore_pattern_p): Likewise.
24834 (umips_load_store_pair_p): Likewise.
24835 (umips_output_load_store_pair): Likewise.
24836 (umips_movep_target_p): Likewise.
24837 (umips_12bit_offset_address_p): Likewise.
24838 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24839 (mips_base_mips16): Rename this...
24840 (mips_base_compression_flags): ...to this. Update all uses.
24841 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24842 (mips_mips16_decl_p): Delete.
24843 (mips_nomips16_decl_p): Delete.
24844 (mips_get_compress_on_flags): New function.
24845 (mips_get_compress_off_flags): New function.
24846 (mips_get_compress_mode): New function.
24847 (mips_get_compress_on_name): New function.
24848 (mips_get_compress_off_name): New function.
24849 (mips_insert_attributes): Support multiple compression types.
24850 (mips_merge_decl_attributes): Likewise.
24851 (umips_12bit_offset_address_p): New function.
24852 (mips_start_function_definition): Emit .set micromips directive.
24853 (mips_call_may_need_jalx_p): New function.
24854 (mips_function_ok_for_sibcall): Add microMIPS support.
24855 (mips_print_operand_punctuation): Support short delay slots and
24857 (umips_swm_mask, umips_swm_encoding): New.
24858 (umips_build_save_restore): New function.
24859 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24860 (was_mips16_p): Remove.
24861 (old_compression_mode): New.
24862 (mips_set_compression_mode): New function.
24863 (mips_set_current_function): Add microMIPS support.
24864 (mips_option_override): Likewise.
24865 (umips_save_restore_pattern_p): New function.
24866 (umips_output_save_restore): New function.
24867 (umips_load_store_pair_p_1): New function.
24868 (umips_load_store_pair_p): New function.
24869 (umips_output_load_store_pair_1): New function.
24870 (umips_output_load_store_pair): New function.
24871 (umips_movep_target_p) New function.
24872 (mips_prepare_pch_save): Add microMIPS support.
24873 * config/mips/mips.h (TARGET_COMPRESSION): New.
24874 (TARGET_CPU_CPP_BUILTINS): Update macro
24875 to use new compression flags and to support microMIPS.
24876 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24877 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24878 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24879 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24880 (ASM_SPEC): Support mmicromips and mno-micromips.
24881 (M16STORE_REG_P): New macro.
24882 (MIPS_CALL): Support TARGET_MICROMIPS.
24883 (MICROMIPS_J): New macro.
24884 (mips_base_mips16): Rename this...
24885 (mips_base_compression_flags): ...to this.
24886 (UMIPS_12BIT_OFFSET_P): New macro.
24887 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24888 (MULTILIB_DIRNAMES): Likewise.
24889 2013-03-20 Richard Biener <rguenther@suse.de>
24891 PR tree-optimization/56661
24892 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
24893 the result does not have to be distinct.
24895 2013-03-20 Richard Biener <rguenther@suse.de>
24897 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
24900 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24901 Steven Bosscher <steven@gcc.gnu.org>
24903 PR rtl-optimization/56605
24904 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
24906 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24909 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
24910 that require movd instead of movq.
24912 2013-03-20 Richard Biener <rguenther@suse.de>
24914 * tree-ssa-structalias.c (struct variable_info): Add pointer
24915 to the first field of an aggregate with sub-vars. Make
24916 this and the pointer to the next subfield its ID.
24917 (vi_next): New function.
24918 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
24919 storedanything_id, integer_id): Increment by one.
24920 (new_var_info, get_call_vi, lookup_call_clobber_vi,
24921 get_call_clobber_vi): Adjust.
24922 (solution_set_expand): Simplify and speedup.
24923 (solution_set_add): Inline into ...
24924 (set_union_with_increment): ... this. Adjust accordingly.
24925 (do_sd_constraint): Likewise.
24926 (do_ds_constraint): Likewise.
24927 (do_complex_constraint): Simplify.
24928 (build_pred_graph): Adjust.
24929 (solve_graph): Likewise. Simplify and speedup.
24930 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
24931 get_constraint_for_component_ref, get_constraint_for_1,
24932 first_vi_for_offset, first_or_preceding_vi_for_offset,
24933 create_function_info_for, create_variable_info_for_1,
24934 create_variable_info_for, intra_create_variable_infos): Adjust.
24935 (init_base_vars): Push NULL for ID zero.
24936 (compute_points_to_sets): Adjust.
24938 2013-03-20 Richard Biener <rguenther@suse.de>
24940 * cfgloop.c (verify_loop_structure): Streamline and avoid
24941 ICEing on corrupt loop tree.
24942 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
24945 2013-03-20 Richard Biener <rguenther@suse.de>
24947 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
24948 check whether an SSA update is needed.
24950 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
24952 * config/mips/constraints.md (T): Rename to...
24956 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
24957 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
24959 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24961 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
24962 (*subsi3_carryin_uxtw): Likewise.
24964 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24966 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
24967 (*rorsi3_insn_uxtw): Likewise.
24969 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24971 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
24972 (*extrsi5_insn_uxtw): Likewise.
24974 2013-03-19 Richard Biener <rguenther@suse.de>
24976 PR tree-optimization/56273
24977 * passes.c (init_optimization_passes): Move second VRP after DOM.
24979 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
24981 * config/i386/i386.md (*movti_internal): Merge from
24982 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
24983 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
24984 nox64 isa attributes.
24986 2013-03-18 Richard Biener <rguenther@suse.de>
24988 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
24990 (merge_node_constraints): Likewise.
24991 (build_succ_graph): Likewise.
24992 (valid_graph_edge): Inline into single caller.
24993 (unify_nodes): Likewise. Use bitmap_set_bit return value
24995 (scc_visit): Fix formatting and variable use.
24996 (do_sd_constraint): Use gcc_checking_assert.
24997 (do_ds_constraint): Likewise.
24998 (do_complex_constraint): Likewise.
24999 (condense_visit): Likewise. Cleanup.
25000 (dump_pred_graph): New function.
25001 (perform_var_substitution): Dump the pred-graph before
25002 variable substitution.
25003 (find_equivalent_node): Use gcc_checking_assert.
25004 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25006 2013-03-18 Richard Biener <rguenther@suse.de>
25008 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25009 Remove cond_expr_stmt_list argument and do not gimplify the
25011 (vect_loop_versioning): Adjust.
25012 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25013 Cleanup to use less temporaries.
25014 (vect_create_data_ref_ptr): Cleanup.
25016 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25018 PR tree-optimization/56635
25019 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25020 require types_compatible_p types.
25022 2013-03-18 Nick Clifton <nickc@redhat.com>
25024 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25025 spurious backslash.
25027 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25028 Add missing line to comment describing function.
25030 2013-03-18 Richard Biener <rguenther@suse.de>
25032 PR tree-optimization/56210
25033 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25034 Handle string / character search functions.
25035 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25037 2013-03-18 Richard Biener <rguenther@suse.de>
25039 PR middle-end/56483
25040 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25041 and implement properly.
25042 * gimple.h (gimple_cond_single_var_p): Remove.
25044 2013-03-18 Richard Biener <rguenther@suse.de>
25046 * tree-data-ref.h (find_data_references_in_loop): Declare.
25047 * tree-data-ref.c (get_references_in_stmt): Use a stack
25048 vector pre-allocated in the callers.
25049 (find_data_references_in_stmt): Adjust.
25050 (graphite_find_data_references_in_stmt): Likewise.
25051 (create_rdg_vertices): Likewise.
25052 (find_data_references_in_loop): Export.
25053 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25054 Compute dependences here...
25055 (vect_analyze_data_refs): ...not here. When we encounter
25056 a non-vectorizable data reference in basic-block vectorization
25057 truncate the data reference vector. Do not bother to
25058 fixup data-dependence information for gather loads.
25059 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25060 of data references, as reported.
25062 2013-03-18 Richard Biener <rguenther@suse.de>
25064 PR tree-optimization/3713
25065 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25066 has_constants and expr.
25067 (stmt_has_constants): Properly valueize SSA names when deciding
25068 whether the stmt has constants.
25070 2013-03-18 Richard Biener <rguenther@suse.de>
25072 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25073 whole function when there is nothing to do.
25074 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25075 * tree-vectorizer.c (vectorize_loops): Update virtual and
25076 loop-closed SSA once.
25077 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25079 2013-03-18 Richard Biener <rguenther@suse.de>
25081 PR middle-end/56113
25082 * domwalk.c (bb_postorder): New global static.
25083 (cmp_bb_postorder): New function.
25084 (walk_dominator_tree): Replace scheme imposing an order for
25085 visiting dominator sons by one sorting them at the time they
25086 are pushed on the stack.
25088 2013-03-18 Richard Biener <rguenther@suse.de>
25090 PR tree-optimization/39326
25091 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25092 (struct mem_ref): Replace mem member with ao_ref typed member.
25093 (MEM_ANALYZABLE): Adjust.
25094 (memref_eq): Likewise.
25095 (mem_ref_alloc): Likewise.
25096 (gather_mem_refs_stmt): Likewise.
25097 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25098 (execute_sm_if_changed_flag_set): Adjust.
25099 (execute_sm): Likewise.
25100 (ref_always_accessed_p): Likewise.
25101 (refs_independent_p): Likewise.
25102 (can_sm_ref_p): Likewise.
25104 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25107 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25108 return 1 even for !unsignedp.
25110 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25112 * config/i386/i386.md (isa): Add x64 and nox64.
25113 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25114 (*pushtf): Enable *roF alternative for x64 isa only.
25115 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25116 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25117 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25118 (*movtf_internal): Merge from *movtf_internal_rex64 and
25119 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25120 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25121 nox64 isa attributes.
25122 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25123 nox64 isa attributes.
25124 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25126 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25128 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25130 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25133 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25134 class names. Remove trailing comma after "ALL_REGS".
25136 2013-03-16 Jan Hubicka <jh@suse.cz>
25138 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25139 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25140 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25141 of cgraph_get_create_node.
25142 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25144 2013-03-16 Jason Merrill <jason@redhat.com>
25147 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25148 with DW_AT_default_value.
25150 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25152 * BASE-VER: Set to 4.9.0.
25154 2013-03-14 Andi Kleen <ak@linux.intel.com>
25157 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25158 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25159 Document _x* TSX intrinsics.
25161 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25162 David Holsgrove <david.holsgrove@xilinx.com>
25164 * configure.ac: Add MicroBlaze TLS support detection.
25165 * configure: Regenerate.
25166 * config/microblaze/microblaze-protos.h
25167 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25168 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25169 * config/microblaze/microblaze.c (microblaze_address_type): Add
25170 ADDRESS_TLS and tls_reloc address types.
25171 (microblaze_address_info): Add tls_reloc.
25172 (TARGET_HAVE_TLS): Define.
25173 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25174 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25175 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25176 load_tls_operand, microblaze_call_tls_get_addr,
25177 microblaze_legitimize_tls_address): New functions.
25178 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25179 (get_base_reg): Use microblaze_tls_symbol_p.
25180 (microblaze_classify_address): Handle TLS.
25181 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25182 label_mentioned_p and microblaze_tls_referenced_p.
25183 (microblaze_legitimize_address): Handle TLS.
25184 (microblaze_address_insns): Handle ADDRESS_TLS.
25185 (pic_address_needs_scratch): Handle TLS.
25186 (print_operand_address): Handle TLS.
25187 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25188 (microblaze_expand_move): Handle TLS.
25189 (microblaze_legitimate_constant_p): Check
25190 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25191 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25192 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25193 (PIC_OFFSET_TABLE_REGNUM): Set.
25194 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25195 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25196 (addsi3, movsi_internal2, movdf_internal): Update constraints
25197 * config/microblaze/predicates.md (arith_plus_operand): Define
25198 (move_operand): Redefine as move_src_operand,
25199 check microblaze_tls_referenced_p.
25201 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25203 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25204 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25206 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25208 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25211 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25213 PR tree-optimization/53265
25214 * common.opt (Waggressive-loop-optimizations): New option.
25215 * tree-ssa-loop-niter.c: Include tree-pass.h.
25216 (do_warn_aggressive_loop_optimizations): New function.
25217 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
25218 if number_of_latch_executions returned constant.
25219 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
25220 early. If number_of_latch_executions returned constant, set
25221 nb_iterations_upper_bound back to it.
25222 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
25224 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
25225 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
25227 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
25228 (MULTILIB_OSDIRNAMES): Set.
25229 * genmultilib: If defaultosdirname doesn't start with :: , set
25230 defaultosdirname2 instead, clear it and emit two . multilib_raw
25231 entries instead of just one.
25233 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
25235 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25236 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
25237 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25238 (SUBTARGET_OVERRIDE_OPTIONS): New.
25240 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
25243 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
25244 (musermode): Convert to Var(TARGET_USERMODE).
25245 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
25246 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
25247 * config/sh/sh.c (sh_option_override): Use
25248 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
25249 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
25251 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
25253 (udivsi3_i4_single, divsi3_i4_single): Use
25254 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
25256 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
25258 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
25261 2013-03-13 Richard Biener <rguenther@suse.de>
25263 PR tree-optimization/56608
25264 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
25265 calls when vectorizing basic-blocks.
25267 2013-03-13 Jakub Jelinek <jakub@redhat.com>
25270 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
25273 2013-03-12 Jakub Jelinek <jakub@redhat.com>
25275 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
25277 2013-03-11 Jan Hubicka <jh@suse.cz>
25280 * lto-streamer-out.c (output_symbol_p): Skip references from
25281 constructors of external variables.
25283 2013-03-11 Jan Hubicka <jh@suse.cz>
25285 PR middle-end/56571
25286 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
25288 * emit-rtl.c (verify_rtx_sharing): Likewise.
25289 (copy_insn_1): Likewise.
25290 * rtl.c (copy_rtx): Likewise.
25292 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
25295 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
25296 output_operand_lossage message.
25298 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
25301 * arm.c (shift_op): Validate RTL pattern on the fly.
25302 (arm_print_operand, case 'S'): Don't use shift_operator to validate
25305 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25308 * config/pa/pa.md (call_value): Check for calls to powf and direct to
25309 new call patterns that clobber %fr12.
25310 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
25311 split and postreload patterns.
25312 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
25313 registers %fr12 and %fr12R as call used.
25315 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
25317 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
25318 (canon_address, record_store, replace_read, check_mem_read_rtx,
25319 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
25320 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
25321 rest_of_handle_dse): Likewise.
25323 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
25325 PR middle-end/56524
25326 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
25328 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
25329 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
25330 (save_optabs_if_changed): Replace with...
25331 (init_tree_optimization_optabs): ...this.
25332 * optabs.c (save_optabs_if_changed): Rename to...
25333 (init_tree_optimization_optabs): ...this. Take the optimization node
25334 as argument. Do nothing if the base optabs are already correct.
25335 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
25336 to recompute optabs.
25337 * function.h (function): Remove optabs field.
25338 * function.c (invoke_set_current_function_hook): Call
25339 init_tree_optimization_optabs. Use the result to initialize
25342 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
25344 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
25345 if GTMA_HAS_NO_INSTRUMENTATION.
25346 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
25347 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
25348 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
25349 * gimple-pretty-print.c (dump_gimple_transaction): Handle
25350 GTMA_HAS_NO_INSTRUMENTATION.
25352 2013-03-08 Jakub Jelinek <jakub@redhat.com>
25354 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
25357 2013-03-08 Marek Polacek <polacek@redhat.com>
25358 Jakub Jelinek <jakub@redhat.com>
25360 PR tree-optimization/56478
25361 * predict.c (is_comparison_with_loop_invariant_p): Change the
25362 type of loop_step to tree.
25363 (predict_loops): Adjust.
25364 (predict_iv_comparison): Perform the computations on double_ints.
25366 2013-03-08 Richard Biener <rguenther@suse.de>
25368 PR tree-optimization/56570
25369 * tree-cfg.c (verify_expr_location_1): Verify locations for
25371 * tree-sra.c (create_access_replacement): Strip locations
25372 from DECL_DEBUG_EXPRs.
25374 2013-03-08 Richard Biener <rguenther@suse.de>
25376 * tree-inline.c (expand_call_inline): Do not associate
25377 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
25378 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
25380 2013-03-08 Richard Biener <rguenther@suse.de>
25382 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
25383 or block changes with -Og. Fix for location / block encoding
25384 changes and PHI arguments with locations.
25386 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
25388 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
25390 (struct output_info): Likewise.
25391 (register_overhead): Remove bad gcc_assert.
25392 (bitmap_find_bit): If there is only a single bitmap element, do not
25393 count a miss as a search.
25394 (print_statistics): Update for counter type changes.
25395 (dump_bitmap_statistics): Likewise. Print headers such that they
25396 are properly lined up with the printed counters.
25398 2013-03-07 Jakub Jelinek <jakub@redhat.com>
25400 PR tree-optimization/56559
25401 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
25402 check that it has only a single use.
25404 2013-03-07 Richard Biener <rguenther@suse.de>
25406 * doc/invoke.texi (fwhole-program): Discourage use in combination
25409 2013-03-06 Jakub Jelinek <jakub@redhat.com>
25411 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
25413 PR tree-optimization/56539
25414 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
25415 instead of GSI_CONTINUE_LINKING as last argument to
25416 force_gimple_operand_gsi. Adjust function comment.
25418 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
25421 PR middle-end/56548
25422 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
25423 promoted mode, convert the result back to the original mode.
25425 2013-03-06 Richard Biener <rguenther@suse.de>
25427 PR middle-end/56294
25428 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
25429 (insert_updated_phi_nodes_compare_uids): New function.
25430 (update_ssa): Sort symbols_to_rename after UID before
25431 traversing it to insert PHI nodes.
25433 2013-03-06 Richard Biener <rguenther@suse.de>
25435 PR middle-end/50494
25436 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
25437 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
25440 2013-02-13 Richard Biener <rguenther@suse.de>
25443 * varasm.c (output_constant_def_1): Get the decl representing
25444 the constant as argument.
25445 (output_constant_def): Wrap output_constant_def_1.
25446 (make_decl_rtl): Use output_constant_def_1 with the decl
25447 representing the constant.
25448 (build_constant_desc): Optionally re-use a decl already
25449 representing the constant.
25450 (tree_output_constant_def): Adjust.
25452 2013-03-06 Joey Ye <joey.ye@arm.com>
25455 * gcc.c (convert_white_space): New function.
25456 (main): Handles white space in function name.
25458 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
25461 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
25462 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
25463 to SH_DIV_CALL_TABLE for TARGET_SH2.
25464 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
25466 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
25467 call-table options.
25469 2013-03-05 Sterling Augustine <saugustine@google.com>
25470 Cary Coutant <ccoutant@google.com>
25473 * dwarf2out.c (resolve_addr): Don't call
25474 remove_loc_list_addr_table_entries a second time for the same
25477 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25480 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
25481 (avoid_complex_debug_insns): New function.
25482 (expand_debug_locations): Call it.
25484 PR rtl-optimization/56484
25485 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
25486 lifetimes of hard registers on small register class machines.
25488 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
25490 * config/microblaze/microblaze-protos.h: Rename
25491 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
25492 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
25494 (microblaze_fast_interrupt_function_p): New function.
25495 (microblaze_is_interrupt_handler): Rename to
25496 microblaze_is_interrupt_variant and add fast_interrupt check.
25497 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
25498 (save_restore_insns): Likewise.
25499 (compute_frame_size): Likewise.
25500 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
25501 (microblaze_globalize_label): Likewise.
25502 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
25503 * config/microblaze/microblaze.md: Use wrapper
25504 microblaze_is_interrupt_variant.
25506 2013-03-05 Kai Tietz <ktietz@redhat.com>
25508 * sdbout.c (sdbout_one_type): Switch to current function's section
25509 supporting cold/hot.
25511 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
25513 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
25516 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25518 PR middle-end/56461
25519 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
25520 if VALGRIND_GET_VBITS is defined, temporarily make object
25521 memory all defined, and restore previous valgrind addressability
25522 and definability afterwards. Free this_object at the end.
25524 PR middle-end/56461
25525 * lra.c (lra): Call lra_clear_live_ranges if live_p,
25526 right before calling lra_create_live_ranges, also call it
25527 when clearing live_p. Only call lra_clear_live_ranges
25528 at the end if live_p.
25530 PR middle-end/56461
25531 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
25533 2013-03-05 Richard Biener <rguenther@suse.de>
25535 PR tree-optimization/56521
25536 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
25539 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
25542 * except.h (remove_unreachable_eh_regions): New prototype.
25543 * except.c (remove_eh_handler_splicer): New function, split out
25544 of remove_eh_handler.
25545 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
25546 warning about running it on many EH regions one at a time.
25547 (remove_unreachable_eh_regions_worker): New function, walk the
25548 EH tree in depth-first order and remove non-marked regions.
25549 (remove_unreachable_eh_regions): New function.
25550 * tree-eh.c (mark_reachable_handlers): New function, split out
25551 from remove_unreachable_handlers.
25552 (remove_unreachable_handlers): Use mark_reachable_handlers and
25553 remove_unreachable_eh_regions.
25554 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
25555 and remove_unreachable_eh_regions.
25557 2013-03-05 Richard Biener <rguenther@suse.de>
25559 PR middle-end/56525
25560 * loop-init.c (fix_loop_structure): Remove loops in two stages,
25561 not freeing them until the end.
25563 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25565 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
25567 2013-03-05 Richard Biener <rguenther@suse.de>
25569 PR tree-optimization/56270
25570 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
25571 of loads after scheduling an SLP instance.
25573 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25575 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
25577 (check_gcc_parallelize): Run guality.exp as a separate job from
25578 vect.exp with unsorted.exp and $(dg_target_exps) separately from
25579 struct-layout-1.exp with stackalign.exp.
25581 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
25583 PR middle-end/56461
25584 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
25585 load_index sbitmap even if some bit in it isn't set.
25587 PR middle-end/56461
25588 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
25589 (discover_iteration_bound_by_body_walk): Change queues to
25590 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
25591 spelling in comment. Call safe_push on queues[bound_index] directly.
25592 Release queues[queue_index] in every iteration unconditionally.
25593 Release bounds vector.
25595 PR middle-end/56461
25596 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
25597 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
25598 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
25601 2013-03-05 Richard Biener <rguenther@suse.de>
25604 * tree-inline.c (remap_blocks_to_null): New function.
25605 (expand_call_inline): When expanding a call stmt without
25606 an associated block inline remap all callee blocks to NULL.
25608 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25610 PR rtl-optimization/56494
25611 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
25612 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
25613 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
25615 PR middle-end/56461
25616 * sel-sched-ir.c (free_sched_pools): Release
25617 succs_info_pool.stack[succs_info_pool.max_top] vectors too
25618 if succs_info_pool.max_top isn't -1.
25621 * opts.c (opts_obstack, opts_concat): Moved to...
25622 * opts-common.c (opts_obstack, opts_concat): ... here.
25624 2013-03-04 Jakub Jelinek <jakub@redhat.com>
25626 PR middle-end/56461
25627 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
25629 2013-03-04 Martin Jambor <mjambor@suse.cz>
25631 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
25632 all appropriate places.
25634 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
25636 PR tree-optimization/56424
25637 * ipa-split.c (split_function): Do not set the RSO flag if result is
25638 not by reference and its type is a register type.
25640 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
25642 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
25643 (microblaze_legitimate_pic_operand): Likewise
25644 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
25645 new function microblaze_legitimate_pic_operand
25646 * config/microblaze/microblaze-protos.h
25647 (microblaze_legitimate_pic_operand): Declare.
25649 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25651 * config/microblaze/predicates.md (call_insn_simple_operand):
25652 New predicate for supported rtx code types.
25653 * config/microblaze/microblaze.md (call_internal1): Use
25654 call_insn_simple_operand predicate.
25656 2013-03-04 Jakub Jelinek <jakub@redhat.com>
25658 PR middle-end/56461
25659 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
25660 partitions.ordered_remove.
25662 PR middle-end/56461
25663 * tree-vect-stmts.c (vectorizable_conversion): Don't call
25664 vec_oprnds0.create (1) for modifier == NONE.
25666 PR middle-end/56461
25667 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
25668 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
25669 vec_oprnds1 right before pushing anything to it for
25672 PR middle-end/56461
25673 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
25674 set nbbs to 0 instead of having separate code path.
25675 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
25676 instead of false as last argument if returning NULL.
25678 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
25680 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
25681 the attribute is now called "target" instead of "option".
25682 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
25683 * doc/tm.texi.in (Target Attributes): Likewise document the correct
25684 attribute/pragma name for TARGET_OPTION_VALID_P and
25685 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
25686 * doc/tm.texi: Regenerated.
25688 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
25690 * config/microblaze/microblaze.c:
25691 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
25692 * config/microblaze/microblaze.h: Add -mxl-reorder to
25694 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
25695 instructions emitted if TARGET_REORDER.
25696 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
25697 or 0 for -m/-mno case, but initialises as 2 to detect default use case
25700 2013-03-01 Xinliang David Li <davidxl@google.com>
25702 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
25705 2013-03-01 Jakub Jelinek <jakub@redhat.com>
25707 PR middle-end/56461
25708 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
25709 vector even when returning true. Fix up function comment formatting.
25711 PR middle-end/56461
25712 * ira-build.c (ira_loop_nodes_count): New variable.
25713 (create_loop_tree_nodes): Initialize it.
25714 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
25716 PR middle-end/56461
25717 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
25718 method on dr_chain and result_chain.
25719 * tree-vect-stmts.c (vectorizable_store): Only call
25720 result_chain.create if j == 0.
25722 PR middle-end/56461
25723 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
25724 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
25725 before overwriting it.
25727 2013-03-01 Tobias Burnus <burnus@net-b.de>
25729 * doc/extended.texi (C Extensions): Change order in @menu
25731 (Other MIPS Built-in Functions): Move last MIPS entry before
25732 "picoChip Built-in Functions".
25733 (SH Built-in Functions): Move after RX Built-in Functions.
25734 * doc/gcc.texi (Introduction): Change order in @menu
25736 * doc/md.texi (Constraints): Ditto.
25737 * gty.texi (Type Information): Ditto.
25738 (User-provided marking routines for template types): Make
25740 * doc/invoke.texi (AArch64 Options): Move before
25741 "Adapteva Epiphany Options".
25743 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
25744 Jakub Jelinek <jakub@redhat.com>
25747 * asan.c (gate_asan): Lookup no_sanitize_address instead of
25748 no_address_safety_analysis attribute.
25749 * doc/extend.texi (no_address_safety_attribute): Rename to
25750 no_sanitize_address attribute, mention no_address_safety_analysis
25751 attribute as deprecated alias.
25753 2013-02-28 Jakub Jelinek <jakub@redhat.com>
25755 PR middle-end/56461
25756 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
25757 type to vec<vec<tree> > *.
25758 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
25759 to be vec<tree> instead of vec<tree> *, set vec_defs
25760 to vNULL and call vec_defs.create (number_of_vects), adjust other
25762 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
25763 vectorizable_condition): Adjust vect_get_slp_defs callers.
25765 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
25767 * config/aarch64/aarch64.c
25768 (aarch64_float_const_representable): Remove unused variable.
25770 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
25772 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
25774 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
25776 * config/aarch64/aarch64-builtins.c
25777 (aarch64_init_simd_builtins): Make static.
25779 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
25781 * config/aarch64/aarch64.c
25782 (aarch64_simd_make_constant): Make static.
25784 2013-02-28 Martin Jambor <mjambor@suse.cz>
25786 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
25787 with no initialization to the RHS of debug statements.
25789 2013-02-28 Martin Jambor <mjambor@suse.cz>
25791 PR tree-optimization/56294
25792 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
25794 (get_access_replacement): Do not call create_access_replacement.
25795 Assert a replacement exists.
25796 (get_repl_default_def_ssa_name): Create the replacement declaration
25799 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25801 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
25802 final_end_function.
25804 2013-02-28 Marek Polacek <polacek@redhat.com>
25806 PR rtl-optimization/56466
25807 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
25808 if we're changing a loop.
25809 (peel_loops_completely): Likewise.
25811 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
25814 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
25816 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
25819 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
25820 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
25821 INTX_FTYPE_FX, FX_FTYPE_INTX.
25822 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
25824 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
25826 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
25827 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
25828 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
25829 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
25830 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
25831 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
25832 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
25833 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
25834 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
25835 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
25836 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
25837 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
25838 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
25839 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
25840 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
25841 (avrxmega6): Increase max flash segments from 5 to 6.
25842 * config/avr/t-multilib: Regenerate.
25843 * config/avr/avr-tables.opt: Regenerate.
25844 * doc/avr-mmcu.texi: Regenerate.
25846 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
25848 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
25849 (avr_device_to_arch): Rename to avr_device_to_ld.
25850 (avr_device_to_as): New prototype.
25851 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
25852 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
25853 * config/avr/driver-avr.c (avr_device_to_as): New.
25854 (avr_device_to_arch): Rename to avr_device_to_ld.
25856 2013-02-27 Jakub Jelinek <jakub@redhat.com>
25858 PR middle-end/56461
25859 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
25860 method on dr_chain and result_chain.
25862 PR middle-end/56461
25863 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
25864 pointer_set_destroy on not_executed_last_iteration.
25866 PR middle-end/56461
25867 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
25869 PR middle-end/56461
25870 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
25871 FOR_EACH_DEFINED_FUNCTION when freeing state.
25873 PR middle-end/56461
25874 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
25876 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
25879 PR middle-end/56461
25880 * ipa-cp.c (decide_whether_version_node): Call vec_free on
25881 known_aggs[i].items and release known_aggs vector.
25883 PR middle-end/56461
25884 * ipa-reference.c (propagate): Free node_info even for alias nodes.
25886 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25888 * config/microblaze/microblaze.c (microblaze_emit_compare):
25889 Use xor for EQ/NE comparisions.
25890 * config/microblaze/microblaze.md (cstoresf4): Add constraints
25891 (cbranchsf4): Adjust operator to comparison_operator.
25893 2013-02-27 Jakub Jelinek <jakub@redhat.com>
25895 PR middle-end/56461
25896 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
25898 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
25899 vec_safe_push, always update *slot.
25900 (redirect_edge_var_map_clear): Use vec_free.
25901 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
25902 (free_var_map_entry): Use vec_free.
25903 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
25904 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
25906 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
25908 PR middle-end/45472
25909 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
25910 when the may_trap_p bit of the exprs being merged differs.
25911 Reorder tests for speculativeness in the logical and operator.
25913 2013-02-27 Jakub Jelinek <jakub@redhat.com>
25915 * incpath.c (add_standard_paths): Use reconcat instead of concat
25916 where appropriate and avoid leaking memory.
25918 * opts.h: Include obstack.h.
25919 (opts_concat): New prototype.
25920 (opts_obstack): New declaration.
25921 * opts.c (opts_concat): New function.
25922 (opts_obstack): New variable.
25923 (init_options_struct): Call gcc_init_obstack on opts_obstack.
25924 (finish_options): Use opts_concat instead of concat
25925 and XOBNEWVEC instead of XNEWVEC.
25926 * opts-common.c (generate_canonical_option, decode_cmdline_option,
25927 generate_option): Likewise.
25928 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
25929 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
25932 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
25933 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
25935 2013-02-26 Jakub Jelinek <jakub@redhat.com>
25937 PR middle-end/56461
25938 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
25940 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
25942 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
25943 (arm_block_move_unaligned_straight): Likewise.
25944 (arm_adjust_block_mem): Likewise.
25946 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
25949 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
25950 temp, cond and label.
25951 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
25954 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
25955 * config/c6x/c6x.h (dbx_register_map): Update declaration.
25958 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
25959 of prologue/epilogue functions.
25962 * config/tilegx/tilegx.c (tilegx_expand_prologue):
25963 Remove unused variable cfa_offset.
25964 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
25967 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
25968 type promotion to unsigned.
25971 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
25972 for HOST_WIDE_INT of 32 bit / same size as int.
25973 (arm_block_move_unaligned_straight): Likewise.
25974 (arm_adjust_block_mem): Likewise.
25977 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
25980 2013-02-26 Marek Polacek <polacek@redhat.com>
25982 PR tree-optimization/56426
25983 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
25985 2013-02-26 Richard Biener <rguenther@suse.de>
25988 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
25989 unused variable loops.
25991 2013-02-26 Jakub Jelinek <jakub@redhat.com>
25993 PR tree-optimization/56448
25994 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
25995 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
25996 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
25997 later operands of the references, or even first operand for
25998 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26000 PR tree-optimization/56443
26001 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26002 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26003 to type_for_mode langhook.
26005 2013-02-25 Matt Turner <mattst88@gmail.com>
26007 * doc/invoke.texi: Document r4700.
26009 2013-02-25 Richard Biener <rguenther@suse.de>
26011 PR tree-optimization/56175
26012 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26014 (simplify_bitwise_binary): ... here. Also guard the conversion
26015 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26017 2013-02-25 Catherine Moore <clm@codesourcery.com>
26020 2013-02-24 Catherine Moore <clm@codesourcery.com>
26021 Maciej W. Rozycki <macro@codesourcery.com>
26022 Tom de Vries <tom@codesourcery.com>
26023 Nathan Sidwell <nathan@codesourcery.com>
26024 Iain Sandoe <iain@codesourcery.com>
26025 Nathan Froyd <froydnj@codesourcery.com>
26026 Chao-ying Fu <fu@mips.com>
26028 * doc/extend.texi: (micromips, nomicromips, nocompression):
26029 Document new function attributes.
26030 * doc/invoke.texi (minterlink-compressed, mmicromips,
26031 m14k, m14ke, m14kec): Document new options.
26032 (minterlink-mips16): Update documentation.
26033 * doc/md.texi (ZC, ZD): Document new constraints.
26034 * configure.ac (gcc_cv_as_micromips): Check if linker
26035 supports the .set micromips directive.
26036 * configure: Regenerate.
26037 * config.in: Regenerate.
26038 * config/mips/mips-tables.opt: Regenerate.
26039 * config/mips/micromips.md: New file.
26040 * constraints.md (ZC, AD): New constraints.
26041 * config/mips/predicates.md (movep_src_register): New predicate.
26042 (movep_src_operand): New predicate.
26043 (non_volatile_mem_operand): New predicate.
26044 * config/mips/mips.md (multimem): New type.
26045 (length): Differentiate between 17-bit and 18-bit branch offsets.
26046 (MOVEP1, MOVEP2): New mode iterator.
26047 (mov_<load>l): Use ZC constraint.
26048 (mov_<load>r): Likewise.
26049 (mov_<store>l): Likewise.
26050 (mov_<store>r): Likewise.
26051 (*branch_equality<mode>_inverted): Add microMIPS support.
26052 (*branch_equality<mode>): Likewise.
26053 (*jump_absolute): Likewise.
26054 (indirect_jump_<mode>): Likewise.
26055 (tablejump_<mode>): Likewise.
26056 (<optab>_internal): Likewise.
26057 (sibcall_internal): Likewise.
26058 (sibcall_value_internal): Likewise.
26059 (prefetch): Use constraint ZD.
26060 * config/mips/mips.opt (minterlink-compressed): New option.
26061 (minterlink-mips16): Now an alias for minterlink-compressed.
26062 (mmicromips): New option.
26063 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26064 (compare_and_swap_12): Likewise.
26065 (sync_add<mode>): Likewise.
26066 (sync_<optab>_12): Likewise.
26067 (sync_old_<optab>_12): Likewise.
26068 (sync_new_<optab>_12): Likewise.
26069 (sync_nand_12): Likewise.
26070 (sync_old_nand_12): Likewise.
26071 (sync_new_nand_12): Likewise.
26072 (sync_sub<mode>): Likewise.
26073 (sync_old_add<mode>): Likewise.
26074 (sync_old_sub<mode>): Likewise.
26075 (sync_new_add<mode>): Likewise.
26076 (sync_new_sub<mode>): Likewise.
26077 (sync_<optab><mode>): Likewise.
26078 (sync_old_<optab><mode>): Likewise.
26079 (sync_new_<optab><mode>): Likewise.
26080 (sync_nand<mode>): Likewise.
26081 (sync_old_nand<mode>): Likewise.
26082 (sync_new_nand<mode>): Likewise.
26083 (sync_lock_test_and_set<mode>): Likewise.
26084 (test_and_set_12): Likewise.
26085 (atomic_compare_and_swap<mode>): Likewise.
26086 (atomic_exchange<mode>_llsc): Likewise.
26087 (atomic_fetch_add<mode>_llsc): Likewise.
26088 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26089 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26090 (umips_save_restore_pattern_p): Likewise.
26091 (umips_load_store_pair_p): Likewise.
26092 (umips_output_load_store_pair): Likewise.
26093 (umips_movep_target_p): Likewise.
26094 (umips_12bit_offset_address_p): Likewise.
26095 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26096 (mips_base_mips16): Rename this...
26097 (mips_base_compression_flags): ...to this. Update all uses.
26098 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26099 (mips_mips16_decl_p): Delete.
26100 (mips_nomips16_decl_p): Delete.
26101 (mips_get_compress_on_flags): New function.
26102 (mips_get_compress_off_flags): New function.
26103 (mips_get_compress_mode): New function.
26104 (mips_get_compress_on_name): New function.
26105 (mips_get_compress_off_name): New function.
26106 (mips_insert_attributes): Support multiple compression types.
26107 (mips_merge_decl_attributes): Likewise.
26108 (umips_12bit_offset_address_p): New function.
26109 (mips_start_function_definition): Emit .set micromips directive.
26110 (mips_call_may_need_jalx_p): New function.
26111 (mips_function_ok_for_sibcall): Add microMIPS support.
26112 (mips_print_operand_punctuation): Support short delay slots and
26114 (umips_swm_mask, umips_swm_encoding): New.
26115 (umips_build_save_restore): New function.
26116 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26117 (was_mips16_p): Remove.
26118 (old_compression_mode): New.
26119 (mips_set_compression_mode): New function.
26120 (mips_set_current_function): Add microMIPS support.
26121 (mips_option_override): Likewise.
26122 (umips_save_restore_pattern_p): New function.
26123 (umips_output_save_restore): New function.
26124 (umips_load_store_pair_p_1): New function.
26125 (umips_load_store_pair_p): New function.
26126 (umips_output_load_store_pair_1): New function.
26127 (umips_output_load_store_pair): New function.
26128 (umips_movep_target_p) New function.
26129 (mips_prepare_pch_save): Add microMIPS support.
26130 * config/mips/mips.h (TARGET_COMPRESSION): New.
26131 (TARGET_CPU_CPP_BUILTINS): Update macro
26132 to use new compression flags and to support microMIPS.
26133 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26134 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26135 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26136 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26137 (ASM_SPEC): Support mmicromips and mno-micromips.
26138 (M16STORE_REG_P): New macro.
26139 (MIPS_CALL): Support TARGET_MICROMIPS.
26140 (MICROMIPS_J): New macro.
26141 (mips_base_mips16): Rename this...
26142 (mips_base_compression_flags): ...to this.
26143 (UMIPS_12BIT_OFFSET_P): New macro.
26144 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26145 (MULTILIB_DIRNAMES): Likewise.
26147 2013-02-25 Tom de Vries <tom@codesourcery.com>
26149 PR rtl-optimization/56131
26150 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26151 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26152 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26154 2013-02-25 Tobias Burnus <burnus@net-b.de>
26156 * doc/invoke.texi (-fsanitize=): Move from optimization
26157 to debugging options.
26159 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26161 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26163 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26164 Alexander Monakov <amonakov@ispras.ru>
26166 PR middle-end/56077
26167 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26168 flush pending lists also on non-jumps. Adjust comment.
26170 2013-02-24 Catherine Moore <clm@codesourcery.com>
26171 Maciej W. Rozycki <macro@codesourcery.com>
26172 Tom de Vries <tom@codesourcery.com>
26173 Nathan Sidwell <nathan@codesourcery.com>
26174 Iain Sandoe <iain@codesourcery.com>
26175 Nathan Froyd <froydnj@codesourcery.com>
26176 Chao-ying Fu <fu@mips.com>
26178 * doc/extend.texi: (micromips, nomicromips, nocompression):
26179 Document new function attributes.
26180 * doc/invoke.texi (minterlink-compressed, mmicromips,
26181 m14k, m14ke, m14kec): Document new options.
26182 (minterlink-mips16): Update documentation.
26183 * doc/md.texi (ZC, ZD): Document new constraints.
26184 * configure.ac (gcc_cv_as_micromips): Check if linker
26185 supports the .set micromips directive.
26186 * configure: Regenerate.
26187 * config.in: Regenerate.
26188 * config/mips/mips-tables.opt: Regenerate.
26189 * config/mips/micromips.md: New file.
26190 * constraints.md (ZC, AD): New constraints.
26191 * config/mips/predicates.md (movep_src_register): New predicate.
26192 (movep_src_operand): New predicate.
26193 (non_volatile_mem_operand): New predicate.
26194 * config/mips/mips.md (multimem): New type.
26195 (length): Differentiate between 17-bit and 18-bit branch offsets.
26196 (MOVEP1, MOVEP2): New mode iterator.
26197 (mov_<load>l): Use ZC constraint.
26198 (mov_<load>r): Likewise.
26199 (mov_<store>l): Likewise.
26200 (mov_<store>r): Likewise.
26201 (*branch_equality<mode>_inverted): Add microMIPS support.
26202 (*branch_equality<mode>): Likewise.
26203 (*jump_absolute): Likewise.
26204 (indirect_jump_<mode>): Likewise.
26205 (tablejump_<mode>): Likewise.
26206 (<optab>_internal): Likewise.
26207 (sibcall_internal): Likewise.
26208 (sibcall_value_internal): Likewise.
26209 (prefetch): Use constraint ZD.
26210 * config/mips/mips.opt (minterlink-compressed): New option.
26211 (minterlink-mips16): Now an alias for minterlink-compressed.
26212 (mmicromips): New option.
26213 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26214 (compare_and_swap_12): Likewise.
26215 (sync_add<mode>): Likewise.
26216 (sync_<optab>_12): Likewise.
26217 (sync_old_<optab>_12): Likewise.
26218 (sync_new_<optab>_12): Likewise.
26219 (sync_nand_12): Likewise.
26220 (sync_old_nand_12): Likewise.
26221 (sync_new_nand_12): Likewise.
26222 (sync_sub<mode>): Likewise.
26223 (sync_old_add<mode>): Likewise.
26224 (sync_old_sub<mode>): Likewise.
26225 (sync_new_add<mode>): Likewise.
26226 (sync_new_sub<mode>): Likewise.
26227 (sync_<optab><mode>): Likewise.
26228 (sync_old_<optab><mode>): Likewise.
26229 (sync_new_<optab><mode>): Likewise.
26230 (sync_nand<mode>): Likewise.
26231 (sync_old_nand<mode>): Likewise.
26232 (sync_new_nand<mode>): Likewise.
26233 (sync_lock_test_and_set<mode>): Likewise.
26234 (test_and_set_12): Likewise.
26235 (atomic_compare_and_swap<mode>): Likewise.
26236 (atomic_exchange<mode>_llsc): Likewise.
26237 (atomic_fetch_add<mode>_llsc): Likewise.
26238 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26239 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26240 (umips_save_restore_pattern_p): Likewise.
26241 (umips_load_store_pair_p): Likewise.
26242 (umips_output_load_store_pair): Likewise.
26243 (umips_movep_target_p): Likewise.
26244 (umips_12bit_offset_address_p): Likewise.
26245 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26246 (mips_base_mips16): Rename this...
26247 (mips_base_compression_flags): ...to this. Update all uses.
26248 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26249 (mips_mips16_decl_p): Delete.
26250 (mips_nomips16_decl_p): Delete.
26251 (mips_get_compress_on_flags): New function.
26252 (mips_get_compress_off_flags): New function.
26253 (mips_get_compress_mode): New function.
26254 (mips_get_compress_on_name): New function.
26255 (mips_get_compress_off_name): New function.
26256 (mips_insert_attributes): Support multiple compression types.
26257 (mips_merge_decl_attributes): Likewise.
26258 (umips_12bit_offset_address_p): New function.
26259 (mips_start_function_definition): Emit .set micromips directive.
26260 (mips_call_may_need_jalx_p): New function.
26261 (mips_function_ok_for_sibcall): Add microMIPS support.
26262 (mips_print_operand_punctuation): Support short delay slots and
26264 (umips_swm_mask, umips_swm_encoding): New.
26265 (umips_build_save_restore): New function.
26266 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26267 (was_mips16_p): Remove.
26268 (old_compression_mode): New.
26269 (mips_set_compression_mode): New function.
26270 (mips_set_current_function): Add microMIPS support.
26271 (mips_option_override): Likewise.
26272 (umips_save_restore_pattern_p): New function.
26273 (umips_output_save_restore): New function.
26274 (umips_load_store_pair_p_1): New function.
26275 (umips_load_store_pair_p): New function.
26276 (umips_output_load_store_pair_1): New function.
26277 (umips_output_load_store_pair): New function.
26278 (umips_movep_target_p) New function.
26279 (mips_prepare_pch_save): Add microMIPS support.
26280 * config/mips/mips.h (TARGET_COMPRESSION): New.
26281 (TARGET_CPU_CPP_BUILTINS): Update macro
26282 to use new compression flags and to support microMIPS.
26283 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26284 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26285 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26286 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26287 (ASM_SPEC): Support mmicromips and mno-micromips.
26288 (M16STORE_REG_P): New macro.
26289 (MIPS_CALL): Support TARGET_MICROMIPS.
26290 (MICROMIPS_J): New macro.
26291 (mips_base_mips16): Rename this...
26292 (mips_base_compression_flags): ...to this.
26293 (UMIPS_12BIT_OFFSET_P): New macro.
26294 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26295 (MULTILIB_DIRNAMES): Likewise.
26297 2013-02-24 Jakub Jelinek <jakub@redhat.com>
26300 * target-globals.c (save_target_globals): For init_reg_sets and
26301 target_reinit remporarily set this_fn_optabs to this_target_optabs.
26303 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
26305 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
26306 * config/aarch64/t-aarch64
26307 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
26309 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
26311 PR inline-asm/56148
26312 * lra-constraints.c (process_alt_operands): Reload operand
26313 conflicting with earlier clobber only if no more other conflicting
26316 2013-02-22 Jakub Jelinek <jakub@redhat.com>
26319 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
26320 if not linking a shared library.
26322 2013-02-22 Seth LaForge <sethml@google.com>
26324 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
26326 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
26328 * config/arm/arm.md (split for extendsidi): Update condition.
26329 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
26330 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
26331 (qhs_zextenddi_cstr): Likewise.
26333 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26335 PR middle-end/56420
26336 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
26337 avoid signed wrapping.
26338 (expand_mult): Handle properly multiplication by
26339 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
26340 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
26341 in the compiler if coeff is HOST_WIDE_INT_MIN.
26342 (expand_divmod): Don't make ext_op1 static, change it's type to
26343 uhwi. Avoid undefined behavior in -INTVAL (op1).
26345 PR rtl-optimization/50339
26346 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
26348 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
26349 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
26350 into splitting_ashiftrt field.
26351 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
26353 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
26356 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
26358 PR middle-end/56108
26359 * trans-mem.c (execute_tm_mark): Do not expand transactions that
26360 are sure to go irrevocable.
26362 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
26364 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
26365 scalars are valid operands.
26367 2013-02-21 Martin Jambor <mjambor@suse.cz>
26369 PR tree-optimization/56310
26370 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
26371 only matching indices and non-negative final offsets.
26372 (intersect_aggregates_with_edge): Pass src_idx to
26373 agg_replacements_to_vector. Pass src_idx insstead of index to
26374 intersect_with_agg_replacements.
26376 2013-02-21 Martin Jambor <mjambor@suse.cz>
26378 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
26379 instead of hard-wired defaults.
26381 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
26383 * doc/invoke.texi (MIPS Options): Update documentation of the
26384 floating-point multiply-accumulate instruction restrictions.
26386 2013-02-21 Kostya Serebryany <kcc@google.com>
26388 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
26389 asan_shadow_offset on x86_64 linux.
26391 2013-02-21 Richard Biener <rguenther@suse.de>
26393 PR tree-optimization/56415
26395 2013-02-11 Richard Biener <rguenther@suse.de>
26397 PR tree-optimization/56273
26398 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
26401 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26404 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
26407 PR inline-asm/56405
26408 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
26409 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
26411 2013-02-20 Jan Hubicka <jh@suse.cz>
26413 PR tree-optimization/56265
26414 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
26415 when target is referenced for first time.
26417 2013-02-20 Richard Biener <rguenther@suse.de>
26419 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
26420 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
26421 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
26422 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
26423 not return anything.
26424 (rename_ssa_copies): Do not remove unused locals.
26425 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
26426 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
26427 * passes.c (execute_function_todo): Do not schedule unused locals
26428 removal if cleanup_tree_cfg did something.
26429 * tree-ssa-live.c (remove_unused_locals): Dump statistics
26430 about the number of removed locals.
26432 2013-02-20 Richard Biener <rguenther@suse.de>
26434 PR tree-optimization/56398
26435 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
26437 2013-02-20 Martin Jambor <mjambor@suse.cz>
26439 PR tree-optimization/55334
26440 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
26441 restricted pointers to arrays.
26443 2013-02-20 Richard Biener <rguenther@suse.de>
26444 Jakub Jelinek <jakub@redhat.com>
26446 PR tree-optimization/56396
26447 * tree-ssa-ccp.c (n_const_val): New static variable.
26448 (get_value): Return NULL for SSA names we don't have a lattice
26450 (ccp_initialize): Initialize n_const_val.
26451 * tree-ssa-copy.c (n_copy_of): New static variable.
26452 (init_copy_prop): Initialize n_copy_of.
26453 (get_value): Return NULL_TREE for SSA names we don't have a
26456 2013-02-20 Martin Jambor <mjambor@suse.cz>
26458 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
26460 2013-02-20 Richard Biener <rguenther@suse.de>
26462 * genpreds.c (write_lookup_constraint): Do not compare first
26463 letter of the constraint again.
26465 2013-02-20 Richard Biener <rguenther@suse.de>
26467 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
26469 (get_use_iv_cost): Terminate hashtable walk when coming across
26472 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
26474 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
26475 reassociation for avx2 targets.
26477 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26479 * config/microblaze/microblaze.c: microblaze_has_clz = 0
26480 Add version check for v8.10.a to enable microblaze_has_clz
26481 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
26482 version and TARGET_PATTERN_COMPARE check
26483 * config/microblaze/microblaze.md: New clzsi2 instruction
26485 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26487 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
26488 function before branching.
26490 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
26492 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
26494 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
26496 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
26498 PR middle-end/55889
26499 * sel-sched.c: Include ira.h.
26500 (implicit_clobber_conflict_p): New function.
26501 (moveup_expr): Use it.
26502 * Makefile.in (sel-sched.o): Depend on ira.h.
26504 2013-02-19 Richard Biener <rguenther@suse.de>
26506 PR tree-optimization/56384
26507 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
26508 (vn_hash_type): Split out from ...
26509 (vn_hash_constant_with_type): ... here.
26510 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
26511 (vn_phi_eq): Compare types from vn_phi_s structure.
26512 (vn_phi_lookup): Populate vn_phi_s type.
26513 (vn_phi_insert): Likewise.
26515 2013-02-19 Jakub Jelinek <jakub@redhat.com>
26517 PR tree-optimization/56350
26518 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
26519 if haven't found reduction or nested cycle operand, rather than
26520 asserting we must find it.
26522 PR tree-optimization/56381
26523 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
26526 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
26527 Jakub Jelinek <jakub@redhat.com>
26530 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
26531 (swap_optab_enable): Same.
26532 (init_all_optabs): Use argument instead of global.
26533 * tree.h (struct tree_optimization_option): New field target_optabs.
26534 * expr.h (init_all_optabs): Add argument to prototype.
26535 (TREE_OPTIMIZATION_OPTABS): New.
26536 (save_optabs_if_changed): Protoize.
26537 * optabs.h: Declare this_fn_optabs.
26538 * optabs.c (save_optabs_if_changed): New.
26539 Declare this_fn_optabs.
26540 (init_optabs): Add argument to init_all_optabs() call.
26541 * function.c (invoke_set_current_function_hook): Handle per
26543 * function.h (struct function): New field optabs.
26544 * config/mips/mips.c (mips_set_mips16_mode): Handle when
26545 optimization_current_node has changed.
26546 * target-globals.h (save_target_globals_default_opts): Protoize.
26547 * target-globals.c (save_target_globals_default_opts): New.
26549 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26552 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
26553 registers %fr12 and %fr12R as call used.
26556 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
26557 and HImode, require all displacements to be an integer multiple of
26559 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
26560 only allow QImode and HImode when reload is in progress and strict is
26561 true. Likewise for symbolic addresses. Use base14_operand to check
26562 displacements in REG+BASE addresses.
26564 2013-02-18 Richard Biener <rguenther@suse.de>
26566 PR tree-optimization/56366
26567 * tree-vect-loop.c (get_initial_def_for_induction): Properly
26568 handle sign-conversion of outer-loop initial induction value.
26570 2013-02-18 Richard Biener <rguenther@suse.de>
26572 PR middle-end/56349
26573 * cfghooks.c (merge_blocks): If we merge a latch into another
26574 block adjust references to it.
26575 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
26576 (verify_loop_structure): Verify that a recorded latch is in fact
26579 2013-02-18 Richard Biener <rguenther@suse.de>
26581 PR tree-optimization/56321
26582 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
26583 order SSA name release and virtual operand unlinking.
26585 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26587 * config/microblaze/microblaze.md (save_stack_block): Define.
26588 (restore_stack_block): Likewise.
26590 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26592 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
26593 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
26594 * config/microblaze/microblaze.c (microblaze_option_override):
26595 Bail out early for PIC modes when target does not support PIC.
26597 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26599 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
26600 Replace with a microblaze version.
26601 (microblaze_trampoline_init): Adapt for microblaze.
26602 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
26605 2013-02-16 Jakub Jelinek <jakub@redhat.com>
26606 Dodji Seketeli <dodji@redhat.com>
26609 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
26610 (instrument_mem_region_access): Do not forget to always put
26611 instrumentation of the of 'base' and 'base + len' in a "if (len !=
26612 0) statement, even for cases where either 'base' or 'base + len'
26613 are not instrumented -- because they have been previously
26614 instrumented. Simplify the logic by putting all the statements
26615 instrument 'base + len' inside a sequence, and then insert that
26616 sequence right before the current insertion point. Then, to
26617 instrument 'base + len', just get an iterator on that statement.
26618 And do not forget to update the pointer to iterator the function
26619 received as argument.
26621 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
26623 PR rtl-optimization/56348
26624 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
26626 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
26628 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
26629 (clean_graph_dump_file): Pass base to start_graph_dump.
26631 2013-02-14 Richard Henderson <rth@redhat.com>
26634 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
26636 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
26638 * collect2-aix.h: Define F_LOADONLY.
26640 2013-02-14 Richard Biener <rguenther@suse.de>
26643 * varasm.c (output_constant_def_1): Get the decl representing
26644 the constant as argument.
26645 (output_constant_def): Wrap output_constant_def_1.
26646 (make_decl_rtl): Use output_constant_def_1 with the decl
26647 representing the constant.
26648 (build_constant_desc): Optionally re-use a decl already
26649 representing the constant.
26650 (tree_output_constant_def): Adjust.
26652 2013-02-14 Dodji Seketeli <dodji@redhat.com>
26655 * asan.c (instrument_builtin_call): Really put the length of the
26656 second source argument into src1_len.
26658 2013-02-13 Jakub Jelinek <jakub@redhat.com>
26660 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
26661 argument. If it is false, don't create edge from then_bb to
26663 (insert_if_then_before_iter): Pass true to it.
26664 (build_check_stmt): Pass false to it.
26665 (transform_statements): Flush hash table only on extended basic
26666 block boundaries, rather than at the beginning of every bb.
26667 Don't flush hash table on nonfreeing_call_p calls.
26668 * tree-flow.h (nonfreeing_call_p): New prototype.
26669 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
26671 2013-02-13 David S. Miller <davem@davemloft.net>
26673 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
26675 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
26678 * ira.c (max_regno_before_ira): Move from ...
26680 (fix_reg_equiv_init): Use max_regno_before_ira instead of
26683 2013-02-13 Jakub Jelinek <jakub@redhat.com>
26685 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
26687 2013-02-13 Richard Biener <rguenther@suse.de>
26690 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
26691 globals in MEM_REFs.
26693 2013-02-13 Richard Biener <rguenther@suse.de>
26695 * loop-init.c (loop_optimizer_init): Clear loop state when
26696 re-initializing preserved loops.
26697 * loop-unswitch.c (unswitch_single_loop): Return whether
26698 we unswitched the loop. Do not verify loop state here.
26699 (unswitch_loops): When we unswitched a loop discover new loops.
26701 2013-02-13 Kostya Serebryany <kcc@google.com>
26703 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
26705 * sanitizer.def: Rename __asan_init to __asan_init_v1.
26707 2013-02-12 Dodji Seketeli <dodji@redhat.com>
26709 Avoid instrumenting duplicated memory access in the same basic block
26710 * Makefile.in (asan.o): Add new dependency on hash-table.h
26711 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
26712 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
26713 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
26714 (free_mem_ref_resources, has_mem_ref_been_instrumented)
26715 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
26716 (get_mem_ref_of_assignment): New functions.
26717 (get_mem_refs_of_builtin_call): Extract from
26718 instrument_builtin_call and tweak a little bit to make it fit with
26720 (instrument_builtin_call): Use the new
26721 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
26722 of is_gimple_builtin_call.
26723 (instrument_derefs, instrument_mem_region_access): Insert the
26724 instrumented memory reference into the hash table.
26725 (maybe_instrument_assignment): Renamed instrument_assignment into
26726 this, and change it to advance the iterator when instrumentation
26727 actually happened and return true in that case. This makes it
26728 homogeneous with maybe_instrument_assignment, and thus give a
26729 chance to callers to be more 'regular'.
26730 (transform_statements): Clear the memory reference hash table
26731 whenever we enter a new BB, when we cross a function call, or when
26732 we are done transforming statements. Use
26733 maybe_instrument_assignment instead of instrumentation. No more
26734 need to special case maybe_instrument_assignment and advance the
26735 iterator after calling it; it's now handled just like
26736 maybe_instrument_call. Update comment.
26738 2013-02-13 Richard Biener <rguenther@suse.de>
26740 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
26741 Fix loop discovery code.
26743 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
26745 PR inline-asm/56148
26746 * lra-constraints.c (process_alt_operands): Match early clobber
26747 operand with itself. Check conflicts with earlyclobber only if
26748 the operand is not reloaded. Prefer to reload conflicting operand
26749 if earlyclobber and matching operands are the same.
26751 2013-02-12 Richard Biener <rguenther@suse.de>
26754 * lto-streamer-out.c (write_symbol): Do not output symbols
26755 for hard register variables.
26757 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
26760 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
26761 (umulsidi3_insn, mulsidi3_insn): New insns.
26763 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
26765 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
26766 (struct tune_params): Add vec_costs field.
26767 * config/arm/arm.c (arm_builtin_vectorization_cost)
26768 (arm_add_stmt_cost): New functions.
26769 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
26770 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
26771 (arm_default_vec_cost): New struct of type cpu_vec_costs.
26772 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
26773 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
26774 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
26775 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
26777 2013-02-12 Richard Biener <rguenther@suse.de>
26780 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
26781 decls again if possible.
26783 2013-02-12 Richard Biener <rguenther@suse.de>
26785 PR middle-end/56288
26786 * tree-ssa.c (verify_ssa_name): Fix check, move
26787 SSA_NAME_IN_FREE_LIST check up.
26789 2013-02-12 Jakub Jelinek <jakub@redhat.com>
26790 Steven Bosscher <steven@gcc.gnu.org>
26792 PR rtl-optimization/56151
26793 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
26794 equal to op0 or op1, and last_insn pattern is CODE operation
26795 with MEM dest and one of the operands matches that MEM.
26797 2013-02-11 Sriraman Tallam <tmsriram@google.com>
26799 * doc/extend.texi: Document Function Multiversioning and "default"
26800 parameter string to target attribute.
26801 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
26802 target attribute parameter is "default".
26803 (ix86_compare_version_priority): Remove checks for target attribute.
26804 (ix86_mangle_function_version_assembler_name): Change error to sorry.
26805 Remove check for target attribute equal to NULL. Add assert.
26806 (ix86_generate_version_dispatcher_body): Change error to sorry.
26808 2013-02-11 Iain Sandoe <iain@codesourcery.com>
26809 Jack Howarth <howarth@bromo.med.uc.edu>
26810 Patrick Marlier <patrick.marlier@gmail.com>
26813 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
26814 define ENDFILE_SPEC as TM_DESTRUCTOR.
26815 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
26817 2013-02-11 Alexander Potapenko <glider@google.com>
26818 Jack Howarth <howarth@bromo.med.uc.edu>
26819 Jakub Jelinek <jakub@redhat.com>
26822 * config/darwin.c (cdtor_record): Rename ctor_record.
26823 (sort_cdtor_records): Rename sort_ctor_records.
26824 (finalize_dtors): New routine to sort destructors by
26825 priority before use in assemble_integer.
26826 (machopic_asm_out_destructor): Use finalize_dtors if needed.
26828 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
26830 PR rtl-optimization/56275
26831 * simplify-rtx.c (avoid_constant_pool_reference): Check that
26832 offset is non-negative and less than cmode size before
26833 calling simplify_subreg.
26835 2013-02-11 Richard Biener <rguenther@suse.de>
26837 PR tree-optimization/56264
26838 * cfgloop.h (fix_loop_structure): Adjust prototype.
26839 * loop-init.c (fix_loop_structure): Return the number of
26840 newly discovered loops.
26841 * tree-cfgcleanup.c (repair_loop_structures): When new loops
26842 are discovered, do a full loop-closed SSA rewrite.
26844 2013-02-11 Richard Biener <rguenther@suse.de>
26846 PR tree-optimization/56273
26847 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
26849 (check_array_ref): Fix missing newline in dumps.
26850 (search_for_addr_array): Likewise.
26852 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
26854 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
26856 2013-02-09 Jakub Jelinek <jakub@redhat.com>
26859 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
26861 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
26863 PR rtl-optimization/56246
26864 * lra-constraints.c (simplify_operand_subreg): Try to reuse
26866 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
26867 constraints are satisfied.
26869 2013-02-08 Jeff Law <law@redhat.com>
26872 * emit-rtl.c (reg_is_parm_p): New function.
26873 * regs.h (reg_is_parm_p): New prototype.
26874 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
26875 callee-clobbered registers.
26877 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
26880 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
26881 If there is no implicit builtin declaration, just return NULL.
26883 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
26885 * config/i386/sse.md (FMAMODEM): New mode iterator.
26886 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
26887 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
26889 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
26891 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
26892 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
26893 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
26895 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26897 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
26898 (microblaze*-*-elf): Likewise.
26899 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
26901 * config/microblaze/microblaze-c.c: Add builtin defines for
26902 _LITTLE_ENDIAN and _BIG_ENDIAN.
26903 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
26904 add to TARGET_DEFAULT flags.
26905 Expand ASM_SPEC and LINK_SPEC.
26906 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
26907 * config/microblaze/microblaze.md: Update extendsidi2 and
26908 movdi_internal instructions to use low-order / high-order reg
26910 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
26911 options and inversemask / mask of LITTLE_ENDIAN.
26912 * config/microblaze/t-microblaze: Expand multilib options to
26913 include mlittle-endian (le) and update exceptions patterns.
26915 2013-02-08 Jakub Jelinek <jakub@redhat.com>
26917 PR rtl-optimization/56195
26918 * lra-constraints.c (get_reload_reg): Don't reuse regs
26919 if they have smaller mode than requested, if they have
26920 wider mode than requested, try to return a SUBREG.
26922 PR tree-optimization/56250
26923 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
26924 if type is unsigned and code isn't MULT_EXPR.
26926 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
26928 PR tree-optimization/56064
26929 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
26930 bits according to mode.
26931 * fixed-value.h (fixed_from_double_int)
26932 (const_fixed_from_double_int): Adjust comments.
26934 2013-02-08 Richard Biener <rguenther@suse.de>
26937 * lto-streamer.h (struct data_in): Remove current_file, current_line
26938 and current_col members.
26939 * lto-streamer-out.c (lto_output_location): Stream changed bits
26940 en-block for efficiency.
26941 * lto-streamer-in.c (clear_line_info): Remove.
26942 (lto_input_location): Cache current file, line and column
26943 globally via local statics. Read changed bits en-block.
26944 (input_function): Do not call clear_line_info.
26945 (lto_read_body): Likewise.
26946 (lto_input_toplevel_asms): Likewise.
26948 2013-02-08 Michael Matz <matz@suse.de>
26950 PR tree-optimization/52448
26951 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
26952 (nt_call_phase): New static.
26953 (add_or_mark_expr): Only mark accesses with newer phase than any
26955 (nonfreeing_call_p): New.
26956 (nt_init_block): Update nt_call_phase, mark blocks as visited.
26957 (nt_fini_block): Keep blocks marked as visited.
26958 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
26960 2013-02-08 Richard Biener <rguenther@suse.de>
26962 * ira.c (ira): Free broken dominator information.
26964 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
26966 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
26968 2013-02-08 Marek Polacek <polacek@redhat.com>
26970 * cfgloop.c (verify_loop_structure): Add more checking of headers.
26972 2013-02-08 Richard Biener <rguenther@suse.de>
26974 PR middle-end/56181
26975 * cfgloop.h (flow_loops_find): Adjust.
26976 (bb_loop_header_p): Declare.
26977 * cfgloop.c (bb_loop_header_p): New function split out from ...
26978 (flow_loops_find): ... here. Adjust function signature,
26979 support incremental loop structure update.
26980 (verify_loop_structure): Cleanup. Verify a loop is a loop.
26981 * cfgloopmanip.c (fix_loop_structure): Move ...
26982 * loop-init.c (fix_loop_structure): ... here.
26983 (apply_loop_flags): Split out from ...
26984 (loop_optimizer_init): ... here.
26985 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
26986 in incremental mode, only remove dead loops here.
26988 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
26991 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
26992 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
26993 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
26994 (*round<mode>3.libgcc): New insns for fixed-modes.
26995 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
26996 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
26997 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
26998 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
26999 implementations. Define to __builtin_avr_absFX,
27000 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27001 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27002 __builtin_avr_countlsFX, respectively.
27003 * config/avr/avr-c.c (target.h): Include it.
27004 (enum avr_builtin_id): New enum.
27005 (avr_resolve_overloaded_builtin): New static function.
27006 (avr_register_target_pragmas): Use it to set
27007 targetm.resolve_overloaded_builtin.
27008 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27009 tree nodes used by DEF_BUILTIN.
27010 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27011 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27012 <AVR_BUILTIN_xxBITS>: Same.
27014 2013-02-08 Richard Biener <rguenther@suse.de>
27016 * cfgloop.c (verify_loop_structure): Properly handle
27017 a loop exiting to another loop header.
27018 * ira-int.h (ira_loops): Remove.
27019 * ira.c (ira_loops): Remove.
27020 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27021 (do_reload): Use loop_optimizer_finalize.
27022 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27023 number_of_loops to access the loop tree.
27024 (more_one_region_p): Likewise.
27025 (finish_loop_tree_nodes): Likewise.
27026 (rebuild_regno_allocno_maps): Likewise.
27027 (mark_loops_for_removal): Likewise.
27028 (mark_all_loops_for_removal): Likewise.
27029 (remove_unnecessary_regions): Likewise.
27030 (ira_build): Likewise.
27031 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27033 2013-02-08 Richard Biener <rguenther@suse.de>
27035 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27036 * ipa-pure-const.c (analyze_function): Avoid calling
27037 mark_irreducible_loops twice.
27038 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27040 2013-02-07 David S. Miller <davem@davemloft.net>
27042 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27044 * var-tracking.c (vt_add_function_parameter): Test the presence of
27045 HAVE_window_save properly and do not remap argument registers when
27046 we have a leaf function.
27048 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27051 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27053 * config/i386/i386.c (ix86_print_operand): Ditto.
27055 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27057 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27059 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27061 PR rtl-optimization/56225
27062 * lra-constraints.c (process_alt_operands): Check that reload hard
27063 reg can hold value for strict_low_part.
27065 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27068 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27069 dwarf2out_end_function.
27070 (in_first_function_p, maybe_at_text_label_p,
27071 first_loclabel_num_not_at_text_label): New variables.
27072 (dwarf2out_var_location): In the first function find out
27073 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27074 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27077 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27079 PR rtl-optimization/56178
27080 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27081 SUBREG of a register. Tidy up related block of code.
27082 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27083 note if the source is a register or a SUBREG of a register.
27085 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27088 * config/rs6000/rs6000.md (ptrm): New mode attr.
27089 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27090 call_value_indirect_aix<pttrsize>,
27091 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27094 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27096 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27097 if -bnortl. Convert to strcmp and strncmp.
27099 2013-02-07 Alan Modra <amodra@gmail.com>
27102 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27103 addresses won't wrap when offsetting.
27104 (rs6000_secondary_reload): Provide secondary reloads needed for
27105 wrapping LO_SUM addresses.
27107 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27109 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27110 MACH, just __MACH__.
27112 2013-02-06 Richard Biener <rguenther@suse.de>
27114 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27115 instead of calling fix_loop_structure.
27117 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27119 PR middle-end/56217
27120 * omp-low.c (use_pointer_for_field): Return false if
27121 lower_send_shared_vars doesn't generate any copy-out code.
27123 2013-02-06 Tom de Vries <tom@codesourcery.com>
27125 PR rtl-optimization/56131
27126 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27127 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27128 of the label is NULL. Add comment.
27130 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27132 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27135 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27136 (STATIC_LIBTSAN_LIBS): Likewise.
27137 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27138 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27139 is defined, don't add anything else beyond that.
27140 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27141 (LINK_COMMAND_SPEC): Use them.
27143 PR tree-optimization/56205
27144 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27145 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27146 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27148 2013-02-05 Richard Biener <rguenther@suse.de>
27150 PR tree-optimization/53342
27151 PR tree-optimization/53185
27152 * tree-vectorizer.h (vect_check_strided_load): Remove.
27153 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27154 not disallow peeling for vectorized strided loads.
27155 (vect_check_strided_load): Make static and simplify.
27156 (vect_analyze_data_refs): Adjust.
27157 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27158 correctly when vectorizing strided loads.
27160 2013-02-05 Richard Biener <rguenther@suse.de>
27162 * doc/install.texi: Refer to ISL, not PPL.
27164 2013-02-05 Jan Hubicka <jh@suse.cz>
27166 PR tree-optimization/55789
27167 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27169 2013-02-05 Jan Hubicka <jh@suse.cz>
27171 PR tree-optimization/55789
27172 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27173 the dead call anyway.
27175 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27178 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27180 2013-02-04 Alexander Potapenko <glider@google.com>
27181 Jack Howarth <howarth@bromo.med.uc.edu>
27182 Jakub Jelinek <jakub@redhat.com>
27185 * config/darwin.c (sort_ctor_records): Stabilized qsort
27186 on constructor priority by using original position.
27187 (finalize_ctors): New routine to sort constructors by
27188 priority before use in assemble_integer.
27189 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27191 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27194 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27195 about visibility on artificial decls.
27196 * config/sol2.c (solaris_assemble_visibility): Likewise.
27198 2013-02-04 Kai Tietz <ktietz@redhat.com>
27201 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27202 argument and improve checking of return-argument types for 16-byte
27204 (ix86_function_value_1): Add additional valtype argument on call
27205 of function_value_64.
27206 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27207 handling infunction_value_64 function.
27209 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27211 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27213 2013-02-04 Richard Biener <rguenther@suse.de>
27215 PR tree-optimization/56188
27216 * tree-ssa-structalias.c (label_visit): Consider case with
27217 initially non-empty points-to set.
27218 (perform_var_substitution): Dump node mapping and clean up.
27220 2013-02-04 Richard Guenther <rguenther@suse.de>
27223 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
27224 node prevail as last resort.
27225 (lto_symtab_merge_decls): Remove guard on LTRANS here.
27226 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
27228 2013-02-04 Richard Biener <rguenther@suse.de>
27230 PR tree-optimization/56113
27231 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
27233 (equiv_class_lookup_or_add): ... this.
27234 (label_visit): Adjust and fix error in previous patch.
27235 (perform_var_substitution): Adjust.
27237 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
27239 * config/sh/divtab.c: Fix formatting and comments throughout the file.
27240 * config/sh/sh4-300.md: Likewise.
27241 * config/sh/sh4a.md: Likewise.
27242 * config/sh/constraints.md: Likewise.
27243 * config/sh/sh.md: Likewise.
27244 * config/sh/netbsd-elf.h: Likewise.
27245 * config/sh/predicates.md: Likewise.
27246 * config/sh/sh-protos.h: Likewise.
27247 * config/sh/ushmedia.h: Likewise.
27248 * config/sh/linux.h: Likewise.
27249 * config/sh/sh.c: Likewise.
27250 * config/sh/superh.h: Likewise.
27251 * config/sh/elf.h: Likewise.
27252 * config/sh/sh4.md: Likewise.
27253 * config/sh/sh.h: Likewise.
27255 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27257 * config/pa/constraints.md: Adjust unused letters. Change "T"
27258 constraint to match_test floating_point_store_memory_operand().
27259 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
27260 (base14_operand): New.
27261 (floating_point_store_memory_operand): New.
27262 (integer_store_memory_operand): Revise to use base14_operand and
27263 reg_plus_base_memory_operand.
27264 (move_dest_operand): Allow symbolic_memory_operands.
27265 (symbolic_memory_operand): Check for LO_SOM.
27266 (symbolic_operand): Change default case to break.
27267 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
27268 CONST_DOUBLE values to be reloaded by putting them into memory when
27269 the destination is a floating point register.
27270 (movdf): Remove code to handle CONST_DOUBLE.
27272 (reload_indf_r1): New.
27273 (reload_insf_r1): New.
27274 Consistently use "Q" and "T" constraints with integer and floating
27275 point move instructions, respectively.
27276 (movdi): Remove FAIL.
27277 Change predicate for source operand unamed DImode move from
27278 general_operand to move_src_operand.
27279 (umulsidi3): Change predicate for destination operand to
27281 Likewise for similar unamed patterns.
27282 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
27283 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
27284 (hppa_legitimize_address): Simplify mask calculation.
27285 (pa_emit_move_sequence): Revised handling of secondary reloads from
27286 REG+D addresses for floating point loads and stores. Directly handle
27287 loading CONST0_RTX (mode) to a floating point register.
27288 (pa_secondary_reload): Handle reloading DF and SFmode constant values
27289 to floating point registers. Don't restrict secondary reloads to
27290 floating point registers to integer modes. Revise some comments and
27292 (TARGET_LEGITIMATE_ADDRESS_P): Define.
27293 (pa_legitimate_address_p): New.
27294 (pa_legitimize_reload_address): New.
27295 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
27296 (STRICT_REG_OK_FOR_BASE_P): New.
27297 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
27298 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
27300 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
27301 Andrew Dixie <andrewd@gentrack.com>
27303 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
27306 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
27308 * expmed.c (extract_bit_field_1): Pass the full width of the
27309 structure to get_best_reg_extraction_insn.
27311 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27314 * configure.ac (use_cxa_atexit): Add AIX.
27315 * configure: Regenerate.
27317 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
27319 2013-02-01 Jakub Jelinek <jakub@redhat.com>
27322 * final.c (need_profile_function): New variable.
27323 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
27324 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
27325 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
27326 notes, targetm.asm_out.function_prologue doesn't emit anything,
27327 HAVE_prologue and profiler should be emitted before prologue,
27328 set need_profile_function instead of emitting it.
27329 (final_scan_insn): If need_profile_function, emit
27330 profile_function on the first NOTE_INSN_BASIC_BLOCK or
27331 NOTE_INSN_FUNCTION_BEG note.
27333 2013-02-01 Richard Henderson <rth@redhat.com>
27335 * config/rs6000/rs6000.md (smulditi3): New.
27338 * config/alpha/alpha.md (umulditi3): New.
27340 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27342 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
27343 (ASM_OUTPUT_ALIGNED_LOCAL): New.
27345 2013-02-01 Richard Biener <rguenther@suse.de>
27347 PR tree-optimization/56113
27348 * tree-ssa-structalias.c (label_visit): Reduce work for
27349 single-predecessor nodes.
27351 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
27353 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
27354 range isn't testing for zero.
27356 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
27358 PR middle-end/56113
27359 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
27361 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
27362 Nick Clifton <nickc@redhat.com>
27364 * config/v850/constraints.md (Q): Define as a memory constraint.
27365 * config/v850/predicates.md (label_ref_operand): New predicate.
27366 (e3v5_shift_operand): New predicate.
27367 (ior_operator): New predicate.
27368 * config/v850/t-v850: Add e3v5 multilib.
27369 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
27370 (v850_gen_movdi): Prototype.
27371 * config/v850/v850.c: Add support for e3v5 architecture.
27372 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
27374 (construct_save_jarl): Add e3v5 long JARL support.
27375 (v850_adjust_insn_length): New function. Adjust length of call
27376 insns when using e3v5 instructions.
27377 (v850_gen_movdi): New function: Generate instructions to move a
27379 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
27380 (CPP_SPEC): Define __v850e3v5__ as appropriate.
27381 (TARGET_USE_FPU): Enable for e3v5.
27382 (CONST_OK_FOR_W): New macro.
27383 (ADJUST_INSN_LENGTH): Define.
27384 * config/v850/v850.md (UNSPEC_LOOP): Define.
27385 (attr cpu): Add v850e3v5.
27386 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
27387 (movdi): New pattern.
27388 (movdi_internal): New pattern.
27389 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
27390 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
27391 (cstoresf4): Likewise.
27392 (cstoredf4): Likewise.
27393 (insv): New pattern.
27394 (rotlso3_a): New pattern.
27395 (rotlsi3_b): New pattern
27396 (rotlsi3_v850e3v5): New pattern.
27397 (doloop_begin): New pattern.
27398 (fix_loop_counter): New pattern.
27399 (doloop_end): New pattern.
27400 (branch_normal): Add e3v5 long branch support.
27401 (branch_invert): Likewise.
27402 (branch_z_normal): Likewise.
27403 (branch_z_invert): Likewise.
27404 (branch_nz_normal): Likewise.
27405 (branch_nz_invert): Likewise.
27406 (call_internal_short): Add e3v5 register-indirect JARL support.
27407 (call_internal_long): Likewise.
27408 (call_value_internal_short): Likewise.
27409 (call_value_internal_long): Likewise.
27410 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
27411 (mloop): New option.
27412 * config.gcc: Add support for configuring v840e3v5 target.
27413 * doc/invoke.texi: Document new v850 specific command line options.
27415 2013-01-31 Paul Koning <ni1d@arrl.net>
27419 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
27420 children if parent is a class.
27421 (prune_unused_types_prune): Don't add DW_AT_declaration.
27423 2013-01-31 Richard Biener <rguenther@suse.de>
27425 PR tree-optimization/56157
27426 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
27427 match up operand with SLP child.
27429 2013-01-31 Jason Merrill <jason@redhat.com>
27432 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
27433 parameters the first time.
27434 (gen_scheduled_generic_parms_dies): Check completeness here.
27436 2013-01-31 Richard Biener <rguenther@suse.de>
27438 PR middle-end/53073
27439 * common.opt (faggressive-loop-optimizations): New flag,
27440 enabled by default.
27441 * doc/invoke.texi (faggressive-loop-optimizations): Document.
27442 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
27443 infer_loop_bounds_from_undefined by it.
27445 2013-01-31 Richard Biener <rguenther@suse.de>
27447 PR tree-optimization/56150
27448 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
27449 visit virtual operands.
27450 (find_uses_to_rename_bb): Likewise.
27452 2013-01-31 Richard Biener <rguenther@suse.de>
27454 PR tree-optimization/56150
27455 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
27456 mixed store non-store stmts.
27458 2013-01-30 Jakub Jelinek <jakub@redhat.com>
27461 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
27462 LIBASAN_EARLY_SPEC is defined.
27463 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
27464 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
27466 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
27469 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
27470 invalid args instead of ICEing on it.
27471 (ix86_valid_target_attribute_tree): Return error_mark_node if
27472 ix86_valid_target_attribute_inner_p failed.
27473 (ix86_valid_target_attribute_p): Return false only if
27474 ix86_valid_target_attribute_tree returned error_mark_node. Allow
27475 target("default") attribute.
27476 (sorted_attr_string): Change argument from const char * to tree,
27477 merge in all target attribute arguments rather than just one.
27478 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
27479 instead of free. Avoid using strcat.
27480 (ix86_mangle_function_version_assembler_name): Mangle
27481 target("default") as if no target attribute is present. Adjust
27482 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
27483 instead of xmalloc and XDELETEVEC instead of free.
27484 (ix86_function_versions): Don't return true if one of the decls
27485 doesn't have target attribute. If they don't and one of the decls
27486 is DECL_FUNCTION_VERSIONED, report an error. Adjust
27487 sorted_attr_string caller. Use XDELETEVEC instead of free.
27488 (ix86_supports_function_versions): Remove.
27489 (make_name): Fix up formatting.
27490 (make_dispatcher_decl): Remove resolver_name and its initialization.
27491 Avoid leaking memory.
27492 (is_function_default_version): Return true if there is
27493 target("default") attribute rather than no target attribute at all.
27494 (make_resolver_func): Avoid leaking memory.
27495 (ix86_generate_version_dispatcher_body): Likewise.
27496 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
27497 * target.def (supports_function_versions): Remove.
27498 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
27499 * doc/tm.texi: Regenerated.
27501 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
27503 PR rtl-optimization/56144
27504 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
27505 for values with side effects.
27507 2013-01-30 Richard Biener <rguenther@suse.de>
27509 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
27510 (sparseset_pop): Likewise.
27511 * cfganal.c (compute_idf): Likewise. Increase work-stack size
27512 to be able to use quick_push in the worker loop.
27514 2013-01-30 Marek Polacek <polacek@redhat.com>
27516 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
27518 2013-01-30 Richard Biener <rguenther@suse.de>
27521 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
27523 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
27525 PR tree-optimization/56064
27526 * fixed-value.c (fixed_from_double_int): New function.
27527 * fixed-value.h (fixed_from_double_int): New prototype.
27528 (const_fixed_from_double_int): New static inline function.
27529 * fold-const.c (native_interpret_fixed): New static function.
27530 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
27531 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
27532 (native_encode_fixed): New static function.
27533 (native_encode_expr) <FIXED_CST>: Use it.
27534 (native_interpret_int): Move double_int worker code to...
27535 * double-int.c (double_int::from_buffer): ...this new static method.
27536 * double-int.h (double_int::from_buffer): Prototype it.
27538 2013-01-30 Richard Biener <rguenther@suse.de>
27540 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
27541 New pointer-map and obstack.
27542 (init_alias_vars): Allocate pointer-map and obstack.
27543 (delete_points_to_sets): Free them.
27544 (find_what_var_points_to): Cache result.
27545 (find_what_p_points_to): Adjust for changed interface of
27546 find_what_var_points_to.
27547 (compute_points_to_sets): Likewise.
27548 (ipa_pta_execute): Likewise.
27550 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
27552 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
27553 * configure: Regenerate.
27554 * config.in: Regenerate.
27555 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
27556 #nobits/#progbits if supported.
27558 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
27561 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
27562 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
27563 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
27565 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27567 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
27568 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
27570 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27572 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
27573 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
27575 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27577 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
27579 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
27580 * config/arm/cortex-a7.md: New bypasses using
27581 arm_mac_accumulator_is_result.
27583 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27585 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
27586 (cortex_a7_neon_mla): Likewise.
27587 (cortex_a7_fpfmad): New reservation.
27588 (cortex_a7_fpmacs): Use ffmas and update required units.
27589 (cortex_a7_fpmuld): Update required units and latency.
27590 (cortex_a7_fpmacd): Likewise.
27591 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
27592 (cortex_a7_neon). Likewise.
27593 (bypass) Update participating units.
27595 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27597 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
27598 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
27600 * config/arm/vfp11.md (vfp_farith): Use ffmas.
27601 (vfp_fmul): Use ffmad.
27602 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
27603 (cortex_r4_fmacd): Use ffmad.
27604 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
27605 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
27606 (cortex_a9_fmacd): Use ffmad.
27607 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
27608 (cortex_a8_vfp_macd): Use ffmad.
27609 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
27610 (cortex_a5_fpmacd): Use ffmad.
27611 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
27612 (cortex_a15_vfp_macd): Use ffmad.
27613 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
27615 2013-01-29 Jason Merrill <jason@redhat.com>
27618 * varasm.c (default_assemble_visibility): Don't warn about
27619 visibility on artificial decls.
27621 2013-01-29 Richard Biener <rguenther@suse.de>
27623 PR tree-optimization/56113
27624 * tree-ssa-structalias.c (equiv_class_lookup): Also return
27626 (label_visit): Free duplicate bitmaps and record the leader instead.
27627 (perform_var_substitution): Adjust.
27629 2013-01-29 Richard Biener <rguenther@suse.de>
27631 PR tree-optimization/55270
27632 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
27633 the CFG, schedule loops for fixup.
27635 2013-01-29 Nick Clifton <nickc@redhat.com>
27637 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
27640 2013-01-28 Leif Ekblad <leif@rdos.net>
27642 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
27643 * config/i386/i386.h (TARGET_RDOS): New macro.
27644 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
27645 * config/i386/i386.c (ix86_option_override_internal): For 64bit
27646 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
27647 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
27648 DEFAULT_LARGE_SECTION_THRESHOLD.
27649 * config/i386/i386.md (R14_REG, R15_REG): New constants.
27650 * config/i386/rdos.h: New file.
27651 * config/i386/rdos64.h: New file.
27653 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
27656 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
27659 2013-01-28 Jakub Jelinek <jakub@redhat.com>
27661 PR rtl-optimization/56117
27662 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
27663 call cselib_lookup_from_insn on the MEM before calling
27664 add_insn_mem_dependence.
27666 2013-01-28 Richard Biener <rguenther@suse.de>
27668 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
27669 to a stmt that didn't have one.
27670 (copy_phis_for_bb): Likewise for PHI arguments.
27671 (copy_debug_stmt): Likewise for debug stmts.
27673 2013-01-28 Richard Biener <rguenther@suse.de>
27675 PR tree-optimization/56034
27676 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
27677 (partition_builtin_p): Adjust.
27678 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
27679 it is the last partition.
27680 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
27681 up the vertex for the definition.
27682 (classify_partition): Classify whether a partition is a
27683 PKIND_REDUCTION, thus has uses outside of the loop.
27684 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
27685 Merge all PKIND_REDUCTION partitions into the last partition.
27686 (tree_loop_distribution): Seed partitions from reductions as well.
27688 2013-01-28 Jakub Jelinek <jakub@redhat.com>
27690 PR tree-optimization/56125
27691 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
27692 pow(x,c) into sqrt(x) * powi(x, n/2) or
27693 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
27694 optimizing for size.
27695 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
27696 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
27699 PR tree-optimization/56094
27700 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
27701 to UNKNOWN_LOCATION while gimplifying expr.
27703 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
27706 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
27707 operand 0 in movabs insn template for -masm=intel asm alternative.
27708 (*movabs<mode>_2): Ditto for operand 1.
27710 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
27713 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
27716 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27718 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
27721 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27723 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
27724 Undef to avoid warning.
27726 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27728 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
27729 * configure: Regenerate.
27731 2013-01-25 Jakub Jelinek <jakub@redhat.com>
27733 PR tree-optimization/56098
27734 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
27735 for stmts with volatile ops.
27736 (cond_store_replacement): Don't optimize if assign has volatile ops.
27737 (cond_if_else_store_replacement_1): Don't optimize if either
27738 then_assign or else_assign have volatile ops.
27739 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
27742 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
27744 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
27746 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
27748 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
27749 missing ':' in asm example.
27751 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
27753 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
27754 entries into lane and laneq entries.
27755 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
27756 Remove AdvSIMD scalar modes.
27757 (aarch64_sq<r>dmulh_laneq<mode>): New.
27758 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
27760 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
27761 builtin implementations to relfect changes in RTL in aarch64-simd.md.
27762 * config/aarch64/iterators.md (VCOND): New.
27765 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
27768 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
27769 Add NULL LIBNAME argument to existing definitions.
27770 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
27771 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
27772 * config/avr/avr.c (DEF_BUILTIN): Same.
27773 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
27774 (avr_expand_builtin): Expand to a vanilla call if a libgcc
27775 implementation is available (DECL_ASSEMBLER_NAME is set).
27776 (avr_fold_absfx): New static function.
27777 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
27778 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
27779 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
27780 AVR_BUILTIN_ABSLLK.
27781 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
27782 (abshk, absk, abslk, absllk): Provide as static inline functions.
27784 2013-01-25 Marek Polacek <polacek@redhat.com>
27786 PR tree-optimization/56035
27787 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
27789 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
27791 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
27792 (*movtf_internal_rex64): Add (!o,C) alternative
27793 (*movxf_internal_rex64): Ditto.
27794 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
27796 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
27798 * doc/invoke.texi: fix typo.
27799 * doc/objc.texi: fix typo.
27801 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
27803 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
27804 for the first two alternatives.
27806 2013-01-24 Diego Novillo <dnovillo@google.com>
27808 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
27809 (ggc-zone.o): Remove.
27810 * configure.ac: Remove option --with-gc.
27811 * configure: Re-generate.
27812 * doc/install.texi: Remove documentation for --with-gc.
27813 * gengtype.c (write_enum_defn): Remove. Update all users.
27814 (write_Types_process_field): Remove generation of gt_e_* argument.
27815 (output_type_enum): Remove. Update all users.
27816 (write_enum_defn): Remove. Update all users.
27817 (enum alloc_zone): Remove. Update all users.
27818 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
27819 * ggc-common.c (ggc_splay_alloc): Remove first argument.
27820 Update all callers.
27821 (struct ptr_data): Remove field TYPE. Update all users.
27822 (gt_pch_note_object): Remove argument TYPE. Update all users.
27823 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
27825 * ggc-none.c (ggc_alloc_typed_stat): Remove.
27826 (struct alloc_zone): Remove.
27827 (ggc_internal_alloc_zone_stat): Remove.
27828 (ggc_internal_cleared_alloc_zone_stat): Remove.
27829 * ggc-page.c (ggc_alloc_typed_stat): Remove.
27830 (ggc_pch_count_object): Remove last argument. Update all users.
27831 (ggc_pch_alloc_object): Remove last argument. Update all users.
27832 (struct alloc_zone): Remove.
27833 * ggc-zone.c: Remove.
27834 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
27835 (struct alloc_zone): Remove.
27836 (ggc_alloc_typed_stat): Remove.
27837 (ggc_alloc_typed): Remove.
27838 (ggc_splay_alloc): Remove first argument.
27839 (rtl_zone): Remove. Update all users.
27840 (tree_zone): Remove. Update all users.
27841 (tree_id_zone): Remove. Update all users.
27842 (ggc_internal_zone_alloc_stat): Remove. Update all users.
27843 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
27844 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
27845 * tree-ssanames.c: Remove references to zone allocator in comments.
27847 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
27849 * config/avr/avr.c (avr_out_fract): Make register numbers that
27850 might be outside of source operand signed.
27852 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
27854 * config/i386/constraints.md (Yf): New constraint.
27855 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
27856 of f constraint to conditionaly disable x87 register preferences.
27857 (*movdf_internal): Ditto.
27858 (*movsf_internal): Ditto.
27860 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
27862 PR inline-asm/55934
27863 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
27864 that have operands with impossible constraints.
27865 Add a FIXME for a speed-up opportunity.
27866 * lra-constraints.c (process_alt_operands): Verify that a class
27867 selected from constraints on asms is valid for the operand mode.
27868 (curr_insn_transform): Remove incorrect comment.
27870 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
27872 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
27873 TOC operand is a valid symbol ref in the constant pool.
27875 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27877 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
27879 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
27882 * config/avr/stdfix.h: New file.
27883 * t-avr (stdfix-gcc.h): New rule to build it.
27884 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
27886 2013-01-23 Kostya Serebryany <kcc@google.com>
27888 * config/darwin.h: remove dependency on
27889 CoreFoundation (asan on Mac OS).
27891 2013-01-23 Jakub Jelinek <jakub@redhat.com>
27894 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
27895 instead of cmpdi_operand for first comparison operand.
27896 Don't assert that comparison operands aren't both constants.
27898 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
27900 * doc/install.texi (Downloading the Source): Update references to
27901 downloading separate components.
27903 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
27905 * doc/extend.texi (__int128): Improve grammar.
27907 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
27910 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
27911 alternative to (o,r).
27912 (*movdi_internal_rex64): Remove (!o,n) alternative.
27913 (DImode immediate->memory splitter): Remove.
27914 (DImode immediate->memory peephole2): Remove.
27915 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
27916 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
27917 alternative to (!o,*r).
27918 (*movtf_internal_sse): New pattern.
27919 (*movxf_internal_rex64): New pattern.
27920 (*movxf_internal): Disable for TARGET_64BIT.
27921 (*movdf_internal_rex64): Remove (!o,F) alternative.
27923 2013-01-22 Jakub Jelinek <jakub@redhat.com>
27925 PR middle-end/56074
27926 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
27927 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
27928 * tree-vect-loop-manip.c (find_loop_location): Also ignore
27929 stmt locations where LOCATION_LOCUS of the stmt location is
27930 UNKNOWN_LOCATION or BUILTINS_LOCATION.
27933 * config/i386/i386.md (UNSPEC_STOS): New.
27934 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
27935 *strsetqi_1): Add UNSPEC_STOS.
27937 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
27940 * doc/invoke.texi: Remove left over -Wsynth example.
27942 2013-01-21 Jakub Jelinek <jakub@redhat.com>
27944 PR tree-optimization/56051
27945 * fold-const.c (fold_binary_loc): Don't fold
27946 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
27947 a narrowing conversion, or widening conversion from signed
27950 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
27952 PR rtl-optimization/56023
27953 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
27954 dependent on debug instruction.
27956 2013-01-21 Martin Jambor <mjambor@suse.cz>
27958 PR middle-end/56022
27959 * function.c (allocate_struct_function): Call
27960 invoke_set_current_function_hook earlier.
27962 2013-01-21 Jakub Jelinek <jakub@redhat.com>
27964 * reload1.c (init_reload): Only initialize reload_obstack
27965 during the first call.
27967 2013-01-21 Marek Polacek <polacek@redhat.com>
27969 * cfgloop.c (verify_loop_structure): Fix up grammar.
27971 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
27973 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
27974 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
27976 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27979 * config/arm/marvell-pj4.md: Update copyright year.
27980 Fix up use of alu to alu_reg and simple_alu_imm.
27982 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
27984 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
27986 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
27989 * lra-constraints.c (curr_insn_transform): Don't reuse original
27990 insn for secondary memory move when memory mode should be different.
27992 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27994 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
27995 atomic_storedi_1): New patterns.
27997 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27999 btver2 pipeline descriptions.
28000 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28002 * config/i386/i386.md (btver2_decode): New type attributes.
28003 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28005 * config/i386/btver2.md: New file describing btver2 pipelines.
28007 2013-01-19 Andrew Pinski <apinski@cavium.com>
28009 PR tree-optimization/52631
28010 * tree-ssa-sccvn (visit_use): Before looking up the original
28011 statement, try looking up the simplified expression.
28013 2013-01-19 Anthony Green <green@moxielogic.com>
28015 * config/moxie/moxie.c (moxie_expand_prologue): Set
28016 current_function_static_stack_size.
28018 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28020 PR tree-optimization/56029
28021 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28022 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28024 2013-01-18 Sharad Singhai <singhai@google.com>
28026 PR tree-optimization/55995
28027 * dumpfile.c (dump_loc): Print location only if available.
28028 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28030 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28033 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28034 secondary memory move.
28035 (inherit_reload_reg): Use rclass instead of cl for
28036 check_secondary_memory_needed_p.
28038 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28040 PR middle-end/56015
28041 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28042 the case where writing real complex part of target modifies op1.
28044 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28046 * config/aarch64/aarch64-simd.md
28047 (aarch64_vcond_internal<mode>): Handle unordered cases.
28048 * config/aarch64/iterators.md (v_cmp_result): New.
28050 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28051 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28053 * config/arm/marvell-pj4.md: New file.
28054 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28055 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28056 (generic_vfp): Likewise.
28057 * config/arm/arm-cores.def: Add marvell-pj4.
28058 * config/arm/arm-tune.md: Regenerate.
28059 * config/arm/arm-tables.opt: Regenerate.
28060 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28061 * doc/invoke.texi: Document marvell-pj4.
28063 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28065 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28067 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28072 * var-tracking.c (negative_power_of_two_p): New.
28073 (global_get_addr_cache, local_get_addr_cache): New.
28074 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28075 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28077 (vt_stack_offset_p): Remove.
28078 (vt_canon_true_dep): Always canonicalize loc's address.
28079 (clobber_overlapping_mems): Make sure we have a MEM.
28080 (local_get_addr_clear_given_value): New.
28081 (val_reset): Clear local cached entries.
28082 (compute_bb_dataflow): Create and release the local cache.
28083 Disable duplicate MEMs clobbering.
28084 (emit_notes_in_bb): Clobber MEMs likewise.
28085 (vt_emit_notes): Create and release the local cache.
28086 (vt_initialize, vt_finalize): Create and release the global
28087 cache, respectively.
28088 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28090 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28092 PR libmudflap/53359
28093 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28094 not found in the symtab.
28096 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28099 PR rtl-optimization/55547
28100 PR rtl-optimization/53827
28103 * alias.c (offset_overlap_p): New, factored out of...
28104 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28105 the conservative special case for symbolic constants. Don't
28106 adjust zero sizes on alignment.
28108 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28110 PR rtl-optimization/52573
28111 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28112 REG_UNUSED for the same register.
28114 2013-01-17 Richard Biener <rguenther@suse.de>
28115 Marek Polacek <polacek@redhat.com>
28117 PR rtl-optimization/55833
28118 * loop-unswitch.c (unswitch_loops): Move loop verification...
28119 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28120 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28121 Set it to true when we're removing a loop from hierarchy tree in
28122 an irreducible region.
28123 (fix_bb_placements): Adjust caller.
28124 (fix_loop_placements): Likewise.
28126 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28128 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28129 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28130 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28131 Remove ID. Adjust comments.
28132 * config/avr/avr-c.c (avr_builtin_name): Remove.
28133 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28134 * config/avr/avr.c (avr_tolower): New static function.
28135 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28136 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28137 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28140 2013-01-17 Jan Hubicka <jh@suse.cz>
28142 PR tree-optimization/55273
28143 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28145 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28148 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28149 store through atomic_store<mode>_1.
28150 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28152 2013-01-17 Martin Jambor <mjambor@suse.cz>
28154 PR tree-optimizations/55264
28155 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28156 for virtual methods.
28157 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28158 virtual methods before inlining is over.
28159 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28161 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28164 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28166 PR rtl-optimization/56005
28167 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28168 pending reads for prefetch.
28170 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28172 * config/aarch64/aarch64.md
28173 (*cstoresi_neg_uxtw): New pattern.
28174 (*cmovsi_insn_uxtw): New pattern.
28175 (*<optab>si3_uxtw): New pattern.
28176 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28177 (*<optab>si3_insn_uxtw): New pattern.
28178 (*bswapsi2_uxtw): New pattern.
28180 2013-01-16 Richard Biener <rguenther@suse.de>
28182 * tree-inline.c (tree_function_versioning): Remove set but
28183 never used variable.
28185 2013-01-16 Richard Biener <rguenther@suse.de>
28187 PR tree-optimization/55964
28188 * tree-flow.h (rename_variables_in_loop): Remove.
28189 (rename_variables_in_bb): Likewise.
28190 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28191 (copy_loop_before): Adjust and delete update-ssa status.
28192 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28193 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28194 (rename_variables_in_loop): Remove.
28195 (slpeel_update_phis_for_duplicate_loop): Likewise.
28196 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28197 use available cfg machinery instead of duplicating it.
28198 Update PHI nodes and perform poor-mans SSA update here.
28199 (slpeel_tree_peel_loop_to_edge): Adjust.
28201 2013-01-16 Richard Biener <rguenther@suse.de>
28203 PR tree-optimization/54767
28204 PR tree-optimization/53465
28205 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28206 (vrp_visit_phi_node): For PHI arguments coming via backedges
28207 drop all symbolical range information.
28208 (execute_vrp): Compute backedges.
28210 2013-01-16 Richard Biener <rguenther@suse.de>
28212 * doc/install.texi: Update CLooG and ISL requirements to
28215 2013-01-16 Christian Bruel <christian.bruel@st.com>
28218 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
28219 (broken_move): Handle UNSPECV_SP_SWITCH_B.
28220 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
28222 2013-01-16 DJ Delorie <dj@redhat.com>
28224 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
28225 (UNSPECV_SP_SWITCH_E): New.
28226 (sp_switch_1): Change to an unspec.
28227 (sp_switch_2): Change to an unspec. Don't use post-inc when we
28230 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
28232 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
28233 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
28234 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
28235 (expand_mem_thread_fence): Ditto.
28236 (expand_mem_signal_fence): Ditto.
28237 (expand_atomic_load): Ditto.
28238 (expand_atomic_store): Ditto.
28240 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
28242 PR rtl-optimization/55547
28243 PR rtl-optimization/53827
28246 * alias.c (memrefs_conflict_p): Set sizes to negative after
28249 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28252 * function.c (thread_prologue_and_epilogue_insns): Always
28253 add crtl->drap_reg to set_up_by_prologue.set, even if
28254 stack_realign_drap is false.
28256 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
28258 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
28259 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
28260 *call): Fix indention.
28262 2013-01-15 Tom de Vries <tom@codesourcery.com>
28265 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
28268 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
28270 PR rtl-optimization/55153
28271 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
28273 2013-01-15 Martin Jambor <mjambor@suse.cz>
28275 PR tree-optimization/55920
28276 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
28277 accesses as grp_to_be_debug_replaced.
28279 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28281 PR tree-optimization/55920
28282 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
28283 there is non-useless type conversion needed from debug rhs to lhs,
28284 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
28286 2013-01-15 Joseph Myers <joseph@codesourcery.com>
28287 Mikael Pettersson <mikpe@it.uu.se>
28290 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
28292 (ASM_OUTPUT_CASE_LABEL): Remove.
28293 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
28294 * final.c (shorten_branches): Update alignment of labels before
28295 jump tables if CASE_VECTOR_SHORTEN_MODE.
28297 2013-01-15 Richard Biener <rguenther@suse.de>
28300 * system.h: Do not include gmp.h for building host tools.
28302 2013-01-15 Richard Biener <rguenther@suse.de>
28304 PR middle-end/55882
28305 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
28306 account for bitpos when computing alignment.
28308 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
28310 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
28311 (ix86_target_macros_internal): Likewise.
28313 * config/i386/i386.c (m_CORE2I7): Removed.
28314 (m_CORE_HASWELL): New macro.
28315 (m_CORE_ALL): Likewise.
28316 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
28317 (initial_ix86_arch_features): Likewise.
28318 (processor_target_table): Initializations for Core avx2.
28319 (cpu_names): New names "core-avx2".
28320 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
28321 PROCESSOR_CORE_HASWELL.
28322 (ix86_issue_rate): New case.
28323 (ia32_multipass_dfa_lookahead): Likewise.
28324 (ix86_sched_init_global): Likewise.
28326 * config/i386/i386.h (TARGET_HASWELL): New macro.
28327 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
28328 (processor_type): New PROCESSOR_HASWELL.
28330 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28332 PR tree-optimization/55955
28333 * tree-vect-loop.c (vectorizable_reduction): Give up early on
28334 *SHIFT_EXPR and *ROTATE_EXPR codes.
28336 PR tree-optimization/48766
28337 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
28338 -ftrapv disable -fwrapv.
28340 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28343 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
28344 etc. to 1 and not to __flash.
28345 Use LL suffix for __INT24_MAX__ with -mint8.
28346 Use ULL suffix for __UINT24_MAX__ with -mint8.
28348 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28350 * config/avr/avr-arch.h
28351 (struct base_arch_s): Use typedef avr_arch_t instead.
28352 (struct arch_info_s): Use typedef avr_arch_info_t instead.
28353 (struct mcu_type_s): Use typedef avr_mcu_t instead.
28354 * config/avr/avr.c: Same.
28355 * config/avr/avr-devices.c: Same.
28356 * config/avr/driver-avr.c: Same.
28357 * config/avr/gen-avr-mmcu-texi.c: Same.
28358 * config/avr/avr-mcus.def: Adjust comment.
28360 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
28362 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
28363 * config/aarch64/iterators.md (VALLDI): New.
28365 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
28366 Andi Kleen <ak@linux.intel.com>
28369 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
28370 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
28373 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28375 * config/avr/avr-stdint.h: Remove trailing blanks.
28376 * config/avr/avr-log.h: Same.
28377 * config/avr/avr-arch.h: Same.
28378 * config/avr/avr-devices.c: Same.
28379 * config/avr/avr-dimode.md: Same.
28380 * config/avr/predicates.md: Same.
28381 * config/avr/avr-c.c: Same. And fix typo.
28383 * config/avr/avr-protos.h: Same. And:
28384 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
28385 (init_cumulative_args): Rename to avr_init_cumulative_args.
28386 (expand_prologue): Rename to avr_expand_prologue.
28387 (expand_epilogue): Rename to avr_expand_epilogue.
28388 (adjust_insn_length): Rename to avr_adjust_insn_length.
28389 (notice_update_cc): Rename to avr_notice_update_cc.
28390 (final_prescan_insn): Rename to avr_final_prescan_insn.
28391 * config/avr/avr.c: Same.
28392 * config/avr/avr.h: Same.
28393 * config/avr/avr.md: Remove trailing blanks.
28394 (prologue): Use avr_expand_prologue.
28395 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
28397 2013-01-14 Richard Biener <rguenther@suse.de>
28399 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
28400 verify_location, collect_subblocks): New functions.
28401 (verify_gimple_in_cfg): Verify that locations only reference
28402 BLOCKs in the functions BLOCK tree.
28404 2013-01-14 Richard Biener <rguenther@suse.de>
28406 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
28408 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
28410 (insert_out_of_ssa_copy_on_edge): Likewise.
28411 (rewrite_close_phi_out_of_ssa): Likewise.
28412 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
28414 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
28415 propagated constants.
28416 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
28419 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28421 * config/avr/avr-modes.def: Add GPL copyright notice.
28423 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
28425 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
28426 MEMMODEL_MASK to determine memory model.
28427 (atomic_store<mode>): Ditto from operands[2].
28428 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
28430 2013-01-13 Jakub Jelinek <jakub@redhat.com>
28433 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
28434 (fold_gimple_assign): Don't call unshare_expr here.
28435 (fold_ctor_reference): Call unshare_expr.
28437 2013-01-13 Terry Guo <terry.guo@arm.com>
28439 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
28440 * doc/fragments.texi: Document MULTILIB_REUSE.
28441 * gcc.c (multilib_reuse): New internal spec.
28442 (set_multilib_dir): Also search multilib from multilib_reuse.
28443 * genmultilib (tmpmultilib3): Refactor code.
28444 (tmpmultilib4): Ditto.
28445 (multilib_reuse): New multilib argument.
28447 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
28449 * Makefile.in: Update copyright.
28451 2013-01-12 Tom de Vries <tom@codesourcery.com>
28453 PR middle-end/55890
28454 * calls.c (expand_call): Check if arg_nr is valid.
28456 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
28458 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
28459 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
28460 documentation. Add missing '__' in front of
28461 __builtin_ia32_packssdw256.
28463 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28466 * config/s390/s390.c (s390_preferred_reload_class): Do not return
28467 NO_REGS for larl operands.
28468 (s390_reload_larl_operand): Use s390_load_address instead of
28471 2013-01-11 Richard Biener <rguenther@suse.de>
28473 * tree-cfg.c (verify_node_sharing_1): Split out from ...
28474 (verify_node_sharing): ... here.
28475 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
28477 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
28479 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
28480 Substitute TREECHECKING.
28481 * configure: Regenerate.
28482 * Makefile.in (TREECHECKING): New.
28484 2013-01-11 Richard Guenther <rguenther@suse.de>
28486 PR tree-optimization/44061
28487 * tree-vrp.c (extract_range_basic): Compute zero as
28488 value-range for __builtin_constant_p of function parameters.
28490 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
28492 Update copyright years.
28494 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
28496 PR rtl-optimization/55672
28497 * lra-eliminations.c (mark_not_eliminable): Permit addition with
28498 const to be eliminable.
28500 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
28502 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
28503 * configure: Regenerate.
28505 2013-01-10 Richard Biener <rguenther@suse.de>
28507 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
28509 2013-01-10 Richard Biener <rguenther@suse.de>
28512 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
28513 locations for virtual PHI arguments.
28514 (rewrite_update_phi_arguments): Likewise.
28516 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
28518 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
28521 2013-01-10 Jakub Jelinek <jakub@redhat.com>
28523 PR tree-optimization/55921
28524 * tree-complex.c (expand_complex_asm): New function.
28525 (expand_complex_operations_1): Call it for GIMPLE_ASM.
28527 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28530 * config/s390/s390.c (s390_symref_operand_p)
28531 (s390_loadrelative_operand_p): Merge the two functions.
28532 (s390_check_qrst_address, print_operand_address): Add parameters
28533 to s390_loadrelative_operand_p invokation.
28534 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
28535 (s390_reload_larl_operand, s390_secondary_reload): Use
28536 s390_loadrelative_operand_p instead of s390_symref_operand_p.
28537 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
28539 2013-01-09 Mike Stump <mikestump@comcast.net>
28541 * dse.c (record_store): Remove unnecessary assert.
28543 2013-01-09 Jan Hubicka <jh@suse.cz>
28545 PR tree-optimization/55569
28546 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
28547 * cfgloop.h (scale_loop_profile): Likewise.
28549 2013-01-09 Jan Hubicka <jh@suse.cz>
28552 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
28554 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
28556 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
28558 PR middle-end/55114
28559 * expr.h (maybe_emit_group_store): Declare.
28560 * expr.c (maybe_emit_group_store): New function.
28561 * builtins.c (expand_builtin_int_roundingfn): Call it.
28562 (expand_builtin_int_roundingfn_2): Likewise.
28564 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
28566 PR rtl-optimization/55829
28567 * lra-constraints.c (match_reload): Add code for absent output.
28568 (curr_insn_transform): Add code for reloads of matched inputs
28571 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
28573 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
28574 attribute of movddup insn to DF.
28575 (*vec_interleave_lowv2df): Ditto.
28576 (vec_dupv2df): Ditto.
28578 2013-01-09 Jan Hubicka <jh@suse.cz>
28580 PR tree-optimiation/55875
28581 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
28582 EVERY_ITERATION parameter.
28583 (number_of_iterations_exit): Check if exit is executed every iteration.
28584 (idx_infer_loop_bounds): Similarly here.
28585 (n_of_executions_at_most): Simplify
28586 to only test for cases where statement is dominated by the
28587 particular bound; handle correctly the "postdominance" test.
28588 (scev_probably_wraps_p): Use max loop iterations info
28589 as a global bound first.
28591 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
28592 Nick Clifton <nickc@redhat.com>
28594 * config/v850/v850.md (cbranchsf4): New pattern.
28595 (cstoresf4): New pattern.
28596 (cbranchdf4): New pattern.
28597 (cstoredf4): New pattern.
28598 (movsicc): Disallow floating point comparisons.
28599 (cmpsf_le_insn): Fix order of operators.
28600 (cmpsf_lt_insn): Likewise.
28601 (cmpsf_eq_insn): Likewise.
28602 (cmpdf_le_insn): Likewise.
28603 (cmpdf_lt_insn): Likewise.
28604 (cmpdf_eq_insn): Likewise.
28605 (cmpsf_ge_insn): Use LE comparison.
28606 (cmpdf_ge_insn): Likewise.
28607 (cmpsf_gt_insn): Use LT comparison.
28608 (cmpdf_gt_insn): Likewise.
28609 (cmpsf_ne_insn): Delete pattern.
28610 (cmpdf_ne_insn): Delete pattern.
28611 * config/v850/v850.c (v850_gen_float_compare): Use
28612 gen_cmpdf_eq_insn for NE comparison.
28613 (v850_float_z_comparison_operator)
28614 (v850_float_nz_comparison_operator): Move from here ...
28615 * config/v850/predicates.md: ... to here. Move GT and GE
28616 comparisons into v850_float_z_comparison_operator.
28617 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
28619 (v850_float_nz_comparison_operator): Likewise.
28621 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28623 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
28624 with calls to gen_insvsi/gen_insvdi.
28626 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28628 * config/i386/i386.c (initial_ix86_tune_features): Set up
28629 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
28631 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
28632 Jakub Jelinek <jakub@redhat.com>
28634 PR tree-optimization/48189
28635 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
28636 If nitercst is 0, don't predict the exit edge.
28638 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28640 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
28641 in asm_fprintf with reg_names.
28642 (aarch64_print_operand_address): Likewise.
28643 (aarch64_return_addr): Likewise.
28644 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
28646 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28648 * config/pa/pa.h (VAL_U6_BITS_P): Define.
28649 (INT_U6_BITS): Likewise.
28650 * config/pa/predicates.md (uint6_operand): New predicate.
28651 (shift5_operand, shift6_operand): Likewise.
28652 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
28654 (lshrdi3): Use shift6_operand.
28655 (shrpsi4, shrpdi4): New insn patterns.
28656 (extzv): Delete expander.
28657 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
28658 predicates in unamed zero extract patterns. Tighten common constraint.
28659 (extv): Delete expander.
28660 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
28661 predicates in unamed sign extract patterns. Tighten common constraint.
28662 (insv): Delete expander.
28663 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
28664 predicates in unamed insert patterns. Tighten common constraint.
28665 Change uint32_operand predicate to uint6_operand predicate in unamed
28666 DImode pattern to insert constant values of type 1...1xxxx.
28668 2013-01-04 Jan Hubicka <jh@suse.cz>
28670 PR tree-optimization/55823
28671 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
28674 2013-01-08 Jakub Jelinek <jakub@redhat.com>
28675 Uros Bizjak <ubizjak@gmail.com>
28677 PR rtl-optimization/55845
28678 * df-problems.c (can_move_insns_across): Stop scanning at
28679 volatile_insn_p source instruction or give up if
28680 across_from .. across_to range contains any volatile_insn_p
28683 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
28685 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
28686 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
28688 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
28689 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
28691 2013-01-08 Jakub Jelinek <jakub@redhat.com>
28694 * asan.c (asan_clear_shadow): New function.
28695 (asan_emit_stack_protection): Use it.
28697 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
28699 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
28700 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
28701 with tab instead of space.
28703 2013-01-08 Nick Clifton <nickc@redhat.com>
28705 * config/rl78/rl78.c (rl78_expand_prologue): Always select
28706 register bank 0 at the start of an interrupt handler.
28707 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
28710 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
28712 * config/aarch64/aarch64-simd.md
28713 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
28714 (aarch64_simd_bsl): Likewise.
28715 (aarch64_vcond_internal<mode>): Likewise.
28716 (vcond<mode><mode>): Likewise.
28717 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
28718 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
28720 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
28722 * config/aarch64/aarch64-builtins.c
28723 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
28725 2013-01-08 Martin Jambor <mjambor@suse.cz>
28728 * tree-sra.c (analyze_access_subtree): Return true also after
28729 potentially creating a debug-only replacement.
28731 2013-01-08 Jakub Jelinek <jakub@redhat.com>
28733 PR middle-end/55890
28734 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
28736 PR tree-optimization/54120
28737 * tree-vrp.c (range_fits_type_p): Don't allow
28738 src_precision < precision from signed vr to unsigned_p
28739 if vr->min or vr->max is negative.
28740 (simplify_float_conversion_using_ranges): Test can_float_p
28741 against CODE_FOR_nothing.
28743 2013-01-08 Jakub Jelinek <jakub@redhat.com>
28744 Richard Biener <rguenther@suse.de>
28746 PR middle-end/55851
28747 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
28748 types instead of just INTEGER_TYPE types.
28750 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
28752 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
28755 2013-01-07 Steve Ellcey <sellcey@mips.com>
28758 * config/mips/mips.opt: Change mad to mmad to match documentation.
28760 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28763 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
28764 .progmemx.data now.
28766 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28769 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
28770 (avr_addrspace_t): Add .section_name field.
28771 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
28773 (avr_addrspace): Same. Initialize .section_name. Remove last
28774 NULL entry. Put __memx into .progmemx.data.
28775 (progmem_section_prefix): Remove.
28776 (avr_asm_init_sections): No need to initialize progmem_section.
28777 (avr_asm_named_section): Use avr_addrspace[].section_name to get
28778 section name prefix.
28779 (avr_asm_select_section): Ditto. And use get_unnamed_section to
28780 retrieve the progmem section.
28781 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
28782 boundary to run over avr_addrspace[].
28783 (avr_register_target_pragmas): Ditto.
28785 2013-01-06 Jakub Jelinek <jakub@redhat.com>
28787 * varasm.c (output_constant_def_contents): For asan_protect_global
28788 protected strings, adjust DECL_ALIGN if needed, before testing for
28790 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
28791 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
28793 (output_object_block): For asan protected decls, emit asan padding
28794 after their contents.
28795 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
28796 (asan_finish_file): Test it here instead.
28798 2013-01-07 Nick Clifton <nickc@redhat.com>
28799 Matthias Klose <doko@debian.org>
28800 Doug Kwan <dougkwan@google.com>
28801 H.J. Lu <hongjiu.lu@intel.com>
28804 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
28806 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
28808 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
28810 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
28812 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
28814 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28817 * doc/install.texi (Cross-Compiler-Specific Options): Document
28820 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
28822 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
28823 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
28824 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
28825 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
28826 vqmovun_high_s64): Fix source operand number and update copyright.
28828 2013-01-07 Richard Biener <rguenther@suse.de>
28830 PR middle-end/55890
28831 * gimple.h (gimple_call_builtin_p): New overload.
28832 * gimple.c (validate_call): New function.
28833 (gimple_call_builtin_p): Likewise.
28834 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
28835 Use gimple_call_builtin_p.
28836 (find_func_clobbers): Likewise.
28837 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
28838 (strlen_optimize_stmt): Likewise.
28840 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
28842 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
28843 (vld1q_dup_*): Likewise.
28844 (vld1_*): Likewise.
28845 (vld1q_*): Likewise.
28846 (vld1_lane_*): Likewise.
28847 (vld1q_lane_*): Likewise.
28849 2013-01-07 Richard Biener <rguenther@suse.de>
28851 * lto-streamer.h (LTO_minor_version): Bump to 2.
28853 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
28855 * config/aarch64/aarch64-protos.h
28856 (aarch64_const_double_zero_rtx_p): Rename to...
28857 (aarch64_float_const_zero_rtx_p): ...this.
28858 (aarch64_float_const_representable_p): New.
28859 (aarch64_output_simd_mov_immediate): Likewise.
28860 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
28861 move immediate case.
28862 * config/aarch64/aarch64.c
28863 (aarch64_const_double_zero_rtx_p): Rename to...
28864 (aarch64_float_const_zero_rtx_p): ...this.
28865 (aarch64_print_operand): Allow printing of new constants.
28866 (aarch64_valid_floating_const): New.
28867 (aarch64_legitimate_constant_p): Check for valid floating-point
28869 (aarch64_simd_valid_immediate): Likewise.
28870 (aarch64_vect_float_const_representable_p): New.
28871 (aarch64_float_const_representable_p): Likewise.
28872 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
28873 (aarch64_output_simd_mov_immediate): New.
28874 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
28875 (*movdf_aarch64): Likewise.
28876 * config/aarch64/constraints.md (Ufc): New.
28877 (Y): call aarch64_float_const_zero_rtx.
28878 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
28880 2013-01-07 Richard Biener <rguenther@suse.de>
28882 PR tree-optimization/55888
28883 PR tree-optimization/55862
28884 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
28885 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
28886 not if it is contained therein.
28888 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28890 * config/avr/t-avr: Typo.
28892 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
28895 * config/avr/t-avr: Don't automatically rebuild
28896 $(srcdir)/config/avr/t-multilib
28897 $(srcdir)/config/avr/avr-tables.opt
28898 $(srcdir)/doc/avr-mmcu.texi
28899 (avr-mcus): New phony target to build them on request.
28900 (s-avr-mlib, s-avr-mmcu-texi): Remove.
28901 * avr/avr-mcus.def: Adjust comments.
28903 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
28905 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
28907 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
28909 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
28911 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
28913 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
28915 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
28917 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
28918 to generate profiling.
28919 * config/rs6000/aix64.h (LIB_SPEC): Same.
28921 2013-01-04 Andrew Pinski <apinski@cavium.com>
28923 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
28925 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
28927 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
28929 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
28931 (ix86_expand_move): Ditto.
28932 (ix86_zero_extend_to_Pmode): Ditto.
28933 (ix86_expand_call): Ditto.
28934 (ix86_expand_special_args_builtin): Ditto.
28935 (ix86_expand_builtin): Ditto.
28937 2013-01-04 Richard Biener <rguenther@suse.de>
28939 PR tree-optimization/55862
28940 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
28941 translating them through PHI nodes.
28943 2013-01-04 Martin Jambor <mjambor@suse.cz>
28945 PR tree-optimization/55755
28946 * tree-sra.c (sra_modify_assign): Do not check that an access has no
28947 children when trying to avoid producing a VIEW_CONVERT_EXPR.
28949 2013-01-04 Marek Polacek <polacek@redhat.com>
28951 PR middle-end/55859
28952 * opts.c (default_options_optimization): Clarify error message.
28954 2013-01-04 Richard Biener <rguenther@suse.de>
28956 PR middle-end/55863
28957 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
28960 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28963 * config/pa/pa.md (movsi): Revert previous change.
28964 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
28967 2013-01-03 Richard Henderson <rth@redhat.com>
28969 * config/i386/i386.c (ix86_expand_move): Always assign to op1
28970 after eliminating TLS symbols.
28972 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
28975 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
28976 * graphite-poly.c (debug_gmp_value): Likewise.
28978 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
28981 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
28982 selected code model, define __code_mode_small__, __code_model_medium__,
28983 __code_model_large__, __code_model_32__ or __code_model_kernel__.
28984 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
28985 xchg temporary register with %k. Declare temporary register as
28987 [__x86_64__]: For medium and large code models, preserve %rbx register.
28989 2013-01-03 Richard Biener <rguenther@suse.de>
28991 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
28992 (dump_subscript): Adjust.
28993 (finalize_ddr_dependent): Do not dump redundant info.
28994 (analyze_siv_subscript): Adjust.
28995 (subscript_dependence_tester): Likewise.
28996 (compute_affine_dependence): Likewise.
28998 2013-01-03 Richard Biener <rguenther@suse.de>
29001 2013-01-03 Richard Biener <rguenther@suse.de>
29003 PR tree-optimization/55857
29004 * tree-vect-stmts.c (vectorizable_load): Do not setup
29005 re-alignment for invariant loads.
29007 2013-01-02 Richard Biener <rguenther@suse.de>
29009 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29010 invariant load do not generate a vector load from the scalar location.
29012 2013-01-03 Richard Biener <rguenther@suse.de>
29014 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29015 for not vectorizing.
29016 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29017 not build INDIRECT_REFs, call get_name once only.
29018 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29019 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29021 2013-01-03 Richard Biener <rguenther@suse.de>
29023 PR tree-optimization/55857
29024 * tree-vect-stmts.c (vectorizable_load): Do not setup
29025 re-alignment for invariant loads.
29027 2013-01-03 Richard Biener <rguenther@suse.de>
29030 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29031 prefer a built-in decl.
29033 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29035 * gcc.c (process_command): Update copyright notice dates.
29036 * gcov.c (print_version): Likewise.
29037 * gcov-dump.c (print_version): Likewise.
29039 PR rtl-optimization/55838
29040 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29041 iv0.step, iv1.step and step.
29043 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29044 Marc Glisse <marc.glisse@inria.fr>
29046 PR tree-optimization/55832
29047 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29048 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29049 integer_{one,zero}_node.
29051 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29054 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29055 * var-tracking.c (reverse_op): Don't add reverse ops to
29056 VALUEs that have already
29057 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29059 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29061 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29063 2013-01-02 Teresa Johnson <tejohnson@google.com>
29065 * dumpfile.c (dump_loc): Print filename with location.
29066 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29067 new location_t parameter to emit complete unroll message with
29068 new dump framework.
29069 (canonicalize_loop_induction_variables): Compute loops location
29070 and pass to try_unroll_loop_completely.
29071 * loop-unroll.c (report_unroll_peel): New function.
29072 (peel_loops_completely): Use new dump format with location
29073 for main dumpfile message, and invoke report_unroll_peel on success.
29074 (decide_unrolling_and_peeling): Ditto.
29075 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29076 by report_unroll_peel.
29077 (decide_peel_completely): Ditto.
29078 (decide_unroll_constant_iterations): Ditto.
29079 (decide_unroll_runtime_iterations): Ditto.
29080 (decide_peel_simple): Ditto.
29081 (decide_unroll_stupid): Ditto.
29082 * cfgloop.c (get_loop_location): New function.
29083 * cfgloop.h (get_loop_location): Declare.
29085 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29087 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29090 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29092 PR middle-end/55198
29093 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29094 BLKmode objects when EXPAND_MEMORY is specified.
29096 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29098 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29100 (fold_builtin_cpu): Do not share cpu model decls across statements.
29102 2013-01-02 Jason Merrill <jason@redhat.com>
29105 * tree.c (build_array_type_1): Revert earlier change.
29107 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29109 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29111 * config/aarch64/aarch64-tune.md: Re-generate.
29113 2013-01-02 Richard Biener <rguenther@suse.de>
29115 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29116 invariant load do not generate a vector load from the scalar location.
29118 2013-01-02 Richard Biener <rguenther@suse.de>
29121 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29122 * configure: Regenerate.
29124 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29126 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29127 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29128 (expand_builtin_int_roundingfn_2): Keep the original target around
29129 for the fallback case.
29131 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29133 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29134 to be clear for sign changes.
29136 2013-01-01 Jan Hubicka <jh@suse.cz>
29138 * ipa-inline-analysis.c: Fix formatting.
29140 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29142 PR tree-optimization/55831
29143 * tree-vect-loop.c (get_initial_def_for_induction): Use
29144 gsi_after_labels instead of gsi_start_bb.
29146 Copyright (C) 2013 Free Software Foundation, Inc.
29148 Copying and distribution of this file, with or without modification,
29149 are permitted in any medium without royalty provided the copyright
29150 notice and this notice are preserved.